Unclassified NEA/CSNI/R(2009)2

Organisation de Coopération et de Développement Économiques

Organisation for Economic Co-operation and Development 23-Sep-2009
___________________________________________________________________________________________
English - Or. English
NUCLEAR ENERGY AGENCY
COMMITTEE ON THE SAFETY OF NUCLEAR INSTALLATIONS
Unclassified
NEA/CSNI/R(2009)2

Risk-Informed Piping Integrity Management

WORKSHOP PROCEEDINGS

Madrid, Spain
2-4 June 2008
Hosted by Consejo de Seguridad Nuclear (CSN)
English - Or. English

JT03270098

Document complet disponible sur OLIS dans son format d'origine

Complete document available on OLIS in its original format
NEA/CSNI/R(2009)2

2
 NEA/CSNI/R(2009)2

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 1st February 1958 under the name of the OEEC European
Nuclear Energy Agency. It received its present designation on 20th 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: [email protected] or by fax (+33-1) 45 24 99 30. 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, 75006
Paris, France, fax (+33-1) 46 34 67 19, ([email protected]) or (for US only) to Copyright Clearance Center (CCC), 222
Rosewood Drive Danvers, MA 01923, USA, fax +1 978 646 8600, [email protected].

3
NEA/CSNI/R(2009)2

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.

4
 NEA/CSNI/R(2009)2

FOREWORD

The Committee on the Safety of Nuclear Installations (CSNI) Working Group on Integrity and Ageing of
Components and Structures (IAGE) has as general mandate to advance the current understanding of those
aspects relevant to ensuring the integrity of structures, systems and components, to provide for guidance in
choosing the optimal ways of dealing with challenges related to the integrity of operating, as well as new
nuclear power plants, and to practice an integrated approach to design, safety and plant life management.

The Working Group comprises three subgroups dealing with (a) integrity and ageing of metal structures
and components, (b) integrity and ageing of concrete structures, and (c) seismic behaviour of components
and structures.

The concept of risk-informed in-service inspection (RI-ISI) has been successfully implemented in several
NEA countries, as reported in the CSNI state-of-the-art report NEA/CSNI/R(2005)9 “Review of
International developments and cooperation on Risk-Informed In-Service-Inspection (RI-ISI) and
Non-destructive Testing (NDT) Qualification in OECD-NEA member countries”.

Various RI-ISI methodologies have been developed, although the predominantly applied methods are those
developed by the Pressurized Water Reactor Owners Group (PWROG)/ASME and by the Electric Power
Research Institute (EPRI) in the USA. There had not been any direct comparisons drawn among different
RI-ISI methodologies applied to an identical scope of components (system, class, etc.). Recommendations
and support for benchmarking various RI-ISI approaches have been addressed by several international
bodies. In response, the CSNI agreed in December 2005 on a proposal to benchmark the different RI-ISI
methodologies (RISMET) in an attempt to identify how they impact reactor safety and whether they lead to
significantly different results.

The RISMET benchmark is a unique comparative study of selected approaches used to set up an ISI
programme. As a result, the knowledge of different approaches and their impact on plant safety is likely to
be enhanced. At the same time, the use of risk-informed ISI will be promoted.

On the other hand, in 2002 the OECD/NEA established the OECD Piping Failure Data Exchange (OPDE)
Joint Project to produce an international database on the piping service experience of commercial nuclear
power plants. The OPDE project also collects information on pipe degradation that could have generic
implications regarding the reliability of in-service inspections.

The OPDE project is therefore improving the quality of data obtained relating to piping degradation
experience, and, in turn, rendering such data more useful in predicting piping failures. The RISMET and
OPDE projects contribute to improving the management of piping inspections.

This report documents the proceedings of a CSNI workshop on risk-informed piping integrity
management. Hosted by the Consejo de Seguridad Nuclear (CSN), the workshop was held in Madrid,
Spain, on 2-4 June 2008. The objectives of this workshop were to examine the results of the RISMET
project and to discuss the results and applications of the OECD Piping Failure Data Exchange (OPDE)
Joint Project along with other related advances taking place in NEA member countries.

5
NEA/CSNI/R(2009)2

ACKNOWLEDGEMENTS

Gratitude is expressed to the Consejo de Seguridad Nuclear (CSN) in Spain for hosting the Workshop and
to Mr. Carlos Castelao for his help, and excellent organisation of the meeting.

Thanks are also expressed to the Workshop Technical Committee, the Session Chairpersons and the
workshop participants for their efforts and cooperation.

Technical Committee

Karen Gott (SSM, Sweden)

Kaisa Simola (VTT, Finland)

Workshop Chairpersons

Carlos Castelao (CSN, Spain)

Arne Eriksson (EC-JRC)

Claude Faidy (EdF, France) – IAGE Metal Sub-Group Chair

Ladislav Horacek (NRI Rez, Czech Republic)

Alejandro Huerta (OECD/NEA)

Tamio Koriyama (JNES, Japan)

Bengt Lydell (Sigma-Phase Inc., USA)

Patrick O’Regan (EPRI, USA)

Jovica Riznic (CNSC, Canada)

Paul Stevenson (Westinghouse, USA)

Zhaojing Zeng (CNSC, Canada)

6
 NEA/CSNI/R(2009)2

SUMMARY AND CONCLUSIONS

I. Introduction

This is a summary report from the “Workshop on Risk-Informed Piping Integrity Management”, held in
Madrid, Spain on 2-4 June 2008. A total of 63 specialists from 12 countries and international organisations
attended. The meeting was sponsored by the Committee on the Safety of Nuclear Installations of the
OECD Nuclear Energy Agency and the European Commission Joint Research Centre (JRC) and was
hosted by the Consejo de Seguridad Nuclear.

The main objectives of the meeting were namely to examine and discuss the results and conclusions of the
OECD/NEA and EC-JRC coordinated risk-informed in-service inspection methodologies benchmark
(RISMET) and to discuss and present the results and applications of the OECD Piping Failure Data
Exchange (OPDE) Joint project along with other related activities sound in NEA member countries.
Participants in the workshop discussed the applications of risk-informed piping integrity management, and
the regulatory-, utility- and industry-related aspects, as well as any related research and development
activities.

The workshop was structured in four technical sessions, each followed by ample time for panel
discussions. The first technical session was devoted to presentation of the RISMET results and
conclusions. The second technical session was intended for discussion of risk-informed in-service
inspection applications in member countries. The third technical session addressed the structure and
development of the OPDE database and its applications. The final, fourth technical session included
presentations on the developments of probabilistic fracture mechanics in support of piping integrity
management.

II. Workshop Background

Risk-informed in-service inspection (RI-ISI) methodologies have been developed in several countries,
although the only widely applied methods are those developed by the Pressurized Water Reactor Owners
Group (PWROG)/ASME and by the EPRI in the USA. Some attempts had been previously made to
compare the different RI-ISI methodologies; however, there had not yet been any direct comparisons
drawn among the use of different RI-ISI methodologies applied to an identical scope of components
(system, class, etc.). Recommendations and support for performing a benchmarking of various RI-ISI
approaches had been delivered by several international groups and committees.

The CSNI and its Integrity and Ageing of Components and Structures Working Group (IAGE WG)
affirmed the need to clearly understand how the selected RI-ISI methodologies may impact reactor safety
and whether they lead to significantly different results, and how the operating experience could then be
used to estimate failure probabilities. In parallel, the European Network for Inspection and Qualification,
ENIQ, and JRC as its operating agent, were developing similar proposals, whilst the necessity for
cooperation between the two organisations was also elaborated.

In consequence, the OECD/NEA CSNI supported the proposal from the working group on Integrity and
Ageing of Components to undertake a joint NEA/EC-JRC initiative for a benchmark study. By means of

7
NEA/CSNI/R(2009)2

benchmarking the different methodologies, it would become possible to identify how they impact reactor
safety and whether they lead to significantly different results. Also, the benchmarking could result in the
identification of those phases in methodologies associated with the greatest potential to affect the
outcome(s), and it might also suggest areas for further improvement.

The RISMET benchmark is a unique comparative study of selected approaches used to set up an ISI
programme. Its results are expected to strengthen knowledge regarding differences in such approaches and
their impact on plant safety, as well as to promote the use of risk-informed ISI.

On the other hand, in 2002, the OECD established the OECD Pipe Failure Data Exchange Project (OPDE)
to produce an international database on the piping service experience applicable to commercial nuclear
power plants. OPDE, which is operated under the umbrella of the OECD Nuclear Energy Agency, collects
pipe failure data, including: service-induced wall thinning, part through-wall cracks, pinhole leaks, leaks,
and rupture/severance (i.e., events involving large leak rates up to and beyond the make-up capacity of
engineered safety systems). The part through-wall events include degradation in excess of code permissible
for pipe wall thinning or crack depth. OPDE also addresses such degradation that could have generic
implications regarding the reliability of in-service inspection.

The OPDE database supports the following activities: 1) trend analysis, including ageing analysis;
2) statistical analyses to determine pipe failure rates and rupture frequencies for use in risk-informed
activities (e.g. loss-of-coolant-accident frequency assessment, internal flooding initiating frequency
assessment, high-energy-line break frequency assessment, RI-ISI ∆-risk assessment); 3) source of data
parameters for input within probabilistic fracture mechanic codes; 4) degradation mechanism analysis
(DMA) in RI-ISI programme development; 5) development of defences against recurring (e.g. systematic)
pipe failures; and 6) exchange of service data in order to pinpoint potential generic implications of a
specific, significant pipe failure.

Therefore, the OPDE project is improving the quality of data pertaining to piping degradation experience
data, and thus yielding such data more efficient in predicting piping failures. Both RISMET and OPDE
projects contribute, amongst other things, to progressing optimised management of piping inspections.

III. Conclusions

Risk-informed in-service inspection (RI-ISI) programme development is widely used in some countries
(for example: Spain and the USA), and based on consequence assessments that rely on plant-specific PSA
model insights and results, RI-ISI has the capability of identifying risk-important inspection locations that
might otherwise be ignored. This is a clear benefit of the RI-ISI approaches, especially in full-scope
applications that not only address ASME Code Class 1 and Class 2 systems, but also consider Code Class 3
and non-Code systems. In some cases reduction in the number of inspections in the primary circuit can be
justified when applying RI-ISI methods, and thus, radiation doses can be significantly reduced. The
resulting economical benefit for the plant by adopting RI-ISI depends on the present ISI scope, rules and
regulations in the specific country. This reduction in scope and dose does not result in the safety margins of
the plant being reduced.
The RI-ISI process itself is a valuable exercise, since it presses the project team to review the piping
degradation potential and identify both direct and indirect consequences of piping failures. This review
may also identify more efficient inspection procedures than are currently used.
It was noted that, in order to benefit the most from a risk-informed approach, the plant-specific PSA model
should be of high quality and also cover low power and shut-down analyses, and internal and external
hazards, including internal flooding and seismic PSA.

8
 NEA/CSNI/R(2009)2

Many RI-ISI initiatives and applications around the world are based on the EPRI and PWROG RI-ISI
methodologies which were originally developed in the US regulatory environment as alternatives to ASME
XI. The regulations are not the same internationally, and therefore, the regulatory authorities have adapted
these methodologies to accommodate the country’s regulatory environment. In addition, so-called
augmented programs are in place in many countries to address some specific degradation mechanisms,
such as flow-accelerated corrosion (FAC), microbiologically influenced corrosion (MIC), and intergranular
stress corrosion cracking (IGSCC). When applying RI-ISI, it is important that the defined owner or
augmented programs are integrated into or coordinated with the RI-ISI programme in a logical manner.
RISMET constituted a good project allowing for the identification of the main differences in the selected
RI-ISI methodologies and assessing how these differences impact reactor safety and whether they lead to
significantly different results. To achieve the project objectives, the benchmark was organised into
Application Groups, responsible for applying various methodologies, and Evaluation Groups, to analyse
the information provided. Four main Evaluation Groups were formed: (1) "Scope of Application";
(2) "Failure Probability Analysis"; (3) "Failure Consequence Analysis"; and (4) "Risk Ranking and Site
Selection". A fifth group, "Regulatory Aspects", which comprised the RISMET members belonging to
regulatory bodies, was formed to address regulatory aspects.
Both the need to document the results comprehensively as well as to disseminate the report to all the NEA
member countries promptly were pointed out, in order for the member countries to be aware of the main
differences in the RI-ISI methodologies and also to improve their inspection programmes.
The OPDE project constitutes a good international project for the collection of pipe degradation and
failures in commercial nuclear power plants in the participating OECD member countries. The database
supports various analyses including trend analysis, aging analysis, statistical analyses to determine pipe
failure rates and rupture frequencies, provision of data parameters for input to probabilistic fracture
mechanics codes, and degradation mechanism analysis in ISI applications.
The necessary procedures and systems are in place for the OPDE project to collect and validate data and
also to ensure that such data is more readily available to the international community (for example,
regarding PSA and RI-ISI applications) and to maintain availability in a controlled manner to the research
community.
It was recommended that the OPDE project continues to ensure a good quality database and that new data
is entered as soon as deemed available.
Probabilistic degradation models are an important element of the probabilistic piping reliability analysis
that needs to be completed with leak and rupture criteria. The pipe failure parameter of interest in RI-ISI
applications is namely the frequency of pipe failure per inspection site and year. Our current
state-of-knowledge with respect to assessments of probability of failure and critical crack size parameters
associated with specific degradation mechanisms (for example, primary water stress corrosion cracking,
PWSCC) point to large uncertainties.
It was noted that a successful implementation of a quantitative structural reliability analysis is strongly
dependent on an in-depth knowledge of structural integrity management and piping system degradation
susceptibilities. Therefore, it was recommended that any application of such structural models should be
performed by practitioners with a deep physical understanding of the degradation mechanism, and
familiarity with identifying and assessing the uncertainties associated with data, models and tools used
(e.g. sensitivity analysis).
The use and application of probabilistic fracture mechanic tools and computer codes should be performed
with caution, by experienced users, with a solid understanding of how they works, their limitations. In
other terms, such devices are not to be treated with limited knowledge, as if they were “black box”. The
predictive power of these models is highly correlated with the underlying assumptions regarding flaw
initiation, flaw growth (propagation), etc. and these aspects have to be clearly understood. It is important

9
NEA/CSNI/R(2009)2

that these computer codes be subjected to extensive validation and verification (V&V) processes of the
physical models, including benchmarking.
It was recommended that the CSNI/IAGE working group develop benchmark problems to increase
confidence in the results of the PFM computer codes and also to address the training and the use of such
computer codes.
Another important aspect to consider is the use of service experience for validation of these structural
computer codes/models. Consistency of code predictions with all the experience data banks, such as
OPDE, is needed and should be an important element of the V&V process.
Continuous applications of international lessons learned from service experience and research and
developments results should be systematically implemented to develop a better understanding of
degradation mechanisms and their associated uncertainties, and thus to further improve and enhance
probabilistic fracture mechanic models and computer codes. It was recommended that international R&D
efforts be encouraged to deepen knowledge in this field.
Some other probabilistic application tools for decision-making aside from those pertaining to piping issues
have been successfully presented such as risk-informed studies for reactor pressure vessels, steam
generators and LBB. It was recommended that efforts be made to compare the results of such applications.
It was also recommended that the NEA/CSNI establish means to ensure that all IAGE benchmarks can be
applicable as a reference by other users in the future.

10
 NEA/CSNI/R(2009)2

Workshop on Risk-Informed Piping Integrity

Madrid, Spain 2-4 June 2008

TABLE OF CONTENTS

FOREWORD 5
ACKNOWLEDGEMENTS 6
SUMMARY AND CONCLUSIONS 7-10
TABLE OF CONTENTS 11-13

OPENING SESSION 15

Piping Integrity Activities at the OECD/NEA 17-25

Mr. A. Huerta, OECD/NEA

Invited Lecture – On the Use of Probabilistic and Reliability Analyses in Risk-Informed 27-45
NPP Piping Integrity Management
F. Simonen and S. Gosselin, PNL, USA

SESSION 1 - RISMET RESULTS AND CONCLUSIONS 47

General Introduction to the RISMET Project 49-60

L. Gandossi, A. Eriksson, EC JRC, and A. Huerta, OECD/NEA

RISMET Project Evaluation of the Scope of Application 61-69

P. Stevenson, Westinghouse

Evaluation of Pipe Failure Probability Analysis in the RISMET Project 71-80

B. Lydell, Sigma-Phase Inc., and T. Schimpfke, GRS

RISMET Project: Comparison of Consequence of Failure Evaluation 81-88

G. Hultqvis, Forsmark and K. Simola, VTT

RISMET Project: Evaluation of Risk Ranking and Site Selection, and Conclusions of 89-99
the Benchmark Exercise
K. Simola, VTT

Regulatory Perspectives for RISMET 101-111

B. Brickstad, SKI, W. Norris, U.S. NRC, and Z. Zeng, CNSC

11
NEA/CSNI/R(2009)2

SESSION 2 - RI-ISI APPLICATIONS 113

Practical Consideration in Developing a Risk-Informed In-Service Inspection Program 115-130

P. O’Regan, EPRI
Application of EPRI Streamlined RI-ISI (Risk-Informed Safety-Based, or RIS_B, 131-144
Codified in ASME Section XI Code Case N-716) to the Entergy Plants
K. W. Hall, Iddeal Solutions, N. G. Cofie and S. T. Chesworth, Structural Integrity
Associates

EDF Alternative Method to Optimise ISI Program 145-152

C. Faidy, EDF–SEPTEN
Current Status and Experience of Application of RI-ISI in Finland 153-167
K. Simola, VTT, O. Hietanen, Fortum Nuclear Services Ltd and P. Kuusinen, Teollisuuden
Voima Oy

RI-ISI Programmes in Spain 169-179

C. Mendoza, F. Olivar and J. M. Figueras, CSN
Study on Risk-Informed In Service Inspection (RI-ISI) for BWR Piping 181-204
T. Koriyama, L. Yinsheng, Y. Hamaguchi, M. Yamashita, M. Hirano, JNES
RI-ISI C.N. Cofrentes Programme 205-216
E. Gutiérrez Fernández, C. Martín-Serrano Ledesm, Iberdrola Ingeniería Y Construcción,
P. Pérez Tejedor, C. Gavilán Moreno, C.N. Cofrentes, C. Cueto-Felgueros García,
Tecnatom S.A

SESSION 3 - OPDE DATABASE AND ITS APPLICATION 217

The OECD Piping Failure Data Exchange Project 219-233

K. Gott, SSM
Application Activities of the OPDE Database 235-243
S. Ogiya, JNES
An Evaluation of Impact of Ageing in OPDE Piping Failure data 245-258
M. D. Pandey and X. X. Yuan, Institute for Risk Research, University of Waterloo,
J. Riznic, CNSC

Progress with the OECD Pipe Failure Data Exchange Project 259-271
K. Gott, SKI, A. Huerta, OECD/NEA, B. Lydell, Sigma-Phase, Inc.

SESSION 4 - PROBABILISTIC FRACTURE MECHANICS 273

Development of Probabilistic Fracture Mechanics Analysis Code for Aged Piping under 275-285
Stress Corrosion Cracking
K. Onizawa and H. Itoh, JAEA
Reliability Parameter Handbook for Piping Components in Nordic Nuclear Power 287-299
Plants
Bengt Lydell, Sigma-Phase Inc., Anders Olsson, Relcon Scandpower AB

12
 NEA/CSNI/R(2009)2

Probabilistic Evaluation in Support of Risk-Informed Pressure Tube Maintenance 301-314

E. Nadeau, AEC, T. Byrne, DRT Technical Services, S. Khajehpour, AEC, D. Leemans
Crosscurrents Research and Policy Consulting

A Structural Reliability Model for the Evaluation of Failure Probabilities of Alloy 315-329
182/82 Welds in PWRs
C. Cueto-Felgueros, Tecnatom
Activities of the PFM Subcommittee of the Japan Welding Engineering Society on 331-353
Probabilistic Fracture Mechanics
N. Maeda, JNES, S. Yoshimura, University of Tokyo, Y. Kanto, University of Ibaraki
NURBIM Fatigue Benchmarking of SRRA Code for Risk-Informed ISI 355-366
E. Milanov, Westinghouse
CANTIA - Probabilistic Code for CANDU Tube Inspection Assessment 367-382
S. T. Revankar and B. Wolf, Purdue University, and J. Riznic, CNSC
Probabilistic Modelling of Primary Water Stress Corrosion Cracking (PWSCC) 383-396
In Piping Dissimilar Metal Butt Welds
B. A. Bishop, Westinghouse
Risk-Informed Leak-Before-Break Reconciliation with Modern Codes and Standards 397-411
for Plant Life Extension
M.J. Kozluk and X. Duan, AECL

LIST OF PARTICIPANTS 414-423

13
NEA/CSNI/R(2009)2

14
 NEA/CSNI/R(2009)2

OPENING SESSION

Piping Integrity Activities at the OECD/NEA

Mr. A. Huerta, OECD/NEA
Invited Lecture – On the Use of Probabilistic and Reliability Analyses in Risk-Informed NPP Piping
Integrity Management
F. Simonen and S. Gosselin, PNL, USA

15
NEA/CSNI/R(2009)2

16
 NEA/CSNI/R(2009)2

17
NEA/CSNI/R(2009)2

18
 NEA/CSNI/R(2009)2

19
NEA/CSNI/R(2009)2

20
 NEA/CSNI/R(2009)2

21
NEA/CSNI/R(2009)2

22
 NEA/CSNI/R(2009)2

23
NEA/CSNI/R(2009)2

24
 NEA/CSNI/R(2009)2

25
NEA/CSNI/R(2009)2

26
 NEA/CSNI/R(2009)2

27
NEA/CSNI/R(2009)2

28
 NEA/CSNI/R(2009)2

29
NEA/CSNI/R(2009)2

30
 NEA/CSNI/R(2009)2

31
NEA/CSNI/R(2009)2

32
 NEA/CSNI/R(2009)2

33
NEA/CSNI/R(2009)2

34
 NEA/CSNI/R(2009)2

35
NEA/CSNI/R(2009)2

36
 NEA/CSNI/R(2009)2

37
NEA/CSNI/R(2009)2

38
 NEA/CSNI/R(2009)2

39
NEA/CSNI/R(2009)2

40
 NEA/CSNI/R(2009)2

41
NEA/CSNI/R(2009)2

42
 NEA/CSNI/R(2009)2

43
NEA/CSNI/R(2009)2

44
 NEA/CSNI/R(2009)2

45
NEA/CSNI/R(2009)2

46
 NEA/CSNI/R(2009)2

SESSION 1
RISMET RESULTS AND CONCLUSIONS

General Introduction to the RISMET Project

L. Gandossi, A. Eriksson, EC JRC, and A. Huerta, OECD/NEA
RISMET Project Evaluation of the Scope of Application
P. Stevenson, Westinghouse
Evaluation of Pipe Failure Probability Analysis in the RISMET Project
B. Lydell, Sigma-Phase Inc., and T. Schimpfke, GRS
RISMET Project: Comparison of Consequence of Failure Evaluation
G. Hultqvis, Forsmark and K. Simola, VTT
RISMET Project: Evaluation of Risk Ranking and Site Selection, and Conclusions of the Benchmark
Exercise
K. Simola, VTT
Regulatory Perspectives for RISMET
B. Brickstad, SKI, W. Norris, U.S. NRC, and Z. Zeng, CNSC

47
NEA/CSNI/R(2009)2

48
 NEA/CSNI/R(2009)2

General Introduction to the Rismet Project

Luca Gandossi, Arne Eriksson JRC–Institute for Energy, Petten, the Netherlands
A. Huerta NEA/OECD, Issy-les-Moulineaux, Paris, France

Abstract

RI-ISI methodologies have been developed in several countries, although the only widely applied methods
are those developed by Westinghouse Owners Group/ASME (WOG) and by EPRI in the USA. The RIBA
project provided a summary and comparison of four different RI-ISI methodologies, but the only case
study was the application of the WOG methodology to two piping systems of Ringhals 4 PWR, and only a
general comparison was made of the WOG approach to the Swedish procedure for RI-ISI. To date, no
direct comparison of several RI-ISI methodologies applied to an identical scope of components had been
made.

To address this shortcoming, the JRC and the OECD/NEA initiated in 2006 a project called RISMET, with
the overall objective to apply various RI-ISI methodologies to the same case (namely, selected piping
systems in one nuclear power plant). The comparative study aimed at identifying the impact of such
methodologies on reactor safety and at studying how the main differences influence the final result (i.e. the
definition of the risk-informed inspection programme).

In the RISMET benchmark project several RI-ISI methodologies were applied to four selected piping
systems in Ringhals 4 PWR plant, selected to cover a broad spectrum of safety classes, degradation
mechanisms, etc. The evaluation of the results was carried out in five groups concentrating on the
following issues: (1) Scope of Application; (2) Failure Probability Analysis; (3) Consequence of Failure
Analysis; (4) Risk ranking and selection of segments/sites to be included in inspection programs; and (5)
Regulatory Aspects. The evaluation included the identification of differences in the RI-ISI applications, the
analysis of the importance of identified differences, and the comparison between RI-ISI and “traditional”
inspection programmes.

This paper gives a general overview of the project, including its background, objectives and organisation.

49
NEA/CSNI/R(2009)2

50
 NEA/CSNI/R(2009)2

51
NEA/CSNI/R(2009)2

52
 NEA/CSNI/R(2009)2

53
NEA/CSNI/R(2009)2

54
 NEA/CSNI/R(2009)2

55
NEA/CSNI/R(2009)2

56
 NEA/CSNI/R(2009)2

57
NEA/CSNI/R(2009)2

58
 NEA/CSNI/R(2009)2

59
NEA/CSNI/R(2009)2

60
 NEA/CSNI/R(2009)2

RISMET Project Evaluation of the Scope of Application

P. Stevenson (Westinghouse)

Abstract

Several piping risk-informed in-service inspection (RI-ISI) methodologies have been developed. To date,
there has been no direct comparison of the various methodologies on the same scope of application. The
RISMET project is a comparison of risk-informed and other piping in-service inspection (ISI)
methodologies on an identical scope. The evaluation of the RISMET project was divided into four main
areas; evaluation of scope of application, analysis of failure probabilities, analysis of consequences, and
evaluation of risk ranking and site selection. The focus of this paper is on the evaluation of the scope of
application.

RI-ISI methodologies can be used to develop full-scope as well as partial scope applications. Thus, one
objective for the evaluation of the scope of application is to assess the impact of implementing various
scopes of application on the ISI results and the technical basis for changes in ISI results, if any, for the
various ISI methodologies. A second objective is to assess the definition of segment boundaries used in
the various methodologies and to determine their impact on the ISI results and the technical basis for
changes in the ISI results, if any. This paper presents the differences between the Pressurised Water
Reactor Owners Group (PWROG) methodology, the Electric Power Research Institute (EPRI)
methodology, ASME (American Society of Mechanical Engineers) Code Case N-761, SKIFS (Swedish
regulatory requirements) and ASME Section XI methodologies compared in the RISMET analysis.

61
NEA/CSNI/R(2009)2

62
 NEA/CSNI/R(2009)2

63
NEA/CSNI/R(2009)2

64
 NEA/CSNI/R(2009)2

65
NEA/CSNI/R(2009)2

66
 NEA/CSNI/R(2009)2

67
NEA/CSNI/R(2009)2

68
 NEA/CSNI/R(2009)2

69
NEA/CSNI/R(2009)2

70
 NEA/CSNI/R(2009)2

Evaluation of Pipe Failure Probability Analysis

in the RISMET Project

Bengt Lydell, Sigma-Phase, Inc., USA

Thomas Schimpfke, Gesellschaft für Anlagen-und Reaktorsicherheit (GRS) mbH, Germany

The treatment of pipe failure probability analysis in the four RISMET applications has been reviewed. The
objective of this evaluation is to document observations, insights and recommendations that can assist an
organization in making sound decisions about technical issues associated with RI-ISI program
development. Ultimately, different organisational cultures, regulatory requirements as well as technical
requirements will influence a decision about how to best pursue a new RI-ISI project.

1. Overview

The evaluation of pipe failure probability analysis included the consideration of the following twelve (12)
technical issues and associated questions:

1. Use of PFM/SRM and specific requirements on codes (e.g., validation and verification), and
including capability to account for different types of degradation mechanisms. What type of
statistics do the SRM models generate?

2. Use of statistical models of pipe failure, including the role and use of service experience data.
What are the quality requirements to be imposed on service experience data? Is it technically
feasible to derive absolute measures of piping reliability based on service experience data?

3. Relationship (or interface) between service experience data and PFM/SRM. Specific issues
addressed include the ranges of applicability of the two approaches. For example, how well does
PFM/SRM predict failures in carbon steel piping subject to general corrosion, erosion-corrosion,
or flow-accelerated corrosion? How well do statistical models extrapolate stress corrosion
cracking data to major structural failure?

4. Use of expert judgment/expert elicitation. Requirements on expert panels, including consistency

issues in the case of RI-ISI program updates. Requirements on the practical use of computer
codes for calculating POF.

5. Treatment of uncertainties; this subtasks addresses the type of statistics that is generated by
respective technical approach.

6. Treatment of different structural failure modes (as defined by peak through-wall flow rates); this
subtask addresses failure modes definitions and how they are addressed in respective
methodology.

7. Probability of flaw detection, POD, and inspection intervals. How does respective methodology
account for the effects of these factors on an inspection scope?

71
NEA/CSNI/R(2009)2

8. Reliability of leak detection and sensitivity of results to leak detection limits. Treatment of
different leak inspection strategies (e.g., frequency of visual inspection and walk down
inspection).

9. Treatment of different degradation mitigation strategies (material, water chemistry).

10. Updating of original failure probability analyses given new service experience, or
implementation of piping design changes or new mitigation strategies.

11. Compatibility of pipe failure probability analysis approach with PSA requirements
(reverse-engineering). Can the pipe failure probability analysis results of RI-ISI directly support
such PSA tasks as loss-of-coolant-accident (LOCA) frequency assessment, or internal flooding
initiating event frequency assessment?

12. Importance of POF in RI-ISI program definition.

2. Summary of insights

This evaluation considered four RI-ISI methodologies. Quantitative structural reliability analysis is not a
pre-requisite for RI-ISI program development. However, a successful implementation of any methodology
is strongly dependent on an in-depth knowledge of structural integrity management and piping system
degradation susceptibilities. The consideration of structural integrity (qualitatively or quantitatively) is one
of several steps needed to develop a RI-ISI program, and it is acknowledged iteratively throughout a
program development process. With some exceptions, a comparison of the EPRI and PWROG
methodologies points to similar POF ranges. The PWROG methodology appears to be more resource
intensive than any of the other RI-ISI methodologies that are included in the RISMET scope of work.
Respective topical report and implementation guidelines address the role and importance of service
experience data in ensuring realistic results and as input to future RI-ISI program updates. Within the scope
of the PWROG-SE application the explicit roles of the plant-specific and industry wide service experience
data could not be assessed, however.

In reviewing the results of the four applications the twelve technical issues concerning pipe failure
probability analysis were considered by the members of the Evaluation Group on Pipe Failure Probability.
Summarized below are the insights and recommendations resulting from the evaluation:

1. Use of PFM/SRM and specific requirements on codes (e.g., validation and verification), and
including capability to account for different types of degradation mechanisms. What type of
statistics do the SRM models generate?
RISMET Perspective: As already stated, application of PFM/SRM codes is not a pre-requisite for
successful RI-ISI program development and implementation. The strengths and limitations of the
currently available PFM/SRM codes are well documented. First-and-foremost, the currently
available structural reliability models are parametric models that support a broad range of sensitivity
studies. The predictive power of these models is highly correlated with the underlying assumptions
about flaw initiation, flaw growth (propagation), etc.

2. Use of statistical models of pipe failure, including the role and use of service experience data. What
are the quality requirements to be imposed on service experience data? Is it technically feasible to
derive absolute measures of piping reliability based on service experience data?
RISMET Perspective: Only one of the four RI-ISI methodologies in the RISMET project included
any explicit consideration of plant-specific or industry service experience data. However, it is noted
that considerable progress has been made in developing statistical models of piping reliability that
directly use service experience data as primary input. A large number of practical applications have

72
 NEA/CSNI/R(2009)2

been performed in support of risk-informed PSA applications. For degradation mechanisms such as
vibration-fatigue and flow-accelerated corrosion the body of service experience data is extensive.
Therefore, predictive piping reliability analysis is not only highly feasible but also recommended
using statistical analysis tools. For a statistical analysis to be sufficiently robust it is necessary to
have access to a validated and sufficiently complete pipe failure event data base that also includes
data on cause and effect relationships for all remedial or mitigation actions that have been performed
on the piping of interest.

3. Relationship (or interface) between service experience data and PFM/SRM. Specific issues
addressed include the ranges of applicability of the two approaches. For example, how well does
PFM/SRM predict failures in carbon steel piping subject to general corrosion, erosion-corrosion, or
flow-accelerated corrosion? How well do statistical models extrapolate stress corrosion cracking
data to major structural failure?
RISMET Perspective: As documented in the NURBIM reports, PFM/SRM is evolving. The results
obtained are highly correlated with underlying assumptions as well as the knowledge and experience
of analysts performing the calculations. It is recommended that analysts involved in PFM/SRM
applications familiarize themselves with the published reports on benchmarking of computer codes
against service experience data.

4. Use of expert judgment/expert elicitation. Requirements on expert panels, including consistency

issues in the case of RI-ISI program updates. Requirements on the practical use of computer codes
for calculating probability of pipe failure (POF).
RISMET Perspective: Expert judgment (in various forms) is an inherent aspect of pipe failure
probability analysis. Applications of the PWROG methodology involve the application of PWROG
user guidelines for the SRRA code. The user guidelines for the PWROG-SE application were not
included in the evaluation scope, however. These plant-specific guidelines were developed because
Ringhals personnel wanted to document any differences relative to the general PWROG guidance so
that they were treating the SRRA input data in a consistent manner for all piping system segments.

5. Treatment of uncertainties; this subtask addresses the type of statistics that is generated by respective
technical approach.
RISMET Perspective: Only the PWROG/PWROG-SE applications include explicit consideration
of uncertainties. Detailed background is documented in WCAP-14572, Revision 1-NP-A.
Depending on type of degradation mechanisms and loading conditions, the uncertainties in estimated
pipe failure probabilities can be substantial and must be properly accounted for in the RI-ISI
program development process.

6. Treatment of different structural failure modes (as defined by peak through-wall flow rates); this
subtask addresses failure modes definitions and how they are addressed in respective methodology.
RISMET Perspective: Different structural failure modes are accounted for in the CC-716, EPRI
and PWROG applications. Although not specifically addressed by RISMET, for plants with a well
developed internal flooding PSA model, the quantitative assessment of pipe failure probability needs
to reflect the unique conditions that are addressed by the internal flooding scenarios. Where SRM
tools are utilized it is essential to verify compatibility between the output and internal flooding
initiating event frequencies.

7. Probability of flaw detection, POD, and inspection intervals. How does respective methodology
account for the effects of these factors on an inspection scope?
RISMET Perspective: The PWROG/PWROG-SE applications account POD and presence leak
detection. As indicated elsewhere, the assumptions about the probability of flaw detection in the
PWROG methodology appear unrealistic.

73
NEA/CSNI/R(2009)2

8. Reliability of leak detection and sensitivity of results to leak detection limits. Treatment of different
leak inspection strategies (e.g., frequency of visual inspection and walkdown inspection).
RISMET Perspective: see #7

9. Treatment of different degradation mitigation strategies (material, water chemistry).

RISMET Perspective: All RI-ISI methodologies include some consideration of different mitigation
strategies. The ability of SRM models to explicitly account for mitigation (e.g., weld overlay,
hydrogen water chemistry, stress relief) is addressed in the open literature but was not addressed
further by the evaluation group.

10. Updating of an original failure probability analysis given new service experience, or implementation
of piping design changes or new mitigation strategies.
RISMET Perspective: Since two of the four RI-ISI methodologies do not include explicit
consideration of service experience data it is unclear how exactly a program update would be
performed given new observations about pipe degradation or failure. It is recognized that the
practical implications (e.g., extent of analytical resources needed) is tied to specific aspects of a
selected RI-ISI methodology (including absolute versus relative risk ranking). The ability of a
quantitative approach to account for new service experience will determine whether an update of
existing analysis is feasible.

11. Compatibility of a pipe failure probability analysis approach with PSA requirements (reverse-
engineering). Can the pipe failure probability analysis results of RI-ISI directly support such PSA
tasks as loss-of-coolant-accident (LOCA) frequency assessment, or internal flooding initiating event
frequency assessment?
RISMET Perspective: The “compatibility issue” is outside the RISMET scope of work. Potential
for synergies between RI-ISI and PSA is related to technical issues #1, #2 and #3. Additionally,
there are a number of significant flood sources beyond piping (e.g. tanks, vessels).

12. Importance of quantitative pipe failure probability analysis in RI-ISI program development.
RISMET Perspective: It is not a pre-requisite of RI-ISI program development to perform
quantitative assessments of pipe failure probability. As explained in the Final Report, many different
factors determine what specific technical approach is selected for RI-ISI program development.

74
 NEA/CSNI/R(2009)2

75
NEA/CSNI/R(2009)2

76
 NEA/CSNI/R(2009)2

77
NEA/CSNI/R(2009)2

78
 NEA/CSNI/R(2009)2

79
NEA/CSNI/R(2009)2

80
 NEA/CSNI/R(2009)2

RISMET Project: Comparison of Consequence of Failure Evaluation

Göran Hultqvist, Forsmark, Kaisa Simola, VTT, Finland

Abstract

In the RISMET benchmark project various RI-ISI methodologies were applied to four selected piping
systems in Ringhals 4 PWR plant. The objective was, through the benchmarking, to identify how different
RI-ISI approaches impact reactor safety and whether they lead to significantly different results. The
evaluation of the results was carried out in four groups concentrating in following issues: 1) Scope of
application; 2) Failure Probability Analyses; 3) Consequence analyses; and 4) Risk ranking, classification
and selection of segments/sites to be included in inspection programs. Evaluation included the
identification of differences in the RI-ISI applications, the analysis of the importance of identified
differences, and the comparison between RI-ISI and “traditional” inspection programmes.

The main tasks of the third evaluation group were to identify and analyse the differences in the process of
evaluating consequence of failure in the various RI-ISI methodologies and to identify which demands are
placed on the Probabilistic Assessment Study (PSA) from a RI-ISI point of view. This paper summarises
the results of such analysis.

81
NEA/CSNI/R(2009)2

82
 NEA/CSNI/R(2009)2

83
NEA/CSNI/R(2009)2

84
 NEA/CSNI/R(2009)2

85
NEA/CSNI/R(2009)2

86
 NEA/CSNI/R(2009)2

87
NEA/CSNI/R(2009)2

88
 NEA/CSNI/R(2009)2

RISMET Project: Evaluation of Risk Ranking and Site Selection, and

Conclusions of the Benchmark Exercise

Kaisa Simola, VTT, Finland

1. Evaluation of risk ranking and inspection site selection

The evaluation group work was organised in two phases. In the first phase, the risk ranking and inspection
site selection principles in various methodologies were reviewed and summarised. The outcome of the first
phase was a summary document serving as background material for the evaluation group work.

The main tasks of the evaluation group were to identify and analyze the differences in the process of risk
ranking and inspection site selection, and in the results. The risk-informed inspection site selections were
compared to a selection based on “traditional” inspection programs. Specific attention was given to the
importance of these differences.

The results were compared both qualitatively and quantitatively. In the qualitative evaluations the
following aspects were considered:

• How well the different methodologies identify the same high-risk sites/segments.
• Where are the major differences?
• How other aspects are included in the selection of sites (high consequence, low probability;
low consequence high probability; other aspects).
• Treatment of possible risk outliers.

Table 1 summarises the number of inspection sites in each system in each application. It can be noticed
that ASME XI application leads to significantly larger number of inspections in the Reactor Coolant
System than any of the other approaches. In the case of the Residual Heat Removal system, SKIFS and the
Code Case N-716 applications result in very few inspections (Class 1 part of the system), while the other
applications result in 20-40 inspections. The detailed analyses of the reasons for the selections are
discussed in the final report at system level.

The treatment of the Condensate System was somewhat problematic, since the main degradation
mechanism is the flow accelerated corrosion (FAC), and the plant has a separate program for FAC
inspections. In practice FAC is excluded from the RI-ISI process in US, and the analysis only points to the
plant’s augmented programme which should be in line with the existing regulations. The system was
however analyzed in the Ringhals RI-ISI project, and some FAC susceptible piping is included in the
Swedish regulatory ISI program.

Quantitative evaluations were made from different perspectives. One comparison was made on the basis of
the delta risk analysis using the PWROG decision criteria. Another comparison of the site selections was
made by using the EPRI risk matrix and bounding values that can be used in the EPRI methodology.

89
NEA/CSNI/R(2009)2

Table 1: Summary of number of inspection sites in applications

RCS (313) RHR (321) MS (411) CS (414) Total

ASME XI 113 30 28 0 171
SKIFS 1994:1 40 1 29 0 70
PWROG 28 21 10 + FAC 24+FAC 83
orig 4 (+4 VT2) (+4 VT2 +FAC)
PWROG-SE 28 35 3 + FAC 0+FAC 66 (+FAC)
full
PWROG-SE 4 30 35 3 + FAC 0+FAC 68 (+FAC)
EPRI base 49 26 2 + FAC 3+FAC 80 (+FAC)
EPRI R4 49 39 15 + FAC 3+FAC 106 (+FAC)
CC N-716 49 4 7 + FAC 0+FAC 60 (+FAC)

Differences between a risk-informed and a deterministic approach are usually easily explained by the
different philosophies in the approaches. The differences between the results of two RI-ISI methodologies
were also identified and explanations for the differences were sought.

2. Conclusions on the benchmark exercise

Even if the scope of the benchmark was limited to four systems, the variety in these systems regarding the
safety class, potential degradation mechanisms and pipe break consequences ensured a good coverage of
issues for a comparative study. The risk-informed methodologies have some significant differences, and
result in slightly different risk ranking and selection of inspection sites. However, the preliminary results of
the benchmark indicate that the risk impact of these differences is small, and the RI-ISI approaches can
identify safety important piping segments that would be ignored by approaches not using PSA. The results
of the RISMET benchmark improve the knowledge on differences in approaches and their impact on plant
safety, and promote the use of risk-informed ISI.

Due to the limited resources and time for the benchmark, some limitations had to be accepted in the study.
The study did not allow the benchmarking of various PSA models. It would be of interest to identify the
effect of limitations in the PSA coverage and degree of detail in RI-ISI results. The benchmark indicated
that assumptions made in PSA analyses on e.g. human actions and system success criteria may have a
significant impact on the quantification of consequences and thus to the risk ranking and inspection site
selection. Another topic that was not addressed in the benchmark was the detailed definition of the
inspection program, including inspection methods and intervals, and considerations of doses. Both of these
issues could be considered as potential topics for further research.

90
 NEA/CSNI/R(2009)2

91
NEA/CSNI/R(2009)2

92
 NEA/CSNI/R(2009)2

93
NEA/CSNI/R(2009)2

94
 NEA/CSNI/R(2009)2

95
NEA/CSNI/R(2009)2

96
 NEA/CSNI/R(2009)2

97
NEA/CSNI/R(2009)2

98
 NEA/CSNI/R(2009)2

99
NEA/CSNI/R(2009)2

100
 NEA/CSNI/R(2009)2

Regulatory Perspectives for RISMET

B Brickstad (Swedish Nuclear Power Inspectorate), W. Norris, (U.S. Nuclear Regulatory Commission),
and Z. Zeng (Canadian Nuclear Safety Commission)

Abstract

Frameworks have been developed for applying probabilistic risk assessment (PRA) in reactor regulation.
Policies have been established by several regulators to address issues related to risk assessment in
risk-informed regulation and risk management. Policies have also been established providing general
guidance on the regulatory use of risk assessment methods in a risk-informed regulatory framework. For a
regulatory decision relying upon risk insights as one source of information, there needs to be confidence in
the PRA results from which the insights are derived. Consequently, the PRA needs to have the proper
scope and technical attributes to give an appropriate level of confidence in the results used in the regulatory
decision-making. From a regulatory perspective, the U.S. and European approaches are in agreement on
most of the items that an RI-ISI program should address. This paper describes the above and a set of
recurrent issues that arise in the use of PRAs for risk-informed decision making. The regulatory actions
that have been taken to begin addressing these recurrent issues are discussed. Finally, further research
efforts that are needed in the field of risk-informed selection of components for inspection are identified.

101
NEA/CSNI/R(2009)2

102
 NEA/CSNI/R(2009)2

103
NEA/CSNI/R(2009)2

104
 NEA/CSNI/R(2009)2

105
NEA/CSNI/R(2009)2

106
 NEA/CSNI/R(2009)2

107
NEA/CSNI/R(2009)2

108
 NEA/CSNI/R(2009)2

109
NEA/CSNI/R(2009)2

110
 NEA/CSNI/R(2009)2

111
NEA/CSNI/R(2009)2

112
 NEA/CSNI/R(2009)2

SESSION 2
RI-ISI APPLICATIONS

Practical Consideration in Developing a Risk-Informed In-Service Inspection Program

P. O’Regan, EPRI
Application of EPRI Streamlined RI-ISI (Risk-Informed Safety-Based, or RIS_B, Codified in ASME
Section XI Code Case N-716) to the Entergy Plants
K. W. Hall, Iddeal Solutions, N. G. Cofie and S. T. Chesworth, Structural Integrity Associates

EDF Alternative Method to Optimise ISI Program

C. Faidy, EDF – SEPTEN
Current Status and Experience of Application of RI-ISI in Finland
K. Simola, VTT, O. Hietanen, Fortum Nuclear Services Ltd and P. Kuusinen, Teollisuuden Voima Oy

RI-ISI Programmes in Spain

C. Mendoza, F. Olivar and J. M. Figueras, CSN
Study on Risk-Informed In Service Inspection for BWR Piping
T. Koriyama, L. Yinsheng, Y. Hamaguchi, M. Yamashita, M. Hirano JNES
RI-ISI C.N. Cofrentes Programme
E. Gutiérrez Fernández, C. Martín-Serrano Ledesm, Iberdrola Ingeniería Y Construcción, P. Pérez
Tejedor, C. Gavilán Moreno, C.N. Cofrentes, C. Cueto-Felgueros García,Tecnatom S.A

113
NEA/CSNI/R(2009)2

114
 NEA/CSNI/R(2009)2

Practical Considerations in Developing a Risk-Informed In-Service Inspection Program

Patrick O’Regan, Electric Power Research Institute (EPRI), USA

1. Background

Risk-informed inservice inspection (RI-ISI) is now being implemented at nearly one hundred nuclear units
in a number of countries with additional RI-ISI applications and pilot studies are also underway in several
more countries. These RI-ISI applications have ranged from focused scope applications (e.g. Class 1
piping only) to full scope applications (e.g. Class 1, 2 and 3 piping). While some of these RI-ISI
applications have been limited to the deterministic inspection program other RI-ISI applications have
integrated the RI-ISI program with additional plant inspection programs (e.g. break exclusion region). As
RI-ISI programs mature, the importance of properly developing and maintaining these programs has
become clear.

2. Discussion

Based upon the experiences, there are a number of practical considerations and lessons learned that
increase the effectiveness of the RI-ISI application as well as streamline the resources necessary to develop
and maintain a RI-ISI program.

Examples of these practical considerations include:

• Interaction with the regulatory body

• Plant management support
• Composition of the RI-ISI team
• Defining existing inservice inspection requirements and bases
• Determination of the scope of the RI-ISI application
• Timing of the application
• Plant-specific probabilistic risk assessment (PRA) requirements
• Operating conditions and history
• Inspection history, access and capabilities
• Proper Documentation
• Living Program Requirements

Each of the above considerations contains an element of “data gathering and processing”. This includes
relatively straightforward tasks such as plant isometric, flow and ISI drawings as well as more complicated
tasks such as the appropriate use of the plant-specific PRA. However, frequently the three most important
considerations are 1) the interaction with the appropriate regulatory bodies, 2) the definition, (including its

115
NEA/CSNI/R(2009)2

technical and licensing basis) of existing in-service inspection requirements and 3) the proper
documentation of the application.

In addition to the above, the plant’s overall risk management philosophy is fundamental to developing a
successful RI-ISI program. This is true not only for development of the initial RI-ISI program but is also
important for the long term viability of the RI-ISI program as well as extension of this technology to other
plant programs and components (e.g. risk-informed repair / replacement, long term operation).

With this in mind, experience has shown that RI-ISI evaluations of piping systems often identify segments
or structural elements as safety-significant that cannot be inspected effectively or that other measures or
means may be more useful in managing plant risk. One example concerns components which are subject
to degradation mechanisms such as vibratory fatigue, which can develop very fast, and usually in
components that are difficult to inspect because of materials or design (e.g., socket welds). Another
example could be piping failures that result in high consequences due to lack of, or insufficient, plant
procedures. A third example could be use of plant modifications to address risk significant sites (e.g.
barriers, thermal transient monitoring) in lieu of inspections. Such occurrences suggest that plant owners
and regulatory bodies should adopt a wider perspective risk management versus in-service inspection
alone.

It is also important to investigate alternative possibilities for mitigation against the risk. It is necessary to
look at the nature of the risk associated with each site (or segment). It may be possible to identify ways
other than inspection to address the risk, such as continuous monitoring, improved leak rate detection,
improved water chemistry treatment and follow-up, load reduction or component replacement.

From a lessons learned perspective and as plants remain in service longer than originally planned (e.g. long
term operation, license renewal), it is essential for RI-ISI to address new degradation mechanisms that may
be discovered while not invalidating the previous inspection program. This would be done as part of the
living program that requires an evaluation update on a periodic basis when new information is obtained
from plant or industry experience. It is this type of approach that will be useful in managing generic aging
issues discovered in piping systems.

Experience has also shown that RI-ISI programs can only be successful if they are cost-effectively
integrated into the plants engineering processes and procedures. As more and more plants have adopted
risk-informed programs the following have been identified as critical in assuring “programmatic” success:

1. Having a risk-informed program with an immature PRA or too sensitive to PRA changes can lead to
large changes in required examinations as a result of the required evaluation update process.
Changes to the inspection population as a result of the updating process should be minimal, unless
there is a substantial physical change to the plant or a new type damage mechanism is identified.

2. The program should be developed, documented and maintained in a method that supports future
updates to the program. Adopting a RI-ISI program at a utility should not add significant burden to
the station. Once the RI-ISI program is in place and functioning, maintaining the program current
should be possible without frequent expenditures of plant manpower and resources.

3. Plant staffs need to have sufficient training and involvement in the development of the RI-ISI
program. They must fully understand, at all levels of station management, what is expected of the
program and what is necessary to maintain the program valid.

4. To manage resources the industry must strive to identify new mechanisms, aggressively investigate,
and quickly develop criteria allowing them to be subsumed by the RI-ISI program, as appropriate.

116
 NEA/CSNI/R(2009)2

3. Conclusion

RI-ISI is an unquestionable success. It has targeted inspections to critical locations while allowing utilities
to decrease resources expended on inspection of non risk significant locations. As the industry continues
to improve (i.e. safety, reliability) it will be possible, and maybe necessary, to build on the risk-informed
ISI success and expand risk informed technologies into other inspection and testing programs.

The next challenge facing the industry is development of RI-ISI programs for new reactor designs. The
new reactor designs have advanced PRA analysis and software as part of the original plant documentation.
However, current RI-ISI processes used by the industry have not been written to accommodate these new
reactor designs where operational experience is limited. Overcoming these issues and allowing the use of
RI-ISI programs for new reactor designs is the next industry goal in RI-ISI technology.

117
NEA/CSNI/R(2009)2

118
 NEA/CSNI/R(2009)2

119
NEA/CSNI/R(2009)2

120
 NEA/CSNI/R(2009)2

121
NEA/CSNI/R(2009)2

122
 NEA/CSNI/R(2009)2

123
NEA/CSNI/R(2009)2

124
 NEA/CSNI/R(2009)2

125
NEA/CSNI/R(2009)2

126
 NEA/CSNI/R(2009)2

127
NEA/CSNI/R(2009)2

128
 NEA/CSNI/R(2009)2

129
NEA/CSNI/R(2009)2

130
 NEA/CSNI/R(2009)2

Application of EPRI Streamlined RI-ISI (Risk-Informed Safety-Based, Or RIS_B, Codified In

ASME Section XI Code Case N-716) to the Entergy South Plants

Kevin W. Hall, Iddeal Solutions, USA

Nathaniel G. Cofie and Scott T. Chesworth, Structural Integrity Associates, USA

1. Background

Traditional EPRI RI-ISI, as codified in ASME Code Cases N-5601 and N-5782, has been approved by the
NRC and proven on numerous plant applications (i.e. greater than 70 percent of the US fleet). In all cases,
application of the EPRI methodology has been proven to significantly reduce the number of piping
inspections with negligible impact on plant risk.

EPRI’s Streamlined RI-ISI approach, which has been codified as ASME Code Case N-7163 (risk-informed
safety-based, or RIS_B), pre-defines high safety significant (HSS) piping components (Class 1, Break
Exclusion Region (BER), and some additional Class 2 piping welds, as well as any piping segment whose
contribution to CDF is > 1E-06 or LERF is > 1E-07, regardless of piping Class) based upon industry
exhaustive experience with RI-ISI applications. As such, the detailed consequence evaluation of the
traditional approach (see above) does not need to be performed; instead, only a failure potential evaluation
of the HSS components is required. The results of this evaluation, along with insights from the plant
internal flooding study, are used to determine the appropriate sampling of the weld population that will
constitute the RIS_B program (approximately 10 percent of the HSS population). The RIS_B technology
results in reduced inspections over a standard Section XI ISI program while maintaining or improving
plant risk. It also provides the additional cost savings over traditional RI-ISI of not having to perform a
consequence evaluation of the in-scope piping. This allows for cost savings both in the initial application
as well as in performing the required periodic RI-ISI program updates.

The failure potential (degradation mechanism) evaluation of the HSS components is identical to that used
in traditional EPRI RI-ISI (CC N-560 and N-578). Degradation mechanisms evaluated include Thermal
Fatigue (TASCS and Thermal Transients), Stress Corrosion Cracking (Intragranular, Transgranular,
External Chloride and Primary Water), Localised Corrosion (MIC, Pitting), and Flow-Assisted Corrosion
(Erosion-Cavitation, FAC).

Once the failure potential evaluation has been performed, elements are selected on the following basis:
• 10% of the HSS piping welds shall be selected for examination, prorated by system
o IGSCC Category B through G exams can be credited toward 10% requirement
o FAC exams not credited toward 10% requirement

1
American Society of Mechanical Engineers Code Case N-560-2, “Alternative Examination Requirements for Class 1, Category
B-J Piping Welds, Section XI, Division 1,” March 28, 2000.
2
American Society of Mechanical Engineers Code Case N-578-1, “Risk-Informed Requirements for Class 1, 2, or 3 Piping,
Method B, Section XI, Division 1,” March 28, 2000.
3
American Society of Mechanical Engineers Code Case N-716, “Alternative Piping Classification and Examination Requirements,
Section XI, Division 1,” April 19, 2006.

131
NEA/CSNI/R(2009)2

• 25% of the population identified as susceptible to each degradation mechanism (DM) or DM

combination shall be selected, including:
o 2/3 of Reactor Coolant Pressure Boundary (RCPB) welds must be selected between the Reactor
Pressure Vessel (RPV) and first isolation valve
o 10% of RCPB welds outside containment shall be selected
o 10% of the welds within the Break Exclusion Region shall be selected

Selections may be adjusted upward or downward to obtain a 10% overall sampling.

2. Status

Plant-specific NRC Safety Evaluation Reports (SERs) for two pilot plant studies, one BWR (Grand Gulf,
an Entergy South plant) and one 2-unit PWR, were received from the NRC in late 2007. In addition, the
plant-specific SER was received for one follow-on plant application, at Waterford Unit 3 (an Entergy
South plant). Two additional follow-on plant applications are in process, both for Entergy South plants:
Arkansas Nuclear One Unit 1 and River Bend Station.

EPRI plans to present Revision 1 of Code Case N-716 to the Working Group on the Implementation of
Risk-Based Examination for consideration during the upcoming 3rd quarter ASME Code Committee
meetings scheduled for August 2008. Code Case N-716-1 will incorporate all of the lessons learned from
the RIS_B pilot and follow-on applications.

It is the author's understanding that the NRC intends to conditionally approve Code Case N-716 in
Revision 16 of Regulatory Guide 1.1474. A draft guide is expected to be issued for public comment
sometime during 2008. Final issuance of Revision 16 of Regulatory Guide 1.147 is expected before the
end of 2008.

3. Plant data

Entergy South Plant Data is shown in Table 1.

Table 1: Entergy South Plant Data

Waterford Unit 3 Arkansas Nuclear Grand Gulf Nuclear River Bend Station
Name (WF3) One Unit 1 (ANO1) Station (GGNS) (RBS)
Entergy Louisiana, Entergy Arkansas, System Energy Entergy Gulf States,
Owner Inc. Inc. Resources Inc.

Reactor Type PWR PWR BWR-6 BWR-6

Combustion
Reactor Manufacturer Engineering Babcock and Wilcox General Electric General Electric

T/G Manufacturer Westinghouse Westinghouse Kraftwerk Union General Electric

Architect/Engineer Ebasco Bechtel Power Bechtel Power Stone and Webster

Commercial Operation Date September 1985 December 1974 July 1985 June 1986

License Expiration Date December 2024 May 2034 November 2024 August 2025

4
Regulatory Guide 1.147, “Inservice Inspection Code Case Acceptability,” American Society of Mechanical Engineers Boiler and
Pressure Vessel Code, Section XI, Division 1, Revision 15, October 2007.

132
 NEA/CSNI/R(2009)2

4. Failure potential results

The failure potential analysis results for Waterford Steam Electric Station Unit 3 are as follows: The
Reactor Coolant System, Charging System, Safety Injection System (which contains the Residual Heat
Removal System), Feedwater System, and Emergency Feedwater System (where branching off the
Feedwater System) all experienced Thermal Fatigue under various normal and upset plant conditions. The
Reactor Coolant System and Safety Injection System also experienced PWSCC, and the Safety Injection
System additionally experienced IGSCC. Some Feedwater and Emergency Feedwater System piping
components were also scheduled for examination under the plant FAC Program.

The failure potential analysis results for Arkansas Nuclear One - Unit 1 are as follows: The Reactor
Coolant System, Makeup and Purification System, Decay Heat Removal System, and Main Steam System
all experienced Thermal Fatigue. The Reactor Coolant System and Decay Heat Removal System also
experienced PWSCC, and the Makeup and Purification System experienced IGSCC. Some Main
Feedwater System piping components were also scheduled for examination under the plant FAC Program.

The failure potential analysis results for Grand Gulf Nuclear Station are as follows: The Reactor Pressure
Vessel System, Feedwater System, Residual Heat Removal System, and High Pressure Core Spray System
all experienced Thermal Fatigue. The Reactor Pressure Vessel System and Reactor Recirculation System
had some IGSCC Category B through G components scheduled for examination under the plant IGSCC
Program. The Residual Heat Removal System also experienced some Crevice Corrosion.

The failure potential analysis results for River Bend Station are as follows: The High Pressure Core Spray
System, Feedwater System, and Reactor Pressure Vessel System all experienced Thermal Fatigue. The
Reactor Pressure Vessel System had some IGSCC Category B through G components scheduled for
examination under the plant IGSCC Program. Also, some Feedwater, Drains, Reactor Core Isolation
Cooling, Main Steam, and Reactor Water Clean-up System piping components were scheduled for
examination under the plant FAC Program.

5. Element selection results

Tables 2 – 5 show the Element Selection Results for Waterford Steam Electric Station Unit 3, Arkansas
Nuclear One - Unit 1, Grand Gulf Nuclear Station, and River Bend Nuclear Power Station, respectively.

Table 2: Element selection results for WF3

Number of ASME Section XI Number of CC N- CC N-716 Total Number of
Weld Class Code Section XI Percentage 716 Exams Percentage Welds
Exams
1 194 26.6 74 10.2 729

2 140 8.4 25 1.5 1674

3 0 0.0 0 0.0 0

Non 0 0.0 0 0.0 4

Totals 334 13.9 99 4.1 2407

133
NEA/CSNI/R(2009)2

Table 3: Element selection results for ANO1

Number of ASME Section XI Number of CC N- CC N-716 Total Number of

Weld Class Code Section XI Percentage 716 Exams Percentage Welds
Exams
1 104 25.1 44 10.6 414
2 145 8.7 17 1.0 1661
3 0 0.0 0 0.0 0
Non 6 100.0 2 33.3 6
Totals 255 12.3 63 3.0 2081

Table 4: Element selection results for GGNS

Number of ASME Section XI Number of CC N- CC N-716 Total Number of

Weld Class Code Section XI Percentage 716 Exams Percentage Welds
Exams
1 248 28.2 99 11.3 880

2 75 8.0 8 0.8 942

3 0 0.0 2 18.2 11

Non 0 0.0 0 0.0 1

Totals 323 17.6 109 5.9 1834

Table 5: Element selection results for RBS

Number of ASME Section XI Number of CC N- CC N-716 Total Number of

Weld Class
Code Section XI Percentage 716 Exams Percentage Welds
1 191 25.1 83 10.9 762

2 120 9.0 0 0.0 1338

3 0 0.0 0 0.0 4

Non 0 0.0 0 0.0 0

Totals 311 14.8 83 3.9 2104

Summary and conclusions

EPRI streamlined RI-ISI (as codified in ASME Code Case N-716) uses exhaustive industry experience to
pre-define HSS piping segments without the need for a detailed consequence of failure evaluation. The
methodology covers most critical piping scope (Class 1, some Class 2 and BER), and requires less effort to
apply and maintain than previous EPRI RI-ISI approaches. As can be seen from the studies reviewed in this
Summary, the end result is a significant reduction of ISI examinations. Since the change from the existing
ASME Code Section XI ISI program to the new RI-ISI program must meet the requirements of US NRC
Regulatory Guide 1.174, it must be shown to result in a risk-decrease, a negligible risk increase, or be
risk-neutral to the plant. Therefore, this methodology can be applied while maintaining or improving plant
safety.

134
 NEA/CSNI/R(2009)2

135
NEA/CSNI/R(2009)2

136
 NEA/CSNI/R(2009)2

137
NEA/CSNI/R(2009)2

138
 NEA/CSNI/R(2009)2

139
NEA/CSNI/R(2009)2

140
 NEA/CSNI/R(2009)2

141
NEA/CSNI/R(2009)2

142
 NEA/CSNI/R(2009)2

143
NEA/CSNI/R(2009)2

144
 NEA/CSNI/R(2009)2

EDF Alternative Method to Optimise ISI Program

C. Faidy (EDF – SEPTEN)

Abstract

EdF has studied different ways to optimise the ISI of the class 1 components in particular; similar process
is done for class 2-3 pipings.

The principle used to-day is:-

− to define criteria for "concerned" locations to fatigue or flaw tolerance

− and inside this list to define "sensitive" locations that have to be inspected

For performance and frequencies different complementary analysis are done on sensitivity studies.

The major differences with RI-ISI are:-

− no direct consideration of safety consequences

− only damage oriented, no particular CDF or LERF in this approach
− more damage rate analysis are done

Why? Again Safety is the 1st criteria to define ISI, but degradation appearance can be a reason to shut
down the plant for repair and replacement, without safety concern.

The paper will develop the criteria and corresponding list of analyses.

145
NEA/CSNI/R(2009)2

146
 NEA/CSNI/R(2009)2

147
NEA/CSNI/R(2009)2

148
 NEA/CSNI/R(2009)2

149
NEA/CSNI/R(2009)2

150
 NEA/CSNI/R(2009)2

151
NEA/CSNI/R(2009)2

152
 NEA/CSNI/R(2009)2

Current Status and Experience of Application of RI-ISI in Finland

Kaisa Simola, VTT, Finland

Ossi Hietanen, Fortum Nuclear Services Ltd, Finland
Petri Kuusinen, Teollisuuden Voima Oy, Finland

The presentation describes the RI-ISI activities in Finland. First, the regulatory requirements related to
risk-informed in-service inspections are summarised. Second, the status of the on-going Finnish RI-ISI
projects and some experiences are reported. Finally some RI-ISI related research activities and
international co-operation are presented.

1. Regulatory requirements

The regulatory requirements are expressed in the Finnish Regulatory Guides, so called YVL-guides. The
regulatory guides that are related to the risk-informed decision making and in-service inspections are:

− YVL 2.8 “Probabilistic safety analysis in safety management of nuclear power plants” and

− YVL 3.8 “Nuclear power plant pressure equipment. In-service inspection with non-destructive
testing methods”

These two guides require the use of risk-informed approaches in development of inspection programmes.

YVL 2.8 that has come into force in May 2003, states as follows1: “The results of PSA shall be used in the
drawing up and development of the inspection programmes of piping as per Guide YVL 3.8. Combining
the information from PSA and the damage mechanisms of pipes and the secondary impacts of damages, the
inspections are focused in such a way that those are weighted temporarily and quantitatively on those pipes
whose risk significance is greatest. While drawing up the risk informed inspection programme, the systems
of classes 1, 2, 3, 4 and EYT (not safety related) must be regarded as a whole. Similarly how far the
radiation doses can be reduced by focusing inspections and optimising inspection periods shall be
regarded.” YVL 3.8, in force since September 2003 states2: “In the drawing up of inspection programmes
for Safety Class 1, 2, 3 and 4 piping and Class EYT (non-nuclear) piping as well as in the development of
inspection programmes for operating plants, risk-informed methods shall be utilised to ascertain the
inclusion in the inspection scope of those components posing the highest risk.”

1
YVL 2.8 “Probabilistic safety analysis in safety management of nuclear power plants”, 28 May 2003,
http://www.stuk.fi/saannosto/YVL2-8e.html
2
YVL 3.8 “Nuclear power plant pressure equipment. In-service inspection with non-destructive testing methods”, 22 Sept. 2003,
http://www.stuk.fi/saannosto/YVL3-8e.html

153
NEA/CSNI/R(2009)2

2. Pilot studies to prepare for RI-ISI applications

In order to test the applicability and resource needs of RI-ISI, pilot studies have been conducted on limited
scopes both at the regulatory body and at the utility side. For instance insights from the STUK pilot study3
were utilised in the revision of the above-mentioned Finnish regulatory guides. Also VTT has conducted
own pilot studies. The pilot studies tested approaches resembling to some extent to the so-called EPRI
methodology. Typical to the studies is the quantitative use of plant specific PSA models. STUK and utility
pilot studies have used a qualitative evaluation of piping failure potential, while VTT has also tested
quantification with structural reliability models.

3. RI-ISI project for Loviisa

A plan for utilising RI-ISI for Loviisa Unit 1 was prepared in 2005, and it was accepted by STUK in July
2006. The principles follow to some extent the ASME XI Supplement R, Method B (“EPRI
methodology”), but with some important differences4:

− The RI-ISI scope is the whole unit, not only for piping systems of safety classes 1, 2 and 3.
Screening is done on the basis of system CDF/LERF: all systems having a contribution larger than
10-6 to CDF or 10-7 to LERF are included in the scope.

− The consequence assessment is based on the results of the extensive PSA analysis, covering
e.g. flood analyses and low power states.

− The failure assessment is qualitative, but using different criteria than the US approach

− An independent expert panel is evaluating the risk ranking, and the panel work is monitored by
STUK’s observers.

The RI-ISI analysis for the Loviisa Unit 1 systems has recently been completed, and the new RI-ISI
programme has been submitted to STUK for review and acceptance. The experience from the RI-ISI
application indicates that several new features will be brought to the future ISI programme:

− Many totally new systems and new portions of the system of the existing programme are included

− Small diameter instrumentation piping of the primary systems are to be inspected, implying the
development of new methods and techniques

− Consequence differences of parallel redundant safety systems and system portions can often be
discovered, mostly due to fire and flood

Risk-informed approach for all piping systems of plant requires a huge amount of work. The benefits of the
work are maybe not gained from the reduction of inspection scope, but by hopefully reducing the total risk.
The radiation exposure seems to be clearly reduced.

3
Mononen, J., I. Niemelä, R. Virolainen, R. Rantala, A. Julin, O. Valkeajärvi, J. Hinttala, A (2000), Pilot Study on Risk-Informed
In-Service Inspection, Proceedings of the 5th International Conference on Probabilistic Safety Assessment and Management
(PSAM5). Osaka, JP, 27 Nov. - 1 Dec. 2000. Kondo, S. & Furuta, K. (eds.). Vol. 2/4, Universal Academy Press. Tokyo (2000),
349 – 355.
4
Hietanen, O., Jänkälä, K., Bergroth, N., Paussu, R. & Nikula, V. (2007), Risk-Informed Methodology of New ISI Program for
Unit 1 of Loviisa NPP, the 6th International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurised
Components, 8 – 10 October, Budapest, Hungary.

154
 NEA/CSNI/R(2009)2

4. RI-ISI activities at Olkiluoto

TVO has started the RI-ISI application for Olkiluoto 1 Unit. The adopted RI-ISI approach follows largely
the same principles as Loviisa RI-ISI: full scope, extensive use of plant PSA model, qualitative analysis of
failure potential, and the use of independent expert panel to review the risk ranking. The classification of
degradation potential is different from both Loviisa and EPRI approaches. Since IGSCC one of the main
expected degradation mechanisms in BWRs, more resolution in classification of IGSCC susceptible piping
is adopted compared to EPRI. Another difference in the approach is that degradation evaluation is done at
weld level already in the risk ranking phase. The RI-ISI analyses utilise the plant pipeline database, which
includes e.g. piping dimension, material and loading information and allows various searches and analyses
to support the evaluation and presentation of RI-ISI results.

5. Public research and international co-operation

RI-ISI is also a research topic in the Finnish Research Programme on Nuclear Power Plant Safety
2007-2010, SAFIR20105. The on-going research work includes studies related to structural reliability
modelling, inspection reliability and RI-ISI follow-up and updating. Many research activities are connected
to international projects and networks. VTT is participating actively to the European Network for
Inspection and Qualification (ENIQ) Task Group Risk6 [7] work and to the OECD/NEA-JRC coordinated
RI-ISI benchmark project RISMET. Also VTT is contributing to the development of an IAEA technical
document on RI-ISI.

5
http://virtual.vtt.fi/safir2010/
6
http://safelife.jrc.nl/eniq/tgr.php

155
NEA/CSNI/R(2009)2

156
 NEA/CSNI/R(2009)2

157
NEA/CSNI/R(2009)2

158
 NEA/CSNI/R(2009)2

159
NEA/CSNI/R(2009)2

160
 NEA/CSNI/R(2009)2

161
NEA/CSNI/R(2009)2

162
 NEA/CSNI/R(2009)2

163
NEA/CSNI/R(2009)2

164
 NEA/CSNI/R(2009)2

165
NEA/CSNI/R(2009)2

166
 NEA/CSNI/R(2009)2

167
NEA/CSNI/R(2009)2

168
 NEA/CSNI/R(2009)2

RI-ISI Programmes in Spain

Carlos. Mendoza, Francisco. Olivar, and José. M. Figueras (CSN)

Consejo de Seguridad Nuclear, Justo Dorado, 11, 28040 Madrid

Abstract

Since 2000 five applications for Risk Informed In-Service Inspection has been proposed by the Spanish
nuclear industry as an acceptable alternative to the ASME Code examination requirements for piping
inspection, including considerations as both deterministic and risk aspects.

After a preliminary stage, which were developed the Spanish criteria and methodology for the preparation
by the industry and for assessment by the CSN of RI-ISI applications, through a joint guideline UNESA
and CSN, which was verified through pilot applications in some systems of BWR and PWR NPP, the
featured programs have used quantitative methodology described in the Guidelines developed jointly by
CSN and UNESA.

Up to now, the RI-ISI applications approved in Spain have always been of partial scope, so that the first
four in the Spanish NPP Almaraz I & II and Asco I & II, all of the PWR type, have maintained their scope
limited to Class 1 piping. The latest application, submitted by Cofrentes NPP, BWR-GE type, has had as
scope classes 1 & 2 piping.

The aim of this paper is to highlight a number of important lessons learned, obtained after a comparison of
the results in the five applications approved in Spain, for the deterministic and probabilistic aspects of
RI-ISI.

169
NEA/CSNI/R(2009)2

170
 NEA/CSNI/R(2009)2

171
NEA/CSNI/R(2009)2

172
 NEA/CSNI/R(2009)2

173
NEA/CSNI/R(2009)2

174
 NEA/CSNI/R(2009)2

175
NEA/CSNI/R(2009)2

176
 NEA/CSNI/R(2009)2

177
NEA/CSNI/R(2009)2

178
 NEA/CSNI/R(2009)2

179
NEA/CSNI/R(2009)2

180
 NEA/CSNI/R(2009)2

Study on Risk-Informed In-Service Inspection (RI-ISI) for BWR Piping

Tamio Koriyama, Li Yinsheng, Yoshikane Hamaguchi, Masahiro Yamashita, Mitsumasa Hirano

Japan Nuclear Energy Safety Organisation, Japan

1. Failure modes and effects analysis of OPDE data

Piping failure incidents for BWRs were analyzed by means of Failure Mode and Effects Analysis (FMEA)
which is a process for identifying the failure modes of specific components and evaluating their effects on
other components and systems. A failure mode can be defined by failure to start, failure to run, or a leak.
This study focuses on such failure to run and leaks, since piping systems are passive and there are no
failures to start. The OECD-NEA Piping Failure Data Exchange (OPDE) Project has collected piping
failure events and kept such events up to date. This study used 1,673 piping failure incidents occurring at
BWRs as contained in the OPDE database (version as of 29 June 2007).

A piping degradation is caused by a mechanism that attacks the piping material and which may adversely
affect piping integrity. Therefore, piping failure events can be categorised by the degradation mechanisms
concerned. As listed in the right column of Table 1, there are 42 types of degradation mechanisms in the
aforementioned OPDE database for BWRs. These degradation mechanisms can be categorized by
characteristics into the 14 types listed in the left column of Table 1.

Table 1: Classification of degradation mechanisms

Classified degradation mechanisms. Degradation mechanisms occurred in the OPDE. Database for BWRs
Erosion Erosion, Erosion-cavitation, Cavitation
Corrosion Corrosion, External Corrosion, Graphitic Corrosion, Bimetallic Corrosion, Contamination
Flow Accelerated Corrosion (FAC) or Erosion-corrosion FAC, Erosion-corrosion
Microbiologically Induced Corrosion (MIC) MIC
Intergranular Stress Corrosion Cracking (IGSCC) IGSCC, Inter Dendritic Stress Corrosion Cracking (IDSCC)
Transgranular Stress Corrosion Cracking (TGSCC) TGSCC, External Chloride Induced Stress Corrosion Cracking (ECSCC)
High Potential Stress Corrosion Cracking (HPSCC), Hydrogen-Induced SCC,
Other SCC
Strain-rate Induced Corrosion Cracking (SICC)
Thermal Fatigue (TF) TF, TF-Cycle, TF-Stratification
High-Cycle Fatigue High-Cycle Fatigue, Vibration
Low-Cycle Fatigue Fatigue, Corrosion Fatigue
Fretting Fatigue Fretting
Over-pressurization, Severe Overloading, Overstress, Water Hammer, Excessive Vibration,
Severe Overloading
External Impact
Construction/Installation Error, Design Error, Fabrication Error, Human Error, Repair/Maintenance Error,
Human Factor (HF)
Welding Error, In-Service Inspection Program failure
Unknown Unknown, Unreported
14 types in total 41 types in total

For effects analysis, the extent of leakage was used for the first-stage decision of loss of relevant system
function or not. Additionally, impacts on operation and event categories described in the OPDE database
for BWRs were referred for the final-stage decision. In such a way, systems having lost their functions due
to pipe failures caused by degradation mechanisms were extracted from the pipe failure events.

181
NEA/CSNI/R(2009)2

Systems subject to piping failure evaluation at a typical BWR5 plant are those subject to Level 1 PSA for
internal events during power operation reflecting the effects of installed Accident Management facilities.
They are shown in Table 2.

Table 2: Target systems

Control Rod Drive Hydraulic Control System (CRDHS) Standby Liquid Control System (SLCS)
Reactor Core Isolation Cooling System (RCIC) High Pressure Core Spray System (HPCS)
Low Pressure Core Spray System (LPCS) Residual Heat Removal System (RHR)
Primary Containment Venting System (PCVS) Power Conversion System (PCS)
High Pressure Core Spray Service Water System (HPCSSW) Residual Heat Removal Service Water System (RHRSW)
Alternative Injection System (AIS) －

The systems exposed to any of the 14 types of degradation mechanisms in the left column of Table 3 can
be identified from the OPDE database for BWRs. And then these piping failure events per system and
degradation mechanism are adopted.

Table 3: Application of piping failure events per system and degradation mechanism

Target systems at BWR5

CRDHS

HPCS

PCVS
SLCS

LPCS
RCIC

CCW

CSW
RHR

PCS

AIS
Erosion { { { {
Corrosion { { { { {
FAC or Erosion-
{ { { { { {
corrosion
MIC {
IGSCC { { { { { { { {
TGSCC { { { { { { { {
Other SCC { { {
TF { { { { { {
High-Cycle Fatigue { { { { { { { { { {
Low-Cycle Fatigue { { { { {
Fretting Fatigue
Severe Overloading { { { { { { {
HF { { { { { { { { {
Unknown { { { { {
{ Failure events recorded in the OPDE database for BWRs
The possibility of occurrence cannot be completely denied in terms of the characteristics of the
degradation mechanism concerned.

182
 NEA/CSNI/R(2009)2

There are two kinds of effects from piping failures: Direct Effect (possibly resulting in loss of relevant
system function), and Indirect Effect (such as pipe whip and spilled water that may cause loss of function
in another system). The Direct effect is evaluated in this study.

2. Evaluation of piping rupture probability

2.1 Piping rupture frequency per system and degradation mechanism

In considering that FMEA and the evaluation described in Chapter 1 focus on the loss of system function,
the types of failure are represented by piping ruptures. The piping rupture frequencies per system and
degradation mechanism to be applied to the BWR5 plant are basically expressed as (number of
occurrences) / (reactor-years of operation), with Bayesian transform being conducted using a likelihood
function to reflect past data collected in Japan (available from the OPDE database for BWRs). Figure 1
shows the calculation flow.

Prior distribution
Distribution pattern of foreign BWRs
(1) Assumption of lognormal distribution
(2) EF = 30
(3) Mean = (number of occurrences) / (reactor-years of operation),
where, number of occurrences = number of rupture events [when system function is lost],
number of occurrences = mean value when one rupture is assumed at 95th percentile

[when system function is not lost]

Likelihood function
Application of domestic BWRs data
(1) Adoption of Poisson distribution
(2) Past rupture frequency = (number of occurrences) / (reactor-years of operation)
where, number of occurrences = number of rupture events
[when system function is lost]
number of occurrences = 0 [when system function is not lost]

Bayesian transform (using ReDAT code)

Posterior distribution
Piping rupture frequencies per system and degradation mechanism of domestic BWRs

Figure 1: Flow of calculating piping rupture frequencies by the Bayesian transform

183
NEA/CSNI/R(2009)2

2.2 Calculation of piping segment rupture probabilities

Piping rupture frequencies per system and degradation mechanism determined through Bayesian transform
are assigned to piping segments in consideration of the location trends of each degradation mechanism
found in the OPDE database for BWRs.

With regard to considerations other than the installation locations of components, such as the
environmental conditions of internal fluids, elbows, weld joints, and weld heat-affected zones, the
associated location trends of degradation mechanisms are organised in terms of pipe size. This is because
information about location trends is not necessarily clear in the OPDE database for BWRs, but information
about pipe size locations is clear in both the OPDE database for BWRs and the BWR5 plant.

Tables 4 and 5 take samples of FAC showing how to consider the location trends and organise them in
terms of pipe size.

Table 4: Methods of reflecting rupture frequencies into piping segments for FAC

Location trends of each degradation mechanism Reflection of rupture frequencies into piping
segments
y Pumps (10%) y Rupture frequencies are assigned to segments
y Downstream side of valve (15%) in each location according to location trends
y Downstream side of orifices (10%) shown at the left (50% of total).
y Heaters, heat exchangers, etc. (15%) The remaining 50% is assigned to each segment
according to the ratio of the numbers of rupture
events in each pipe size shown in Table5.

Table 5: Numbers of piping degraded conditions in terms of pipe sizes for FAC

Nominal diameter Numbers of degraded conditions Ratio

600mm～ 6 3.6%
300mm～600mm 35 21.0%
80mm～300mm 63 37.7%
～80mm 63 37.7%

Piping segment rupture frequencies are multiplied by the test interval or mission time of the relevant
system to get the probabilities. They are used in the evaluation of effects of accident sequences.

3. Evaluation of effects on accident sequences

Level 1 PSA for internal events during power operation is utilized as a basis for evaluating the effects of
ruptures in piping segments on accident sequences, based on the Westinghouse Owners Group (WOG)
methodology 1 in the U. S. This base model used the domestic data on the failure rates of components
(excluding pipes). The base model has no modelling of ruptures in piping segments due to much

1
K. R. Balkey et al., “Westinghouse Owners Group Application of Risk-Informed Methods to Piping Inservice Inspection Topical
Report,” WCAP-14572, Revision 1-NP-A, February 1999.

184
 NEA/CSNI/R(2009)2

lower rupture probabilities than those of other components. Actually, about 40% of segments were
excluded due to truncation limits in the case of modelling them as basic events in the PSA, according to
another JNES analysis.

Failures of components having equivalent effects on accident sequences as that of ruptures in piping
segments are selected as surrogates. In Figure 2, which shows the piping segments diagram of the High
Pressure Core Spray System, the HPCS pump is a surrogate component for the piping segment P04, P07 to
P13, P14-1, P14-2, or P15 to P23.

When some ruptures in piping segments cause initiating events and one of the relevant initiating events is
called Z, the core damage frequency (CDF) is divided by the occurrence frequency of initiating event X to
determine the Conditional Core Damage Probability (CCDP) by using the following equation:

CCDPIE = Z = CDFIE = Z / FREQIE = Z

where,

CCDPIE = Z : CCDP when a rupture of the target piping segment becomes a cause of initiating
event Z.

CDFIE = Z : CDF when a rupture of the target piping segment becomes a cause of initiating
event Z. (/reactor-year)

FREQIE = Z : Initiating event occurrence frequency for initiating event Z (/reactor-year)

The CCDP determined by using above equation is then multiplied by the rupture frequency of the piping
segment as shown in the following equation to determine the CDF associated with a rupture in the relevant
piping segment.

CDFPB , IE = Z = FR PB , IE = Z ⋅ CCDPIE = Z

where,

185
NEA/CSNI/R(2009)2

Figure 2: Piping segment diagram of High Pressure Core Spray System (operating)

186
 NEA/CSNI/R(2009)2

CDFPB , IE = Z : CDF associated with a rupture in a piping segment that causes initiating
event Z. (/reactor-year)

FR PB , IE = Z : Rupture frequency of the piping segment where the rupture causes initiating
event Z (/reactor-year)

When ruptures in piping segments cause only a loss of function in the mitigating systems, the base CDF is
subtracted from CDF when the relevant surrogate component has failed (rupture probability = 1) to
determine an increase in CDF, which means an increase in CDF that occurs when the rupture probability of
a piping segment changes from 0 to 1. The following equation is used for this calculation:

∆CDFPB = CDFPB =1 − CDFBASE

where,

CDFPB =1 : CDF when a target piping segment is damaged (/reactor-year)

CDFBASE : CDF when the piping segment is integral (/reactor-year)

CDFPB=1 is calculated by assuming that the rupture probability of the surrogate component modeled in
Level 1 PSA for internal events during power operation is 1, and using general-purpose code NUPRA1 for
the analysis of fault trees and event trees. CDFBASE is derived from the base model that has no modeling of
ruptures in piping segments, and thus it is the CDF when the rupture probability of the piping segment is 0.
The increase in CDF determined by above equation is then multiplied by the rupture probability of the
piping segment to determine the CDF associated with a rupture in the relevant piping segment, as shown in
following equation.

CDFPB = FPPB ⋅ ∆CDFPB

where,

CDFPB : CDF associated with a rupture in the piping segment that only causes a loss of
function in mitigating systems (/reactor-year)

FPPB : Rupture probability of the target piping segment

∆CDFPB : Difference in CDF between when the target piping segment is ruptured and when
it is integral (/reactor-year)

When some ruptures in piping segments simultaneously cause initiating events and loss of function in
mitigating systems and thereby showing synergetic effects, one of the relevant initiating events is assumed
to be called Z and a surrogate component is selected from the mitigating systems. Using the event tree for
this initiating event Z, the CDF when the surrogate component in the mitigating systems has failed (failure
probability = 1) is calculated by means of the NUPRA code. This CDF is divided by the occurrence
frequency of initiating event Z to determine the CCDP by using the following equation:

1
“WinNUPRA Version 2.0 USER’S MANUAL,” SCIENTECH Inc., April 2000.

187
NEA/CSNI/R(2009)2

CCDPIE = Z , seg =1 = CDFIE = Z , seg =1 / FREQ IE = Z

where,

CCDPIE = Z , seg =1
: CCDP when a rupture of the target piping segment becomes a cause of
initiating event Z and loss of function in mitigating systems, and the surrogate
component in the mitigating systems fails.

CDFIE = Z , seg =1
: CDF when a rupture of the target piping segment becomes a cause of
initiating event Z, and the surrogate component in the mitigating systems fails.
(/reactor-year)

FREQIE = Z : Initiating event occurrence frequency for initiating event Z (/reactor-year)

The CCDP determined by using the above equation is then multiplied by the rupture frequency of the
piping segment as shown in the following equation to determine the CDF associated with a rupture in the
relevant piping segment.

CDFPB , IE = Z = FR PB , IE = Z ⋅ CCDPIE = E , seg =1

where,

CDFPB , IE = Z
: CDF associated with a rupture in a piping segment that simultaneously
causes initiating event Z and loss of function in mitigating systems, thereby showing a
synergetic effect (/reactor-year)

FR PB , IE = Z
: Rupture frequency of the piping segment where the
rupture simultaneously causes initiating event Z and loss of function in mitigating
systems, thereby showing a synergetic effect (/reactor-year)

In the relevant BWR5 plant, the initiating events assumed to be caused by piping ruptures are a large
LOCA, medium LOCA, and small LOCA. These corresponding piping segment rupture frequencies are
then added up. Since there are no accident sequences that can lead to core damage when a small LOCA
occurs, it is conservatively assumed that the frequency obtained here is the upper limit (95th percentile) of
lognormal distribution, and that the mean of the lognormal distribution is the occurrence frequency of a
small LOCA. The reactor-years in the OPDE database for BWRs are adopted. Those of a large and
medium LOCA are extrapolated in accordance with WASH-1400.2 That of a large LOCA is assumed to be
a tenth of that of a small LOCA; that of a medium LOCA is assumed to be the geometric average of those
of a large LOCA and small LOCA.

Table 6 lists the evaluation results of the frequencies of LOCA in comparison with estimated values by
means of the previous methodology. By assuming that one occurrence of a small LOCA was the upper
limit of lognormal distribution and the average value of the lognormal distribution was adopted for the
occurrence of a small LOCA, the previously estimated values were determined by dividing the average

2
“Reactor Safety Study: An Assessment of Accident Risks in U. S. Commercial Nuclear Power Plants,” WASH-1400
(NUREG-75/014), U. S. Nuclear Regulatory Commission, October 1975.

188
 NEA/CSNI/R(2009)2

value of the lognormal distribution by the reactor-years of BWR operation in the U.S. and Japan. Those of
a large and medium LOCA were extrapolated in accordance with WASH-1400. The estimated values are
almost equivalent to the previously estimated values.

Table 6: Frequencies of LOCA

BWR5 (Unit: / reactor-year)

Previous evaluation This evaluation
LOCA frequency Large 1.7E-05 Large 1.7E-05
Medium 5.4E-05 Medium 5.5E-05
Small 1.7E-04 Small 1.7E-04

The CDF derived from Level 1 PSA for internal events during power operation, which is used as the base
model in the relevant BWR5 plant, is 3.6E-09/reactor-year.

The CDF caused by piping ruptures is determined as the sum of the core damage frequencies caused by
ruptures in individual piping segments for all piping segments as shown in the following equation. The
CDFBASE _ LINE
sum becomes the baseline for evaluating risk significance.

CDFBASE _ LINE = ∑ CDFPB

ALL

where,

CDFBASE _ LINE
: CDF caused by piping ruptures (/reactor-year)

CDFPB CDF caused by ruptures of individual piping segments (/reactor-year)

The CDF caused by piping ruptures in the relevant BWR5 plant as determined by using the above
CDF BASE _ LINE
equation, , is 6.8E-10/reactor-year. This is lower than the base CDF of 3.6E-09/reactor-year.
This indicates that the ratio of contribution is low even when also considering the CDF due to piping
ruptures in the base PSA model.

4. Risk significance indicators

Fussell-Vesely (FV) importance, Risk Achievement Worth (RAW), and Risk Reduction Worth (RRW) can
be used as risk significance indicators for the validation of inspection objects, items, and other aspects of
systems and components. With p denoting the piping rupture probability, the FV importance, RAW, and
RRW of the relevant piping segment are expressed by equations below based on their definitional
equations.

CDFBASE _ LINE − CDFPB , p =0

FV =
CDFBASE _ LINE

CDFPB , p =1
RAW =
CDFBASE _ LINE

189
NEA/CSNI/R(2009)2

CDFBASE _ LINE
RRW =
CDFPB , p =0
where,
CDFPB , p =0 : CDFPB when rupture probability of the target piping segment is 0
CDFPB , p =1 : CDFPB when rupture probability of the target piping segment is 1

From above equations, the FV importance and RRW have a mathematical correlation as shown in the
equation below.

1
RRW =
1 − FV

Greater RRW and FV result in greater effect of reduced CDF by deceased piping rupture probability when
chances to prevent a rupture are increased by an inspection which can detect degradations. They can be
used as indicators for selecting a piping segment for which there is an urgent need to strengthen the
management of inspections. This study uses RRW because it indicates the ratio of CDF in a simpler
expression.

RAW indicates the extent of an increase in CDF when a relevant piping segment has been ruptured by
degradation. Larger RAW calls attention during an inspection to a higher increase in CDF. It is an
indicator in terms of expecting careful oversight or treatment through inspections.

5. Categorisation criteria and restriction criteria

A U. S. industry guideline3 says, “RRW may be used directly to identify risk significant Structures,
Systems, and Components (SSC). An SSC would probably be considered risk significant if its RRW
exceeds 0.5% of the overall Core Damage Frequency (RRW > 1.005).” and “An SSC would probably be
considered risk significant if its RAW shows at least a doubling of the overall Core Damage Frequency.”
The categorization criteria and restriction criteria of RRW and RAW above are cited in the EPRI PSA
Applications Guide,4 which is referenced in the WOG methodology. Note that RAW is used on a relative
basis in the WOG methodology.

The assumption of the relevant piping segment having high safety significance when RRW exceeds 1.005
means, on the other hand, that the safety significance is low when the ratio of contribution of CDF due to a
rupture in the relevant piping segment to the total core damage frequency due to a pipe rupture
CDFBASE _ LINE
( ) is 0.5% or less. This value is much lower than 1% and thus can be considered appropriate
as the categorization criteria and restriction criteria for the risk significance indicators of piping segments.
The WOG methodology specifies that risk significance should be left to an expert panel when RRW is in a
range of 1.001 to 1.004.

Based on the considerations above, this study adopts the following criteria for RRW values as the
categorization criteria and restriction criteria for the risk significance indicators of piping segments.

RRW > 1.005 :High safety significance

3
“Industry Guideline for Monitoring the Effectiveness of Maintenance at Nuclear Power Plants,” Nuclear Energy Institute,
NUMARC 93-01, Revision 3, July 2000.
4
“PSA Applications Guide,” Electric Power Research Institute, EPRI TR-105396, August 1995.

190
 NEA/CSNI/R(2009)2

1.001 ≤ RRW ≤ 1.005 :Medium safety significance

RRW < 1.001 :Low safety significance

How to treat RAW as a significance indicator is a task for the future.

In addition to risk significance indicators, and because a degradation mechanism may determine the
effectiveness and necessity of inspections, rupture importance as determined by degradation mechanisms is
used to classify piping segments. Piping segments with known and active degradation mechanisms such as
FAC or those likely to cause a rupture are classified as high rupture importance, while other piping
segments are classified as low rupture importance.

Based on the study of the OPDE database for BWRs and the WOG methodology selecting FAC, IGSCC,
TF, and High-Cycle Fatigue etc., as known and active degradation mechanisms or those likely to cause a
rupture, high rupture importance degradation mechanisms are selected as listed in Table 7. Table 7 also
lists those classified as low rupture importance.

Table 7: Categorisation of rupture importance

Degradation mechanisms classified

High rupture Erosion, Corrosion, FAC or Erosion-corrosion, MIC, IGSCC, TGSCC, TF,
importance High-Cycle Fatigue
Low rupture Other SCC, Low-Cycle Fatigue, Fretting Fatigue, Severe Overloading, HF,
importance Unknown

As shown in Figure 3, risk significance can be classified into various ranges by using a graph with the
horizontal axis denoting RRW, the vertical axis denoting rupture importance, the categorization criteria of
the risk significance indicator (meaning RRW = 1.001 and RRW = 1.005), and the categorisation of
rupture importance as listed in Table 7.

Figure 3 also shows the evaluation results of the risk significance ranges obtained. Piping segments related
to LOCA occurrences as initiating events or piping segments of Primary Containment Venting System are
included in Risk significance ranges 1, 2, 5, and 6 in which safety significance is high or medium, and
indicate relatively high significance in terms of safety. For almost all piping segments, RAW far exceeds 2,
which is the categorisation criteria and restriction criteria in U. S. industry. It substantiate that RAW tends
to be higher when the rupture probability is assumed to be one because of the much lower piping rupture
probability than that of components.

6. Study on RI-ISI program for piping

6.1 Inspection method

Current inspections require leak examinations for almost all pipes and add non-destructive examinations
for pipes deterministically judged as being high in safety significance.

191
NEA/CSNI/R(2009)2

Figure 3: Evaluation results of risk significance ranges

192
 NEA/CSNI/R(2009)2

Also for RI-ISI, leak examinations are assumed to be required for almost all pipes. And pipes belonging to
Risk Significance Range 1, 2, and 5 (See Figure 3) are assumed to be those with high safety significance.
And, to the piping segments in Risk Significance Range 1, 2, and 5, leak examination and non-destructive
examination are applied.

Even though the changes from the current examinations to RI-ISI gave a reduction of piping segments
subject to both leak and non-destructive examinations and the reduction represents around 29% of the total
piping segments in the BWR5 plant, further investigation will be needed reflecting the view that consensus
standards developed by academics and societies, etc. may require that Risk Significance Range 3 be
subject to non-destructive examinations in consideration of its degradation mechanism.

6.2 Extent-of-examination for piping structural elements

The ratio of piping structural elements subject to non-destructive examination to the total number of piping
structural elements in a certain piping segment is defined as the extent-of-examination. The segments P17
and P18 shown in Figure 2 for High Pressure Core Spray System were selected as object segments
included in Risk Significance Range 1.

The Perdue-Abramson model5, that is the statistical sampling method in the WOG methodology, was
adopted. The Perdue-Abramson model is used to determine the minimum extent-of examination not
exceeding the target leak frequency provided by the WOG methodology based on operating experience.

The resulted extent-of-examination in this trial analysis was 20% for P17 and 26% for P18 as shown in
Figure 4, respectively. They showed lower or slightly higher than the extent-of-examination 25% which is
specified in the current inspection procedure. This trial analysis contributes to enhancing a full
understanding of extent-of-examination in terms of risk.

7. Issues to be addressed

There are issues to be addressed below.

1. How to treat a RAW as a significance indicator is to be assessed.

2. Risk impact is to be assessed after the implementation of RI-ISI.
3. The use of Probabilistic Fracture Mechanics model is to be tried for piping segment rupture
probability analyses. The effects of probability of detection are to be considered. By these activities,
JNES will perform the RI-ISI studies applying a full scope of the WOG methodology.
4. Evaluations of indirect effects from piping ruptures are to be assessed.
5. Uncertainty or incompleteness in the PSA is to be considered.
6. Applications of consensus standards developed by academics and societies, etc. may be considered
to Risk Significance Range 3 which has high rupture importance and low safety significance.
7. Assessments of seismic events are to be needed. Using seismic PSA, piping segment safety significance
is to be identified and risk significance indicators, categorization and restriction criteria to be investigated.
8. Aging effects in the evaluation of piping segment rupture probability are to be addressed.

5
“Technical Elements of Risk-Informed In-service Inspection Programs for Piping (Draft Report),” NUREG-1661, U. S. Nuclear
Regulatory Commission, January 1999.

193
NEA/CSNI/R(2009)2

1.4E-04
Segment leak frequency (per segment per year)

Leak frequency of segment P18

Leak frequency of segment P17
1.2E-04
Target segment leak frequency

1.0E-04

8.0E-05

6.0E-05

4.0E-05

2.0E-05

0.0E+00
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60%
Segment extent-of-examination (%)

Figure 4: Relations between extent-of-examination and leak frequencies in piping segments

194
 NEA/CSNI/R(2009)2

195
NEA/CSNI/R(2009)2

196
 NEA/CSNI/R(2009)2

197
NEA/CSNI/R(2009)2

198
 NEA/CSNI/R(2009)2

199
NEA/CSNI/R(2009)2

200
 NEA/CSNI/R(2009)2

201
NEA/CSNI/R(2009)2

202
 NEA/CSNI/R(2009)2

203
NEA/CSNI/R(2009)2

204
 NEA/CSNI/R(2009)2

RI-ISI C.N. Cofrentes Program

Eduardo Gutiérrez Fernández, Cristina Martín-Serrano Ledesma - Iberdrola Ingeniería Y Construcción,

Porfirio Pérez Tejedor, Carlos Gavilán Moreno - C.N. Cofrentes
Carlos Cueto-Felgueros García - Tecnatom S.A

Abstract

Risk Informed In-Service Inspection (RI-ISI) is a new program of In-Service Inspection that arises to
provide an acceptable alternative to the requirements currently in force (ASME–Section XI), including
considerations of risk promoted both by the regulatory organisms and by the industry. These considerations
take into account industry experience: failures due to different degradatory mechanisms (FAC, SCC,
vibration, thermal fatigue, etc) were studied at an important number of plants. Principal results from this
study are the following:

• 1036 pipe failures with leaks below 5 g.p.m.

• 69 failures with leaks above 50 g.p.m.

Most of these failures were not due to fatigue. ASME XI was focused to deal with fatigue at design and the
thing is that most of the failures have been caused for other degradatory phenomena: vibration,
erosion-corrosion, etc.

Thus, the aspects both related to the safety and the exploitation costs are rationalised and optimised
obtaining clear advantages and allowing to focus the inspection program on the locations with major
impact on the safety, allowing, at the same time, the optimisation of resources and the minimisation of the
exposure to the radiation.

The program uses the quantitative methodology described in the Guide developed by CSN-UNESA, that is
applicable to:

• NDE programs at class 1, 2, 3 pipes and non-class pipes.

• ASME XI and augmented programs.
• Sample size, inspection method and inspection frequency.

And its principal elements are the following:

• Proposed ISI program changes definition.

• Engineering assessment
• Control and implementation programs development.
• Change definition.

205
NEA/CSNI/R(2009)2

Cofrentes NPP has been the pioneer in Spain in the use of the quantitative methodology for the set of the
Class-1 and 2 piping.

The objective of the presentation is to emphasize some of the methodological important aspects of it, as
well as to stand out the problems associated and its solutions, some of which are explained below:

− Development of the methodology for Identification of Indirect Consequences in Class-2 piping, that
requires a specific analysis of each line evaluating possible damages for spraying, flooding, pipe whip,
and in particular the treatment of break exclusion areas.

− Postulated degradation mechanisms and their treatment according to the probabilistic fracture
mechanics code (WinPRAISE) once analysed and evaluated by the Technical Panel. The
erosion-corrosion degradatory phenomena has been treated individually, not using WinPRAISE code.

− Problems associated to the stresses analysis of Class-2 small piping which receives a particular
treatment from the one used for the Class-1.

− Defense in-depth aspects taking into account at the project, in particular those that are aimed to avoid
possible couplings between degradatory mechanisms: for example, between fatigue and
erosion-corrosion.

− Other important items:

• Cofrentes PRA is used as an instrument of Risk-Informed Categorisation.

• Check and validation by the Plant Expert Panel in the Integrated Decision-making Process, assisted
by a Technical Panel formed by responsible in charge of degradatory mechanisms identification,
WinPRAISE input data validation, etc, and that are specialist in stress analysis, in-service
inspection, software related to this project treatment, etc.

• An informatics special structure support for this particular project has been needed due to the project
organisation complexity.

• RI-ISI is a dynamic program:

o The program will be revised with a ten years frequency that meets Section XI ASME Code
intervals.
o Feedback of the program with new relevant information.
o The R-ISI program should be re-adjusted if significant changes are performed at the plant and / or
as a response to NRC or CSN bulletins or generic letters.

Some important conclusions of RI-ISI program at Cofrentes are the following:

ASME Program
RI-ISI PROGRAM
Section XI
Scope (locations number) 2840 6458
Inspection selection 365 131

• A 65% weld inspection reduction is achieved (both class 1 and 2).

206
 NEA/CSNI/R(2009)2

• At the following table, we can see the different regions related to the combination of safety and failure
potentiality (HSS= High Safety Significance, HFI= High Failure Influence, LSS= Low Safety
Significance, LFI=Low Failure Influence). At the second column segments number within every of
the different regions are shown. At the third column, number of welds taken into account at every
region is indicated and at the fourth one, number of ASME Section XI Inspections is able to be seen.

Safety Segments Number RI-ISI Program Selected Elements ASME Section XI.
Importance Inspections Selection
HSS-HFI 42 109 115
(Region 1)
HSS-LFI 0 - -
(Region 2)
LSS-HFI 176 10 36
(Region 3)
LSS-LFI 510 12 214
(Region 4)
Total 728 131 365

• Class 1 contribution to Nuclear Damage Frequency (NDF) is 93% and Class 2 contribution is 7%:

• At the following figure, the most important degradatory phenomena contribution to the Core Damage
Frequency (CDF) could be observed: the erosion-corrosion mechanism represents 51% of the Core
Damage Frequency, IGSCC represents 23% (only 37 of the 728 segments are susceptible).

% FDN por
IGSC C
mecanismos
2 3%

E/ C
5 1%

Fa t ig a
26%

• 98% of segments susceptible to fatigue and with significant contribution to the Core Damage
Frequency are out of the actual ASME Section XI scope current requirements.
• Risk is not altered because a reduction at both FDN (-9, 84E-09) and LERF (-3, 22E-10) is achieved.
• The program meets the guidance objectives: inspection programs optimisation without losing safety.
• The most important degradatory mechanism for its failure potentiality is Erosion-Corrosion.

Finally the results obtained for the Class-1 and 2 piping and the analysis of the Costs/Benefits of the
project will be emphasised at the presentation.

207
NEA/CSNI/R(2009)2

208
 NEA/CSNI/R(2009)2

209
NEA/CSNI/R(2009)2

210
 NEA/CSNI/R(2009)2

211
NEA/CSNI/R(2009)2

212
 NEA/CSNI/R(2009)2

213
NEA/CSNI/R(2009)2

214
 NEA/CSNI/R(2009)2

215
NEA/CSNI/R(2009)2

216
 NEA/CSNI/R(2009)2

SESSION 3
OPDE DATABASE AND ITS APPLICATIONS

The OECD Piping Failure Data Exchange Project

K. Gott, SSM
Application Activities of the OPDE Database
S. Ogiya, JNES
An Evaluation of Impact of Ageing in OPDE Piping Failure data
M. D. Pandey and X. X. Yuan, Institute for Risk Research, University of Waterloo J. Riznic, CNSC

Progress with the OECD Pipe Failure Data Exchange Project

K. Gott, SKI, A. Huerta, OECD/NEA, B. Lydell, Sigma-Phase Inc.

217
NEA/CSNI/R(2009)2

218
 NEA/CSNI/R(2009)2

The OECD Piping Failure Data Exchange Project

Karen Gott
Swedish Radiation Safety Authority, Sweden

OPDE project

Structural integrity of piping systems is essential for plant safety and operability. Thus information on the
degradation and failure of piping components and systems has been collected worldwide to provide
systematic feedback to regulation and research in the areas of non-destructive testing, in-service inspection
programmes and both deterministic and probabilistic safety assessments. This work has had a national
focus and reporting has been in the native language and therefore in many cases not readily available for
the international forum.

In 1994 the Swedish Nuclear Power Inspectorate started a five year project to explore the viability of
creating an international pipe failure database. The results of this effort were presented at international
meetings and it became evident that there was sufficient interest in the work that OECD/NEA took over
formal operation of the project under the coordination of its Committee on the Safety of Nuclear
Installations. The starting point for the OPDE project was the in-kind contribution from the Swedish
Nuclear power Inspectorate of a Microsoft Access database with approximately 2,300 records for the
period 1970 to 1998. The first term of the project ran from May 2002 – 2005, the second term is almost
completed and a third term is in preparation and will run from 2008 – 2011.

The OPDE project collects data from events involving metallic piping components of the primary system,
main process and standby safety systems, and support systems (ASME Code Class 1, 2 and 3 or
equivalent). The database includes events which have led to leakage, the need for remedial action, exceed
the design code or are interpreted as structurally significant. Also included are events in non-safety related
piping which if they leak can lead to common-cause initiating events such as flooding of vital plant areas.

The OPDE project is supported by eighteen organisations from twelve member countries: Belgium,
Canada, Czech Republic, Finland, France, Germany, Japan, Korea (Republic of) Spain, Sweden,
Switzerland and United States of America. The project is controlled by the Project Review Group
comprised of the National Coordinators from the twelve member countries and with the support of a
project secretary from OECD/NEA and a clearinghouse. The group meets at least once a year to review the
progress of the project and draw up guidelines for the near future. During the first term of the project the
main effort was concentrated to database structure and verification of the data already in the database. In
addition countries for which there was minimal data in the SKI database concentrated on populating the
database with data starting with the most recent events and continuing as far as possible back in time.

Organisation and management of OPDE

At the end of the first term agreement had been reached on the detailed structure of the database and a web
interface was developed by NEA for data input. At the same time the entire database was transferred from
the clearinghouse to the secure servers at NEA headquarters. It is anticipated that in the near future the

219
NEA/CSNI/R(2009)2

database will be downloaded directly from the web and no longer be distributed on CDs as the default
option. It is proposed that in the coming term of the project the database be made more user-friendly.

Quality assurance and data quality are two central aspects to the usefulness of a database. The OPDE
project has a Quality Assurance Programme to establish the organisational and technical measures for
quality assurance and monitoring the work to ensure high quality of the end product. All activities in the
project are covered and all project participants must agree to comply. To ensure consistency in the data the
Coding Guidelines have been developed. The quality of the database reflects the reporting requirements in
the different countries and also that the requirements have changed over time. The initial event can have
been recorded in several different ways from plant internal documentation to formal reporting on the
national or international level. In principle it is conceived that in most countries data input will in not be
made by the national coordinator but by the utilities or their representatives. The national coordinator will
perform the first quality check before submitting the record to the clearing house for a final quality check.

Awareness of the importance of good documentation has in general improved over the years, so it is often
difficult to verify the information from events in the 1970s and even 1980s. This also means that it is
difficult to complete all the fields in the database for these early events. The quality of the data should not
be confused with the completeness of the information. A completeness index has been introduced in part
for database management purposes but also as an indication of the usefulness of the individual records for
different applications. The completeness index distinguishes between records for which more information
can reasonably be expected to become available and those considered to be complete.

Another central aspect of this type of project is accessibility to the database, and this has been discussed at
great length within the Project Review Group. The current status is as follows: all organisations which
contribute to the database may also have access to a password protected copy of the latest version. A
“light” version is available on request from a national coordinator for use in specific applications by
organisations contracting the work to a third party. The light version contains no information which will
enable the identification of the source outside their own country. Only a very few copies of the light
version have been made as yet, but the project intends to evaluate the usefulness and is also considering
adding more fields, such as an edited version of the event narrative.

Since the end of the first term the Program Review Group has given the host of their meetings the
possibility of inviting other members of their organisation and also utility representatives to part of the
meeting when presentations of the database and applications are given. This is part of the effort to
encourage the use of the database and has been appreciated. The clearing house can also hold seminars in
member countries to help utility representatives train in the inputting of data.

Database structure

The OPDE database is currently a relational database in Microsoft Access. Data entry is managed by input
forms, tables, roll-down menus and database relationships. Database searches and applications are
performed through queries that utilise the tables and data relationships.

220
 NEA/CSNI/R(2009)2

Figure 1: The opening screen of the OPDE database

Figure 2: Log in screen for data entry

221
NEA/CSNI/R(2009)2

Figure 3: Input forms

Figure 4: Database tables

222
 NEA/CSNI/R(2009)2

In the latest version of the OPDE database there are a total of 3,546 records distributed as follows: Belgium
8 records, Canada 101, Czech Republic 14, Finland 40, France 58, Germany 278, Japan 94, Korea
(Republic of) 49, Spain 30, Sweden 351, Switzerland 55 and United States of America 2,465.

At the first workshop held within the project at the end of 2004 in Seoul a list of potential applications for
the OPDE database was compiled. They fall into five categories (some applications can be included in
more than one category but were only listed once in the compilation). The categories are:

• Applications related to analysis of materials degradation

• Applications related to code development
• Applications related to PSA
• Applications related to risk informed regulation
• Applications of practical or operational significance

Concluding remark

The OPDE project has produced a database of piping failures and degradation which can be used for a
wide variety of applications. The webOPDE input site provides a more convenient method for submitting
data and performing quality checks. For the continued updating of the database it is important that national
routines are established for example to make it possible for utilities to report directly and in a timely
manner. The enthusiasm, constructiveness and openness of the National Coordinators has been impressive
and as we continue the project I personally hope that this will continue and that the database will be used
extensively in many countries.

223
NEA/CSNI/R(2009)2

224
 NEA/CSNI/R(2009)2

225
NEA/CSNI/R(2009)2

226
 NEA/CSNI/R(2009)2

227
NEA/CSNI/R(2009)2

228
 NEA/CSNI/R(2009)2

229
NEA/CSNI/R(2009)2

230
 NEA/CSNI/R(2009)2

231
NEA/CSNI/R(2009)2

232
 NEA/CSNI/R(2009)2

233
NEA/CSNI/R(2009)2

234
 NEA/CSNI/R(2009)2

Application Activities of the OPDE Database

Shunsuke Ogiya, Japan Nuclear Energy Safety Organisation (JNES), Japan

1. Introduction

The Japan Nuclear Energy Safety Organisation (hereinafter referred to as “JNES”) is performing the
various sorts of activities improve nuclear safety regulation, in order to ensure nuclear safety in
cooperation with the regulatory authority, the Nuclear and Industrial Safety Agency (hereinafter referred to
as “NISA”).

The application activities of the OECD database in JNES are categorized into the following four types of
application activities that are performed or planned:

1) Development of countermeasures to prevent piping failure occurrence including improvement of

operating experience feedback.

2) Qualitative evaluations of failure trends and patterns applied to aging management.

3) Risk informed applications in support of development of RI-ISI program or probabilistic safety

assessment (hereinafter referred to as “PSA”) to determine pipe failure rates and rupture
frequencies.

4) Advanced applications supporting material science research, i.e. source of data parameters for
input to probabilistic fracture mechanics (hereinafter referred to as “PFM”) codes.

2. Application to prevent piping failure occurrence including improvement of operating

experience feedback

2.1 Nuclear safety information process flow

The various activities on application of nuclear safety information are being carried out by JNES. The
nuclear safety information includes information related to incident & failure occurred at domestic and
overseas nuclear facilities, operating performances at NPPs, radiation exposure information, component
reliability, inspection information, overseas safety information and so forth. This extensive information has
been collected and processed as a database of nuclear safety information. This information has been
adequately analysed and evaluated corresponding to its application and its users, and delivered to the
public, the NISA and relevant organisations in overseas countries.

2.2 Collection of the information on incident & failure

The domestic information consists of events reported in accordance with laws and events opened to the
public by licensees voluntarily. The incident & failure information from overseas countries includes
information provided by regulatory bodies in overseas countries and relevant information from the
activities of international organisations. The former mainly consist of USNRC incident & failure

235
NEA/CSNI/R(2009)2

information, the information analysed and evaluated by USNRC for events, and incident & failure
information from Europe and Asia countries. The latter mainly consist of INES information, IRS
information, information from the NEA Information Exchange Project such as OPDE, and the information
analyzed and evaluated by NEA and IAEA activities for significant events,

2.3 Analysis and evaluation of the information on incident & failure

Four examples of activities are presented regarding analysis and evaluation of the information on incident
& failure. The first example is use for supporting to NISA in their response at occurrence of incident &
failure in Japan, such as providing information on similar events and following up of horizontal
development. The second one is to derive lessons learned from the information on incident & failure in
Japan and other countries. The review meeting on this kind of safety information is held by NISA and
JNES jointly, and necessity of reflecting the resulting lessons learned to the safety regulation is discussed
in the review meeting. The third one is to speed up the response to incident & failure by cause factor
analyses. The response is made to be quick and accurate by preparing templates of fault tree of
representative events in advance and applying them when similar event occurs. The forth one is trend
analysis, including ageing analysis for the purpose of regulatory review and inspection.

The analysis and evaluation of the information and data on the piping incident & failure also are carried
out in the same manner as mentioned above.

3. Application to Ageing management

In October 2003, the implementation of Ageing Management Evaluation is re-classified as one of the
regulatory requirements. Matters to be implemented as the Aging Management Evaluation are: a) to
analyse the impacts of technically conceivable ageing phenomena on components and structures of NPPs
with safety functions at a time within 30 years after commissioning, and to technically evaluate the
possibility for prevention of the loss of function of the components and structures due to ageing
phenomena under the present maintenance activities provided to them, b) to extract new maintenance
measures from the technical evaluation results to develop the ten-year maintenance program. c) to
re-evaluate the ten-year maintenance program with a ten-year interval and NISA decided to confirm the
status of the implementation of these items.

In August 2005, the Ageing Countermeasure Examination Committee under the Nuclear and Industrial
Safety Sub-committee issued the "Enhancement of Measures for Ageing Management at NPPs". In
response to the report, NISA prepared the “Ageing Management Implementation Guideline for NPPs” and
the “Standard Review Plan for Ageing Management Technical Evaluation of NPPs “in December 2005
and, JNES issued the "The compilation of Technical Information/Document for Ageing Management
Implementation ", which provided the standards, view points and review points when NISA and JNES
assess and review licensee's technical evaluation reports and long-term maintenance programs.

Ageing management is an important aspect of ensuring the safety and integrity of NPPs. In order to reflect
important technical information and outcomes of operating experience feedback activities on occurred events
and safety researches, effectively and efficiently, into regulations or actual ageing managements, it is important
to exchange and share the up-to-date technical findings including events data through international cooperation
such as OPDE and SCAP. "The compilation of Technical Information/Document for Ageing Management
Implementation " consists of “Ageing-related Technical Evaluation Review Manual”, “Collection of Event
Information at domestic and overseas NPPs”, and “up-to-date technical knowledge/information on ageing
management” The knowledge and information from OPDE and SCAP are feed backing to this “Compilation of
Technical Information/Document” to enhance technical basis for ageing management implementation. The

236
 NEA/CSNI/R(2009)2

“Review Manual” includes the “Manual” for Stress Corrosion Cracking (SCC), Fatigue, and Thinning of
Piping, which are related to piping ageing.

4. Application to risk informed regulation

In Japan, the Nuclear Safety Commission (hereinafter referred to as “NSC”) decided the “Basic Policies on
Introduction of Risk Informed Nuclear Safety Regulation” in November 2003. NISA, in response to the
decision, issued the "Fundamental Concepts of Risk Informed Nuclear Safety Regulations" providing the
way of thinking for RIR, and "Short-term Implementation Plans of RIR" was published by NISA/JNES in
May 2005. The studies on RIR were promoted in accordance with this Plan. The implementation plan was
revised based on the progress since then in January 2007. The “Guideline for RIR” and the “PSA Quality
Guideline”, both for trial use, were issued by NISA in January 2007.

In the “Short-Term RIR Implementation Plans”, the following studies related to piping are specified for
enhancement of inspection strategies focused on the safety significant activities; a) to evaluate safety
significance of components/piping at typical PWR & BWR based on the results of PSA, for contribution to
selecting of safety significant components/piping, b) to evaluate relevance of the component failure mode
in PSA model, corresponding to the function and degradation mode of components/piping based on
operating experience. And the “Implementation Plans” also includes the activities for data collection and
preparation of parameters for the use in PSA, such as “Piping failure rates”, “Frequency of occurrence of
initiating event”, and “Parameters associated with common cause failures” using the database like OPDE
and SCAP.

In accordance with the above-mentioned circumstance, the needs of risk informed applications or PSA
applications to determine piping failure rates and rupture frequencies and the applications supporting to
PFM codes have been increasing.

Regarding PFM code, there are several application needs as follows; (1) provisional estimation of LOCA
frequencies, (2) Seismic PSA considering ageing (evaluation of piping failure probability at earthquake
considering ageing), (3) RI-ISI evaluation, etc. JNES are performing improvement and verification
activities of PFM code.

237
NEA/CSNI/R(2009)2

238
 NEA/CSNI/R(2009)2

239
NEA/CSNI/R(2009)2

240
 NEA/CSNI/R(2009)2

241
NEA/CSNI/R(2009)2

242
 NEA/CSNI/R(2009)2

243
NEA/CSNI/R(2009)2

244
 NEA/CSNI/R(2009)2

An Evaluation of Impact of Ageing in OPDE Piping Failure Data

M. D. Pandey and X. X. Yuan, Institute for Risk Research, University of Waterloo

J. Riznic, Canadian Nuclear Safety Commission

Abstract

The accurate estimation of piping failure frequency is an important task to support the probabilistic risk
assessment and risk-informed in-service inspection of nuclear power plants. In particular, in the planning
of inspection and maintenance programs, an important question arises about the effect of plant ageing on
the pipe failure frequency. Naturally, if the failure rate were found to increasing with ageing due to
progressive deterioration, more stringent inspection and replacement programs would be required to
maintain the plant safety.

In the literature, several probabilistic models have been reported to quantify the piping failure frequency,
and they have been applied to data contained in OPDE database. In most of the previous studies, the pipe
failure rate is considered as a constant and the effect of ageing is not included in the analysis.

This paper proposes a stochastic point process model that incorporates both a time dependent trend and
plant specific (or cohort) effects on the failure rate. A rigorous likelihood based statistical method is
proposed for estimating the model parameters. The paper presents a case study to analyse PWR and BWR
Class 1 pipe failure data given in the OPDE Database. The paper shows that the consideration of both
ageing and cohort effect is important.

245
NEA/CSNI/R(2009)2

246
 NEA/CSNI/R(2009)2

247
NEA/CSNI/R(2009)2

248
 NEA/CSNI/R(2009)2

249
NEA/CSNI/R(2009)2

250
 NEA/CSNI/R(2009)2

251
NEA/CSNI/R(2009)2

252
 NEA/CSNI/R(2009)2

253
NEA/CSNI/R(2009)2

254
 NEA/CSNI/R(2009)2

255
NEA/CSNI/R(2009)2

256
 NEA/CSNI/R(2009)2

257
NEA/CSNI/R(2009)2

258
 NEA/CSNI/R(2009)2

Progress with the OECD Pipe Failure Data Exchange Project

Karen Gott, Swedish Nuclear Power Inspectorate, Sweden

Alejandro Huerta, OECD Nuclear Energy Agency, France
Bengt Lydell, Sigma-Phase, Inc., USA

Structural integrity of piping systems is important for nuclear power plant safety and operability. In
recognition of this, information on degradation and failure of piping components and systems is collected
and evaluated by regulatory agencies, international organisations (e.g., OECD/NEA and IAEA) and
industry organisations worldwide to provide systematic feedback for example to reactor regulation and
research and development programs associated with non-destructive examination (NDE) technology,
in-service inspection (ISI) programs, leak-before-break evaluations, risk-informed ISI, and probabilistic
safety assessment (PSA) applications involving passive component reliability.

Several OECD Member countries have agreed to establish the OECD-NEA Piping Failure Data Exchange
Project (OPDE) to encourage multilateral co-operation in the collection and analysis of data relating to
piping failure events in nuclear power plants. The project was formally launched in May 2002 under the
auspices of the OECD/NEA. Organisations producing or regulating more than 80% of nuclear energy
generation worldwide contribute data to OPDE. Currently twelve countries have signed the OPDE
agreement (Belgium, Canada, Czech Republic, Finland, France, Germany, Korea (Republic of), Japan,
Spain, Sweden, Switzerland and United States of America).

This paper describes the current status of the OPDE database after six years of development, and gives
some insights about the collected service experience data applicable to the four systems in the RISMET
scope of work (Condensate System, Decay Heat Removal System, Main Steam System, and Reactor
Coolant System).

Also described are insights and results from a project that addressed the validation of probabilistic fracture
mechanics (PFM) codes. In 2006 the Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory
Commission sponsored an application of data on cracking, leak and rupture events from nuclear power
plant operating experience to estimate failure frequencies for piping components that had been previously
evaluated using the PRO-LOCA and PRAISE probabilistic fracture mechanics (PFM) computer codes.1
The calculations had addressed the failure mechanisms of PWSCC, IGSCC and thermal fatigue for
materials and operating conditions that were known to have failed components. The first objective was to
benchmark the calculations against field experience. A second objective was a review of uncertainties in
the treatments of the data from observed failures and in the structural mechanics models. The OPDE
database was applied to estimate failure frequencies. Because the number of reported failure events was
small, there were also statistical uncertainties in the estimates of frequencies. Comparisons of predicted
and observed failure frequencies showed that PFM codes correctly predicted relatively high failure
probabilities for components that had experienced field failures. However, the predicted frequencies tended
to be significantly greater than those estimated from plant operating experience. A review of the PFM

1
Work supported by the Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington D.C. NRC
JCN N6019. The results of this work are documented in Technical Letter Report PNNL-16625: Probabilistic Fracture Mechanics
Evaluation of Selected Passive Components, May 2007.

259
NEA/CSNI/R(2009)2

models and inputs to the models showed that uncertainties in the calculations were sufficiently large to
explain the differences between the predicted and observed failure frequencies.

1. Introduction

Structural integrity of piping systems is important for nuclear power plant safety and operability. In
recognition of this, information on degradation and failure of piping components and systems is collected
and evaluated by regulatory agencies, international organisations (e.g., OECD/NEA and IAEA) and
industry organisations worldwide to provide systematic feedback for example to reactor regulation and
research and development programs associated with non-destructive examination (NDE) technology,
in-service inspection (ISI) programs, leak-before-break evaluations, risk-informed ISI, and probabilistic
safety assessment (PSA) applications involving passive component reliability.

Several OECD Member countries have agreed to establish the OECD-NEA Piping Failure Data Exchange
Project (OPDE) to encourage multilateral co-operation in the collection and analysis of data relating to
piping failure events in nuclear power plants.

The project was formally launched in May 2002 under the auspices of the OECD/NEA. Organisations
producing or regulating more than 80% of nuclear energy generation worldwide contribute data to OPDE.
Currently twelve countries have signed the OPDE agreement (Belgium, Canada, Czech Republic, Finland,
France, Germany, Korea (Republic of), Japan, Spain, Sweden, Switzerland and United States of America).

This paper describes the current status of the OPDE database after 7 years of development, and gives some
insights based on the 3,800 piping events in the database.

Reviews of service experience with safety-related and non safety-related piping systems have been
ongoing ever since the first commercial nuclear power plants came on line in the 1960’s. For example, in
1975 the U.S. Nuclear Regulatory Commission established a first Pipe Crack Study Group (PCSG)
charged with the task of evaluating the significance of stress corrosion cracking in boiling water reactors
(BWRs) [1] and pressurised water reactors (PWRs) [2]. Service experience review was a key aspect of the
work by the PCSG. As another example, major condensate and feedwater piping failures (e.g., the U.S.
plants Trojan in 1985 and Surry-2 in 1986) due to flow accelerated corrosion (FAC) resulted in similar
national [3,4] and international initiatives [5,6]to learn from service experience and to develop mitigation
strategies to prevent recurrence of such pipe failures. Furthermore, early indications of the significance of
thermal fatigue phenomena evolved in the 1970s, and, again, systematic reviews of the service experience
enabled the introduction of improved piping design solutions, NDE methods, and operating practices [7].
Dr. S.H. Bush made important contributions to the work by the PCSG, and he also published numerous
journal articles and technical reports in the field of piping service experience and reliability spanning
almost four decades, for example [8-13].

In parallel with these focused efforts to evaluate service experience data and to correlate the occurrence of
material degradation with piping design, operational parameters and non-destructive examination (NDE)
techniques, initiatives have been presented to establish an international forum for the systematic collection
and exchange of service experience data on piping. An obstacle to the use of the database by other
countries with national qualitative and quantitative pipe failure information is that criteria and
interpretations applied in the collection and analysis of events and data differ among the various countries.
A further impediment is that the descriptions of reported events and their root causes and underlying
contributing factors, which are important to the assessment of the events, are usually written in the native
language of the countries where the events were observed.

260
 NEA/CSNI/R(2009)2

To overcome these obstacles, the preparation for the OECD Pipe Failure Data Exchange (OPDE) Project was
initiated in 1994 by the Swedish Nuclear Power Inspectorate (SKI). In 1994 SKI launched a 5-year R&D
project to explore the viability of creating an international pipe failure database [14,15]. During this period SKI
hosted meetings to present results of the R&D and to discuss the principles of database development and
maintenance. In September 1996 SKI organised the “Initial Meeting of the International Cooperative Group on
Piping Performance” with participants from thirteen countries. Again, in September 1997 SKI organised the
“International Seminar on Piping Reliability” [16]; this time with participants from eleven countries. As a
follow-up to these meetings, OECD/NEA in 2000 and 2001 organized exploratory meetings to evaluate and
decide on the prospects for an international cooperative effort. The Committee on the Safety of Nuclear
Installations (CSNI) endorsed the establishment of such an effort. Since May 2002, the OECD/NEA has
formally operated the project under the coordination of the CSNI. The first term of the Project covered years
2002-2005 and the second term covers the period 2005-2008. A third term (2008-2011) has been approved.

2. OPDE objective & scope

OPDE has established an international database on pipe degradation and failures in commercial nuclear
power plants in OECD Member Countries. The database supports the following activities:

• Trend analysis, including ageing analysis.

• Statistical analyses to determine pipe failure rates and rupture frequencies for use in risk-informed
activities (e.g., loss-of-coolant-accident frequency assessment, internal flooding initiating frequency
assessment, high-energy-line-break frequency assessment, RI-ISI ∆-risk assessment).
• Source of data parameters for input to probabilistic fracture mechanics codes.
• Degradation mechanism analysis (DMA) in RI-ISI program development.
• Development of defenses against recurring (e.g., systematic) pipe failures.
• Exchange of service data in order to pinpoint potential generic implications of a specific, significant
pipe failure.

OPDE addresses typical metallic piping components of the primary system, main process and standby
safety systems, and support systems (i.e., ASME Code Class 1, 2 and 3, or equivalent, piping). It also
covers non safety-related piping, which if leaking could lead to common-cause initiating events such as
flooding of vital plant areas. For example, raw water systems such as non-essential service water,
circulating water or fire protection system could be significant flood sources given a pipe break. In other
words, the OPDE database covers degradation and failure of high-energy and moderate-energy piping as
well as safety-related and non safety-related piping.

Included in the database are events that require remedial action (e.g., replacement, weld overlay repair)
with or without reactor shutdown to cold shutdown condition. The types of degradation or failure include
service induced inside diameter pipe wall thinning and non through-wall cracking as well as pressure
boundary breaches such as pinhole leaks, leaks, severance and major structural failures (pipe ‘breaks’ or
‘ruptures’). For pipe flaws that do not penetrate the pipe wall or weld/weld heat affected zone the OPDE
work scope encompasses degradation exceeding design code allowable for wall thickness or crack depth as
well as such degradation that could have generic implications regarding the reliability of NDE/ISI
techniques. In summary, the following types of degradation and failures are included:

• Non through-wall defects (e.g., cracks, wall thinning) interpreted as structurally significant and/or
exceeding design code allowable.

261
NEA/CSNI/R(2009)2

• Through-wall defects without active leakage; leakage may be detected following a plant operational
mode change involving depressurization and cool-down, or as part of preparations for NDE.
• Small leaks (e.g., pinhole leaks, drop leakage) resulting in temporary or permanent repair.
• Leaks (e.g., leak rate within Technical Specification limits).
• Large leaks (e.g., through-wall flow rates in excess of Technical Specification limits).
• Severance (pressure boundary failure attributed to external impact or vibration fatigue).
• Rupture (major structural failure).

In May 2002 the starting point for the Project was an in-kind contribution by SKI in the form of an
international pipe failure database in Microsoft® Access. This database included pipe failure data for the
period 1970 to 1998, and it contained approximately 2,300 records. During the first term of the project the
emphasis was on validating the content of the SKI in-kind contribution, improving and streamlining the
database structure and data input format, and populating the database with new failure data for the period
1999 to the present, as well as with pre-1998 records. Signatory countries can use the database content to
generate their own qualitative and quantitative piping reliability insights. In the future an international
co-operation for quantification of piping reliability parameters may be established separately should the
participating organisations wish to do so.

3. Project organisation

The OPDE Project Review Group (PRG) controls the project with support from the OECD/NEA project
secretariat and a Clearinghouse. Each participating country is represented by a National Coordinator. The
OECD/NEA is responsible for administering the project according to OECD rules. To assure consistency
of the data contributed by the National Coordinators the Project operates through the Clearinghouse. The
Clearinghouse verifies whether the information provided by the National Coordinators complies with the
OPDE Coding Guidelines (OPDE-CG). To encourage data exchange the Project has adopted the
IAEA/NEA Incident Reporting System (IRS) procedures [17] for protecting confidential information.

The PRG runs the Project and meets at least once per year. The PRG responsibilities include but are not
limited to the following types of decisions:

• Secure the financial and technical resources necessary to carry out the Project.
• Nominate the OPDE Project chairperson.
• Define the information flow (public information and confidential information).
• Approve the admittance of new members.
• Nominate project task leaders (lead countries) and key persons for the PRG tasks.
• Define the priority of the task activities.
• Monitor the progress of the Project and task activities.
• Monitor the work of the Clearinghouse and quality assurance.

OPDE 1st Term (2002-2005) participating countries and organisations are listed below; for the 2nd Term
(2005-2008) participation remains intact:

• Belgium (Electrabel and Tractebel).

262
 NEA/CSNI/R(2009)2

• Canada (Canadian Nuclear Safety Commission).

• Czech Republic (Nuclear Research Institute, REZ).
• Finland (Radiation and Nuclear Safety Authority, TVO and Fortum).
• France (Institut de Radioprotection et de Sûreté Nucléaire and Electricté de France).
• Germany (Gesellschaft für Anlagen- und Reaktorsicherheit).
• Japan (Japan Nuclear Energy Safety Organisation).
• Korea (Republic of) (Korea Institute of Nuclear Safety and Korea Atomic Energy Research Institute).
• Spain (Consejo de Seguridad Nuclear).
• Sweden (Swedish Nuclear Power Inspectorate).
• Switzerland (Swiss Federal Nuclear Safety Inspectorate).
• United States of America (Nuclear Regulatory Commission and Electric Power Research Institute).

4. Quality assurance & data quality

The OPDE Quality Assurance Program (OPDE-QAP) establishes the organisational and technical
principles and measures for quality assurance and monitoring of the work during operation of the OPDE
Project to ensure high quality of the end product (the database with companion reports). The QAP applies
to all activities in the project and is to be followed by all project participants.

To achieve the objectives established for the OPDE database, a Coding Format has been developed. This
Coding Format is reflected in the Coding Guidelines. The Coding Guidelines builds on established pipe
failure data analysis practices and routines that acknowledge the unique aspects of passive component
reliability in heavy and light water reactor operating environment (e.g., influences by material and water
chemistry).

For an event to be considered for inclusion in the OPDE database it must undergo an initial screening for
eligibility. The objective of this initial screening is to go beyond the abstracts of event reports to ensure
that only pipe degradation and failures according to the work scope definition are included in the database.

Data quality is affected from the moment the service data is recorded at a nuclear power plant, interpreted,
and finally entered into a database system. The service data is recorded in different types of information
systems ranging from work order systems, via ISI databases and outage summary reports, to licensee event
reports or reportable occurrence reports. Consequently the details of a degradation event or failure tend to
be documented to various levels of technical detail in these different information systems. Building an
OPDE database event record containing the full event history often entails extracting information from
multiple sources. For example, a licensee event report (or equivalent) would include a narrative description
of the cause and consequence of a degraded or failed component, and a metallographic evaluation report
would include details on crack morphology, crack size and orientation.

The term “data quality” is an attribute of the processes that have been implemented to ensure that any
given database record (including all of its constituent elements, or database fields) can be traced to the
source information. The term also encompasses “fitness-for-use”, that is, the database records should
contain sufficient technical detail to support database applications.

263
NEA/CSNI/R(2009)2

5. Application of OPDE failure data to validate PFM results

Technical Letter Report PNNL-16625 documents an application of service experience data on cracking,
leak and rupture events from nuclear power plant operating experience to estimate failure frequencies for
piping components that had been previously evaluated using the PRO-LOCA and PRAISE probabilistic
fracture mechanics (PFM) computer codes. The calculations had addressed the failure mechanisms of
stress corrosion cracking, intergranular stress corrosion cracking and fatigue for materials and operating
conditions that were known to have failed components. The first objective was to benchmark the
calculations against field experience. A second objective was a review of uncertainties in the treatments of
the data from observed failures and in the structural mechanics models. An extended version of the OPDE
database was applied to estimate failure frequencies. Because the number of reported failure events was
small, there were also statistical uncertainties in the estimates of frequencies. Comparisons of predicted
and observed failure frequencies showed that PFM codes correctly predicted relatively high failure
probabilities for components that had experienced field failures. However, the predicted frequencies tended
to be significantly greater than those estimated from plant operating experience. A review of the PFM
models and inputs to the models showed that uncertainties in the calculations were sufficiently large to
explain the differences between the predicted and observed failure frequencies.

Comparisons with field experience showed that both PFM codes are capable of predicting high failure
probabilities for the components for which operating conditions are known to have produced field failures.
Calculated probabilities have been conservative and in many cases are significantly higher than the
probabilities estimated from field experience. Evidently uncertainties and modelling assumptions have
conservatively impacted the calculated failure probabilities. A number of factors have been identified that
contribute to the conservative nature of the predictions. Additional evaluations are needed to identify the
specific elements of the PFM model and/or inputs that should be refined to achieve better agreement
between the predictions and plant operating experience.

264
 NEA/CSNI/R(2009)2

6. References

[1] U.S. Nuclear Regulatory Commission, 1975, Investigation and Evaluation of Cracking in
Austenitic Stainless Steel Piping of Boiling Water Reactor Plants, NUREG-75/067, Washington
(DC).
[2] U.S. Nuclear Regulatory Commission, 1980, Investigation and Evaluation of Cracking Incidents
in Piping in Pressurized Water Reactors, NUREG-0691.
[3] Brown, E.J., 1984, Erosion in Nuclear Power Plants, AEOD/E4-16, Office for Analysis and
Evaluation of Operational Data, U.S. Nuclear Regulatory Commission, Washington (DC).
[4] Cragnolino, G., Czajkowski, C., and Shack, W.J., 1988, Review of Erosion-Corrosion in
Single-Phase Flows, NUREG/CR-5156, U.S. Nuclear Regulatory Commission, Washington
(DC).
[5] IAEA, 1990, Corrosion and Erosion Aspects in Pressre Boundary Components of Light Water
Reactors, Proc. Specialist Meeting Organized by the IAEA, IWG-RRPC-88-1, Vienna (Austria).
[6] OECD Nuclear Energy Agency, 1995, Specialist Meeting on Erosion and Corrosion of Nuclear
Power Plant Materials, NEA/CSNI/R(94)26, Issy-les-Moulineaux (France).
[7] OECD Nuclear Energy Agency, 1998, Experience with Thermal Fatigue in LWR Piping Caused
by Mixing and Stratification, Specialist Meeting Proceedings, NEA/CSNI/R(98)8,
Issy-les-Moulineaux (France).
[8] Bush, S.H., 1976, “Reliability of Piping in Light-Water Reactors,” Nuclear Safety, 17:568-579.
[9] Bush, S.H., 1985, “Statistics of Pressure Vessel and Piping Failures,” Pressure Vessel and Piping
Technology 1985: A Decade of Progress, American Society of Mechanical Engineers, New York
(NY), pp 875-893.
[10] Bush, S.H., 1992, “Failure Mechanisms in Nuclear Power Plant Piping Systems,” Journal of
Pressure Vessel Technology, 114:389-395.
[11] Bush, S.H. et al, 1996, Piping Failures in United States Nuclear Power Plants: 1961-1995, SKI
Report 96:20, Swedish Nuclear Power Inspectorate, Stockholm (Sweden).
[12] Bush, S.H. and Chockie, A.D., 1996, Failure Frequencies and Probabilities Applicable to BWR
and PWR Piping, SKI Report 96:39, Swedish Nuclear Power Inspectorate, Stockholm (Sweden).
[13] Bush, S.H. and Chockie, A.D., 1998, Nuclear Reactor Piping Failures at U.S. Commercial
LWRs: 1961-1997, TR-110102 (Licensed Material), Electric Power Research Institute, Palo Alto
(CA).
[14] Nyman, R. et al, 1995, Reliability of Piping System Components, SKI Report 95:58, Swedish
Nuclear Power Inspectorate, Stockholm (Sweden).
[15] Nyman, R. et al, 1997. Reliability of Piping System Components. Framework for Estimating
Failure Parameters from Service Data, SKI Report 97:26, Swedish Nuclear Power Inspectorate,
Stockholm (Sweden).
[16] Nyman, R., 1997, Proc. of the International Piping Reliability Seminar, SKI Report 97:32,
Swedish Nuclear Power Inspectorate, Stockholm (Sweden).
[17] IAEA/NEA-IRS, 1998. Reporting Guidelines. Feedback from Safety Related Operating
Experience for Nuclear Power Plants, International Atomic Energy Agency, Vienna (Austria).
[18] Simonen, F.A. et al, 2007, Probabilistic Fracture Mechanics Evaluation of Selected Passive
Component, PNNL-16625, Pacific Northwest National Laboratory, Richland, (WA).

265
NEA/CSNI/R(2009)2

266
 NEA/CSNI/R(2009)2

267
NEA/CSNI/R(2009)2

268
 NEA/CSNI/R(2009)2

269
NEA/CSNI/R(2009)2

270
 NEA/CSNI/R(2009)2

271
NEA/CSNI/R(2009)2

272
 NEA/CSNI/R(2009)2

SESSION 4
PROBABILISTIC FRACTURE MECHANICS

Development of Probabilistic Fracture Mechanics Analysis Code for Aged Piping under Stress
Corrosion Cracking
K. Onizawa and H. Itoh, JAEA
Reliability Parameter Handbook for Piping Components in Nordic Nuclear Power Plants
Bengt Lydell, Sigma-Phase, Inc., Anders Olsson Relcon Scandpower AB
Probabilistic Evaluation in Support of Risk-Informed Pressure Tube Maintenance
E. Nadeau, AEC, T. Byrne, DRT Technical Services, S. Khajehpour, AEC, D. Leemans, Crosscurrents
Research and Policy Consulting

A Structural Reliability Model for the Evaluation of Failure Probabilities of Alloy 182/82 Welds in
PWRs
C. Cueto-Felgueros, Tecnatom
Activity of the PFM Subcommittee of the Japan Welding Engineering Society on Probabilistic
Fracture Mechanics
N. Maeda, JNES, S. Yoshimura, University of Tokyo, Y. Kanto, University of Ibaraki
NURBIM Fatigue Benchmarking of SRRA Code for Risk-Informed ISI
E. Milanov, Westinghouse
CANTIA- Probabilistic Code for CANDU Tube Inspection Assessment
S. T. Revankar and B. Wolf, Purdue University, and J. Riznic, CNSC
Probabilistic Modelling of Primary Water Stress Corrosion Cracking (PWSCC)
In Piping Dissimilar Metal Butt Welds
B. A. Bishop, Westinghouse
Risk-Informed Leak-Before-Break Reconciliation with Modern Codes and Standards for Plant Life
Extension
M.J. Kozluk and X. Duan, AECL

273
NEA/CSNI/R(2009)2

274
 NEA/CSNI/R(2009)2

Development of Probabilistic Fracture Mechanics Analysis Code

for Aged Piping Under Stress Corrosion Cracking

Kunio Onizawa and Hiroto Itoh, Nuclear Safety Research Centre, Japan Atomic Energy Agency, Japan

Abstract

Stress corrosion cracking (SCC), which affects the structural integrity of reactor component, has been observed
at some piping joints and core shrouds made by Austenitic stainless steel in BWR plants. It is well known that
the SCC behaviour is significantly scattered depending upon the various conditions; material, piping geometry,
welding technique, environmental condition, and so on. Since probabilistic fracture mechanics (PFM) analysis
method treats such scatter and uncertainties in the structural integrity evaluation, it is, therefore, useful to apply
the PFM analysis to the evaluation of the piping integrity considering the SCC behaviour. In JAEA, we have
been developing the PFM analysis code for aged piping based on the latest knowledge on age ing degradation
and analytical methods. The PFM analysis code, called as PASCAL-SP, evaluates the probabilities of failure
such as leakage and rupture of piping at aged welded joints based on the Monte Carlo method. Figure 1 shows
the flow chart of the PASCAL-SP code. The analysis code conforms to regulations of Nuclear and Industrial
Safety Agency (NISA) and the Codes for Nuclear Power Generation Facilities – Rules on Fitness-for-Service
for Nuclear Power Plants – of the Japan Society of Mechanical Engineers (JSME). Although many factors have
some influence on failure probabilities in the PFM analysis, we are mainly focusing on the effects of in-service
inspection and welding residual stress distribution. Accuracy of flaw detection and sizing error for in-service
inspection are modelled by the data from national research projects such as Ultrasonic Test & Evaluation for
Maintenance Standards (UTS). Residual stress distribution, which largely affects SCC behaviour, is determined
by parametric FEM analyses based on welding experiments. In this paper, the effects of inspection intervals and
degrees of in-service inspection on failure probability are presented by the proposed PFM analysis code. The
effect of uncertainty of welding residual stress on failure probability is also discussed.

Start
Analytical Condition Random Variables
・Environment, Geometry Analytical Condition Setting ・Residual Stress
・Operating Stress, Residual ・Crack Growth Rate
Stress, etc Sampling of Random ・Accuracy of Flaw
Variables Detection and Sizing, etc

Simulation of Plant Operation

Monte Carlo Method

Failure
Load of Evaluation of Crack Growth Judgment
Operating Stress, ・Leak
Residual Stress ・Break
Operation Time
Occurrences of Event
・In-Service Inspection（ISI）
・Earthquake
・Transient event

Evaluation of Failure Probabilities

Stop

Figure 1: Flow chart and schematic image of PFM analysis code for piping: PASCAL-SP

275
NEA/CSNI/R(2009)2

276
 NEA/CSNI/R(2009)2

277
NEA/CSNI/R(2009)2

278
 NEA/CSNI/R(2009)2

279
NEA/CSNI/R(2009)2

280
 NEA/CSNI/R(2009)2

281
NEA/CSNI/R(2009)2

282
 NEA/CSNI/R(2009)2

283
NEA/CSNI/R(2009)2

284
 NEA/CSNI/R(2009)2

285
NEA/CSNI/R(2009)2

286
 NEA/CSNI/R(2009)2

Reliability Parameter Handbook for Piping Components in

Nordic Nuclear Power Plants

Bengt Lydell, Scandpower Risk Management, Inc., USA

Anders Olsson, Relcon Scandpower AB

The Nordic PSA Group (NPSAG)1 has undertaken to develop a piping reliability parameter data handbook
for use in risk-informed applications that involve the consideration of structural integrity of piping systems.
The scope of the handbook is to establish high quality reliability parameters that account for the Nordic
and worldwide service experience with safety-related and non-safety-related piping systems in a consistent
and realistic manner. The scope of the Handbook covers all major safety-related and non-safety related
piping systems in commercial BWR and PWR plants. Included in the handbook will be recommended
piping reliability parameters, including sets of pipe failure rates and conditional failure probabilities
organised by plant type, system and component boundary (e.g., base metal, expander/reducer, tee, and
weld-HAZ). Also included in the Handbook will be guidelines for data specialisation using Bayesian
methodology.

While the work to develop the handbook is expected to be finalized during 2009, the planning for its
preparation has been underway for well over ten years. An important step towards the handbook
development project has been the international cooperative effort through the OECD Nuclear Energy
Agency to create an event database on the service experience with piping in commercial nuclear power
plants; an event database, which provides the necessary input to the work with the NPSAG handbook.
Practical applications involving the statistical estimation of pipe through-wall leak and rupture frequencies
have been pursued in parallel with the database development. Experiences from these practical applications
form the basis for the NPSAG project to establish a piping reliability parameter handbook for use by the
practicing PSA engineer.

The focus of this technical paper is on the lessons learned from the practical piping reliability analyses and
how this body of experience is utilised in the handbook development by the NPSAG. In the paper, special
consideration is given uncertainty analysis, development of justifications for chosen prior distributions,
treatment of influences of leak detection and in-service inspection strategies, and treatment of influences of
different degradation mitigation initiatives. The technical requirements for the NPSAG project are
documented in SKI Report 2008:01 (January 2008, available on the Internet at www.ski.se).

1
Formed in December 2000, NPSAG consists of representatives from all Nordic nuclear utilities and Studsvik Nuclear. The
NPSAG industry forum monitors developments in Probabilistic Safety Assessment, shares insights and results from practical
applications and promotes R&D.

287
NEA/CSNI/R(2009)2

1. Introduction

This paper describes the progress of an ongoing Nordic R&D project to develop a piping reliability
parameter handbook (the “R-Book” project). The paper also summarizes the results and insights from a
pilot project to define the content and outline of the proposed handbook. Comments and recommendations
for the R-Book development process were solicited from Nordic and international experts. Detailed
information about technical considerations for how to derive realistic pipe failure rates from the available
service experience data is documented in SKI Report 2008:01 [1]. Work is currently underway to produce
the R-Book and a first edition is scheduled for release to project sponsors during the second half of 2009.
During 2008 interim results including demonstration of analysis tools will be presented to Nordic industry
representatives.

2. Objective and background

The objective of the project is to utilise the OECD Nuclear Energy Agency “OECD Pipe Failure Data
Exchange Project” (OPDE) database to derive piping component failure rates and rupture probabilities for
input to internal flooding probabilistic safety assessment, high-energy line break” (HELB) analysis,
risk-informed in-service inspection (RI-ISI) program development, and other activities related to PSA.
This new R&D project is funded by member organisations of the NPSAG (Forsmark AB, OKG AB,
Ringhals AB, and the Swedish Nuclear Power Inspectorate).

The history behind the current effort to produce a handbook of piping reliability parameters goes back to
1994 when SKI funded a 5-year R&D project to explore the viability of establishing an international
database on the service experience with piping system components in commercial nuclear power plants.
An underlying objective behind this 5-year program was to investigate the different options and
possibilities for deriving pipe failure rates and rupture probabilities directly from service experience data as
an alternative to probabilistic fracture mechanics. The R&D project culminated in an international piping
reliability seminar held in the fall of 1997 in Sigtuna (Sweden) [2] and a pilot project to demonstrate an
application of the pipe failure database to the estimation of loss-of-coolant-accident (LOCA) frequency
(SKI Report 98:30 [3]).

A particularly important outcome of the 5-year project was a decision by SKI to transfer the pipe failure
database including the lessons learned to an international cooperative effort under the auspices of the
OECD Nuclear Energy Agency. Following on information exchange and planning meetings that were
organized by the OECD Nuclear Energy Agency during 2000 – 2001, the “OECD Pipe Failure Data
Exchange Project” (OPDE) was officially launched in May 2002. Today (June 2008), the OPDE is
supported by organisations from twelve countries.

Since the completion of the original piping reliability R&D project in 1998, a very large number of
practical pipe failure database applications have been completed, some of which are referenced in [1]. The
insights and lessons learned from these practical applications, including the experience gained from the
OPDE project, form the basis for developing the “R-Book.” An important observation from prior
applications is the need to ensure that reports on pipe degradation and failure as recorded in a database are
fully validated and that the event populations that result from database queries are sufficiently complete.
The overall strategy for work currently underway is to:

• Demonstrate analysis methods and techniques by selecting two or three plant systems and to
generate qualitative and quantitative reliability information about pipe damage and degradation
susceptibilities. The objective of this initial analysis task is to develop a common understanding of
how to best implement the chosen technical approach.

288
 NEA/CSNI/R(2009)2

• Organise a seminar to allow participating NPSAG member organisations to comment on the

available draft material.
• Implement any recommendations for enhancements and to proceed with the statistical estimation
process for remaining plant systems in the work scope.

3. Technical requirements for the R-Book

The process of defining the technical requirements for the R-Book have benefited from recommendations
made by the intended users of the Handbook. During 2007 a questionnaire was developed and distributed
to several organisations with a request for comments. This questionnaire included a large number of
questions pertaining to specific analytical issues and structure and content of the Handbook.

4. Applicability and level of detail

The main purpose of the R-Book is to provide data for PSA. The derived failure parameters will reflect the
degradation and failure characteristics that are unique to passive components such as piping. Therefore, the
failure parameters will be representative of observed transitions from a degraded state (e.g., wall thinning,
non-through-wall crack) to a failed state – that is a through-wall flaw producing a liquid or steam phase
leak or flow rate of certain magnitude. The Handbook will provide tabulations of failure rates for an initial
defect for different system organised by pipe size and type of material.

For each initial defect the conditional probability for a leakage will be calculated and by this a frequency
for different levels of leakages can be presented. Conditional pipe failure probabilities will be developed
for the uniquely defined consequences of structural failure using a technical approach as documented in
PVP2007-26281 [4]. For different types of piping different levels of leakage will be presented.

In order to make it possible to correlate a leak rate to a corresponding pipe break diameter the data tables in
the R-Book will also contain a column with this information. Frequency of “structural failure” will be
estimated on the basis of the resulting through-wall flow rate (kg/s).

With respect to leak rates, it is important to note that the database OPDE in itself do not contain any
explicit information whether a certain leak did exceed limits given by the Technical Specification or not.
This kind of information is however used in the leak rate definition as explained in the Coding Guideline
of OPDE [5].

In order to have a sufficient statistically significant data material, the parameter estimation process will
pool the Nordic and international service experience data in an “appropriate” manner. Data specialisations
will be performed as needed using Bayesian methodology. The raw data presentations will be carefully
crafted to convey to users of the Handbook what the reasons are for any variability in observed pipe
degradation and failure. These raw data presentations will show the Nordic service experience versus the
international service experience. Where variability in counts of degradation and failure is noticeable can
usually be explained by different reporting criteria as well as different design, fabrication and inspection
practices. While the Handbook will give “recommended” pipe failure rates and rupture frequencies, it will
also include enough qualitative and quantitative information about damage and degradation susceptibilities
to allow a user to pursue further data specializations using the R-Book methodology. Not only will the
R-Book include information on “failure counts”, it will also provide exposure term data that account for
weld populations and pipe lengths for the different systems in the work scope. Piping system design
variability will be accounted for by presenting Median, Lower Bound and Upper Bound estimates of
component counts as derived from piping system design information (i.e., isometric drawings).

289
NEA/CSNI/R(2009)2

Failure data in the R-Book will be presented for different types of piping components and material (carbon
steel, stainless steel, low alloy steel, nickel base material), according to the information available in the
OPDE database. The level of detail with respect to this is expected to be as follows:

• Welds
• Base metal
• T-joints
• Bends/elbows
• Expanders/reducers

If a more detailed differentiation is needed it will be up to each user to proceed with this.

5. Traceability

Data will be extracted from the OPDE database using queries in MS Access. Each query used will be given
a unique ID and be saved, and retained for future reference as appendices in the Handbook. Information
about the queries and version of the OPDE database used will be sufficient in order to reproduce the input
data. If needed, the queries can be expanded in order to also list the individual failure reports in OPDE that
was the result of each query.

Recorded in OPDE is any degraded condition that requires some kind of corrective measure to be taken
(repair or replacement). The database includes “precursor events” (non-through-wall flaws) as well a
through-wall flaws that generate active leakage. OPDE is continually growing with approximately 200
events per year. A new version of OPDE is released every six month.

6. Reliability parameters

The methodology to be used for the R-Book is described in detail in Chapter 2 in Reference [1]. This
methodology has been subjected to independent reviews by the Los Alamos National Laboratory (LANL),
the University of Maryland (UoM), and Korea Energy Research Institute (KAERI). Since its original
implementation in 1998 [7], the methodology has been extensively applied and enhanced to address a
broader class of piping reliability cases. It is implemented in Microsoft® Excel with Crystal Ball® for
uncertainty propagation. An advantage of this implementation is that all calculations will be traceable.
Parameters that will be presented in the first issue of the R-Book are listed in Table 1.

Pipe failure rate estimation involves querying a database for event populations (number of failures) and
corresponding exposure terms or component populations (number of components from which the failure
data are collected). Beyond these basic sets of information and depending on the specific type of
risk-informed application, additional supporting and specialised information on pipe failure is needed.
Database development must go hand-in-hand with practical applications to ensure that structure and
content is sufficiently complete and compatible with the needs of analysts.

290
 NEA/CSNI/R(2009)2

Table 1: Piping Reliability Parameters

Symbol Description
λik Frequency for an initial defect (calculated)
Conditional probability for a leak consequence given the initial
Pik
defect (calculated)
nik Number of events (result from query)
Portion of the total piping component population in a system that is
fik susceptible to certain degradation or damage mechanism (based on
OPDE and RI-ISI Degradation Mechanism Assessments)
Ni Number of piping components in population (results from query)
Exposure time, based on number of reactor years (from plant
Ti
population database)

Also included in Reference [1] is the result of a survey of pipe failure databases. It provides insights about
database structures, database content and the importance of data validation. Can the results of applications
of existing databases be trusted?

Numerous pipe failure databases have been developed to support risk-informed applications. Beyond
fulfilling a one-time objective, most databases have not been subjected to continuous or periodic updates,
however. A lack of validation of data records influences the validity of derived reliability parameters; this
topic is addressed further in Appendix A in Reference [1].

7. Plant systems

The proposed scope of the R-Book is given below in terms of plant systems for which pipe failure data
parameters will be derived. Table 2 presents the proposed work scope, which reflects intended
risk-informed PSA applications. The systems that are listed in Table 2 cover the full range of risk-informed
PSA applications (LOCA frequency estimation, HELB evaluations, internal flood PSA, RI-ISI).

Table 2: Plant Systems Considered in R-Book Project

OPDE Nordic
Description
Generic Designations
ADS BWR Primary Depressurization System (BWR) 314
AFW Auxiliary Feedwater System 327
CC Component Cooling Water System 711/712
COND Condensate System 414/430 (1)
CRD Control Rod Drive (Insert/Removal/Crud Removal) 354
CS Containment Spray System 322
CVC Chemical & Volume Control System (PWR) 334
CW Circulating Water System 443
EXT Steam Extraction System 419/423
FPS Fire Protection System 762
FW Main Feedwater System 312/415 (2)
HPCS High Pressure Core Spray (BWR) --
HPSI High Pressure Safety Injection (PWR) --
LPCS Low Pressure Core Spray (BWR) 323
LPSI Low Pressure Safety Injection (PWR) 321 (LPSI)

291
NEA/CSNI/R(2009)2

Table 2 Cont’d: Plant Systems Considered in R-Book Project

MS Main Steam System 311/411 (3)

MSR Moisture Separator Reheater System 422
RCS Reactor Coolant System (PWR) 313
RHR Residual Heat Removal System 321
RR Reactor Recirculation System (BWR) 313
RPV-HC RPV Head Cooling System (BWR) 326
RVLIS Reactor Vessel Level Indication System (BWR) 536
RWCU Reactor Water Cleanup System (BWR) 331
SFC Spent Fuel Pool Cooling System 324
S/G Blowdown Steam Generator Blowdown System (PWR) 337
SLC Standby Liquid Control System (BWR) 351
SW Service Water System 712/715
Notes:
1. 414 for F1/F2/R1/R2/R3/R4 and 430 for O1/O2/O3
2. 312 for O1/O2/O3 and 415 for F1/F2/R1/R2/R3/R4. Also note that 312 is the
designation for steam generators in Ringhals-2/3/4
3. 311 for O1/O2/O3 411 for F1/F2/R1/R2/R3/R4

For each plant system that is addressed by the R-Book relevant qualitative information on the service
experience will be presented. The qualitative information will be organised according to a template as
given by Table 3.

Table 3: Scope of R-Book

Plant System – e.g., BWR 313 Event History (Failure Count)

Degradation Mechanism (DM#) 1970-1979 1980-1989 1990-1999 2000-2007
Worldwide
DM1
Nordic
Worldwide
DM2
Nordic
Worldwide
DM3
Nordic
Worldwide
DM4
Nordic
Notes:
a – Mitigation program
b – Water chemistry
c – Material (e.g., typical types, material compositions)
d – Ageing effects (including effects of power uprate projects)
e – Non-destructive examination (NDE) practices

In a set of notes (“a” through “e” in Table 3) addresses key piping reliability influence factors. These notes
provide additional information on the conditions that are judged to be of significance with respect to the
number of observed defects. With this information as background the user of the R-Book can form
conclusions about different service conditions, including inspection and mitigation, and their observed
effects on the number of defects that are recorded in OPDE. These conditions might for instance be ageing
effects, effects of change of material, but also change in non-destructive examination (NDE) methods.

292
 NEA/CSNI/R(2009)2

The influence factors on piping performance are interrelated. For example, a power uprate may cause
increased wear effects on secondary system piping. But mitigation programs (e.g., replacement of original
carbon steel piping with piping of low alloy steel) and improved NDE could offset a projected (or
assumed) increase in observed failure rate.

8. Conclusions

The usefulness of any component failure data collection depends on the way by which a stated purpose is
translated into database design specifications and requirements for data input and validation, access rules,
support and maintenance, and QA. Developing the R-Book is an important step to verify the content and
quality of the OPDE database. Also it is important that interested parties strive against harmonised ways of
creating reliability data to be used in safety analyses.

9. References

[1] Anders Olsson and Bengt Lydell. “Reliability Data for Piping Components in Nordic Nuclear Power
Plants “R-Book” Project Phase I”, SKI Report 2008:01 (January 2008).
[2] Ralph Nyman (Editor), Seminar on Piping Reliability, SKI Report 97:32 (October 1997).
[3] Bengt Lydell, “Failure Rates in Barsebäck-1Reactor Coolant Pressure Boundary Piping, An
Application of a Piping Failure Database”, SKI Report 98:30 (May 1999).
[4] Bengt Lydell, The Probability of Pipe Failure on the Basis of Operating Experience, ASME
PVP2007-26281 (July 2007).
[5] OECD Nuclear Energy Agency, OPDE 2006:1 Coding Guideline (OPDE-CG), OPDE PR01,
Version 1.0, Issy-les-Moulineaux (France), 2006.
[6] Bengt Lydell, Alejandro Huerta and Karen Gott, Progress with the International pipe Failure Data
Exchange Project, ASME PVP-26278 (July 2007).
[7] Karl Fleming et al, Piping System Reliability and Failure Rate Estimation Models for Use in Risk
Informed In-Service Inspection Applications, TR-110161, Electric Power Research Institute, Palo
Alto (CA) (December 1998).

293
NEA/CSNI/R(2009)2

294
 NEA/CSNI/R(2009)2

295
NEA/CSNI/R(2009)2

296
 NEA/CSNI/R(2009)2

297
NEA/CSNI/R(2009)2

298
 NEA/CSNI/R(2009)2

299
NEA/CSNI/R(2009)2

300
 NEA/CSNI/R(2009)2

Probabilistic Evaluation in Support of Risk-Informed Pressure Tube Maintenance

Eric Nadeau, Atomic Energy of Canada Limited, Canada

Tom Byrne, DRT Technical Services, Canada
Siavash Khajehpour, Atomic Energy of Canada Limited, Canada
Dirk Leemans, Crosscurrents Research and Policy Consulting (Presently at NSS), Canada

Background

CANDU 6 nuclear reactors are comprised of 380 fuel channels (FC), each having a pressure tube (PT)
installed concentrically inside a calandria tube (CT). Four helicoidal annular springs, also known as
spacers, provide support to the PT and prevent contact between the two tubes.

In 1983, the PT in channel G16 at the Pickering A Unit 2 ruptured unexpectedly. It was found that, in the
older CANDU design, the loose fitting spacers might move from their design locations resulting in long
unsupported spans. This eventually leads to the PT contacting the CT, with the contact spreading axially
over time. Because the moderator on the outside of the CT is much cooler (about 70oC) than the PT
(>260 oC), this results in a cold region in the PT at the location of contact.

During the operation of the reactor, some of the hydrogen isotope formed by the corrosion process is
absorbed into the PT in the form of deuterium. Also, at the time of manufacturing, there was a certain
amount of initial hydrogen (Hinit) present in the PT material. The combination of the initial hydrogen and
the deuterium added during operation is called the hydrogen equivalent concentration (Heq). It is called
equivalent because each deuterium atom has an equivalent effect of a hydrogen atom (protium).

If the Heq in the PT exceeds a limit called the blister formation threshold (BFT), the hydrogen can
precipitate in the form of hydrides at the cold spot to form zirconium hydride blisters. Once the blister
depth reaches a critical value, the blister may crack resulting in an initiation site for delayed hydride
cracking (DHC). The resulting crack growth can lead to the rupture of the PT.

Since blisters can only form when contact occurs between the PT and the CT in regions where the bulk Heq
exceeds BFT, the industry has been managing the prevention of blister formation by ensuring that, at all
times, at least one of these conditions is not met. Specifically, the industry has been relying on the Spacer
Location and Repositioning (SLAR) program to prevent PT/CT contact, and in certain cases where contact
cannot be prevented, ensuring through measurements that the bulk Heq in the contact region remains below
BFT during operation.

Probabilistic methodology

The selection of channels for the SLAR maintenance has historically been made using a deterministic
methodology for blister susceptibility assessments based on a set of inputs and acceptance criteria designed
to provide a conservative evaluation of the reactor core. While this methodology has been very successful
in preventing contact and blisters, it suffers from a number of limitations, the main ones being:

301
NEA/CSNI/R(2009)2

1. Neither the relative risk of PT failure associated with blister formation in individual channels nor the
overall risk to the reactor core can be calculated.

2. There is no single “worst-case” analysis because of the multi-parametric nature of the problem.

3. It is very difficult to select and prioritise channels for inspection, because the relative importance of
violating different acceptance criteria is subjective.

These limitations may lead to unnecessarily large inspection and maintenance programs. Therefore, to
assist the CANDU operators with a risk-informed fuel channel management program, AECL developed
a probabilistic methodology to assess the probability of pressure tube failure resulting from contact and
blisters. This new tool provides the operators with the information needed for ranking individual fuel
channels and making comparisons with other plant components. It can be used in combination with other
available information for the development of an optimised inspection strategy.

The probabilistic assessments were divided into two parts:

1. A probabilistic contact assessment.

2. A probabilistic assessment of blister formation and growth to a critical size, leading to PT failure.

The probability of contact between the PT and the CT tube in a CANDU fuel channel was estimated
using the Monte-Carlo method with 1000 realisations (or trials) per channel. Random sampling was
performed on input parameters affecting the FC bending deformation and thus the PT to CT contact. The
main distributed variables are: PT creep, end rotations, garter spring locations and geometrical dimensions.

Input distributions were based on data from in-reactor inspections and laboratory investigations. In
particular, the distributions of the PT creep and end-slopes were calibrated against PT/CT gaps measured
in the reactor. Two different benchmark methods were developed to achieve this goal. The potential
spacer movement was simulated in all applicable cases. The failure criterion was also distributed.

Blister assessments were carried out at each of the axial locations where contact was predicted for each of
the 1000 realisations per channel. The blister calculations used numerical integration of closed form
solutions, performed in Excel with macros rather than the Monte-Carlo method, resulting in accurate and
efficient results with extremely small CPU time. The calculations accounted for the variability in both the
DUR and equivalent blister depth at which delayed hydride cracking (DHC) is initiated. Other variables
such as the Hinit and BFT were considered as deterministic (i.e. they were not sampled with a distribution).

PT failure was considered to occur when a blister grows to a critical size leading to DHC initiation. This
definition is somewhat conservative as DHC initiation is a necessary but not sufficient condition for PT
rupture.

Sensitivity studies were performed to test the robustness of the probabilistic results against potential
changes or errors in the input distributions. It was found that the results remain within a relatively narrow
range. Most importantly, the base case and sensitivity studies consistently identify the same high-risk
channels with the expected addition or deletion of channels to the list depending on each sensitivity case.

302
 NEA/CSNI/R(2009)2

303
NEA/CSNI/R(2009)2

304
 NEA/CSNI/R(2009)2

305
NEA/CSNI/R(2009)2

306
 NEA/CSNI/R(2009)2

307
NEA/CSNI/R(2009)2

308
 NEA/CSNI/R(2009)2

309
NEA/CSNI/R(2009)2

310
 NEA/CSNI/R(2009)2

311
NEA/CSNI/R(2009)2

312
 NEA/CSNI/R(2009)2

313
NEA/CSNI/R(2009)2

314
 NEA/CSNI/R(2009)2

A Structural Reliability Model for the Evaluation of Failure Probabilities of Alloy

182/82 Welds in PWRS

C. Cueto-Felgueroso (Tecnatom S.A.)

Abstract

In 2005 a R+D program was launched in Spain aimed at obtaining an enhanced Structural Reliability
Model (SRM) to be used in the development of Risk-Informed In-Service Inspection applications for
piping. The program was funded by the CSN, Tecnatom and UNESA. The result of the program was a
new version of the WinPRAISE code, named WinPRAISE07, which incorporates several new capabilities
in comparison with its predecessor. One of these features is a module for the calculation of failure
probabilities in Alloy 182/82 welds in PWR plants as described below.

The model for weld metals is based on the one used for base metal. For Alloy 600 the probability of crack
initiation under PWR conditions has been estimated from the literature. The functional form for the initiation
time is
1
= Ae −Q / RT σ n (1)
tI

This equation is applicable for stresses greater than some threshold, with the threshold value being
approximately equal to the yield strength.

Equation (2) is also used for the initiation time in the weld metal Alloy 182. The most complete set of data
found in the literature was in reference [3] and was used as the starting point in the data analysis. The data
represents initiation time as a function of the stress.

The probabilistic treatment of initiation of PWSCC cracks in WinPRAISE07 is similar to that for IGSCC in
austenitic materials in pcPRAISE [4] and previous versions of the code [5]. One difference, however, is that the
dependence of initiation from site to site can be described by within-heat and heat-to-heat variances.

The growth of a crack following initiation is also addressed in the code. The crack growth rates are
represented by the equation:
Q 1 1 
− ⋅ − 
da R T T
= Ce  ref  ( K − K th ) β
dt (2)

Values of m, Q,  and KIscc are constants and C is a random variable. The distribution of C is based on
reference [6] for Alloy 600 base metal and on reference [7] for the weld metals 82 and 182.

Probability of crack initiation due to PWSCC, as well as probabilities of through wall cracks, big leaks and
ruptures were evaluated for representative components in PWR plants:
− RPV Hot Leg Nozzle to safe-end welds.
− RPV Cold Leg Nozzle to safe-end welds.
− Pressuriser nozzle to safe-end welds.

315
NEA/CSNI/R(2009)2

References

[1] D. O. Harris and D. Dedhia, “WinPRAISE 07. Expanded PRAISE Code in Windows”. Engineering
Mechanics Technology, Inc. Technical Report TR-07-1, January 2007.

[2] Amzallag C., Le Hong S., Pagès C. and Gelpi A. “Stress Corrosion Life Assessment of Alloy 600
PWR Components”, Proceedings of 9th International Conference on Environmental Degradation of
Materials in Nuclear Power Systems – Water Reactors”, TMS, 1999, pp. 243-250.

[3] Amzallag C., Boursier J.-M., Pagès C and Gimond C. “Stress Corrosion Life Assessment of Alloy 182
and 82 Welds Used in PWR Components”, Proceedings of 10th International Conference on
Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors”, NACE, 2002.

[4] Harris D. O. and Dedhia D. “Theoretical and User’s Manual of pc-PRAISE. NUREG/CR-5864” (1992).

[5] Harris D. O. and Dedhia D. “WinPRAISE 98. PRAISE Code for Windows”. Engineering Mechanics
Technology, Inc. Technical Report TR-98-4-1, April 1998.

[6] White G. A., Hickling J. and Mathews L. K. “Crack Growth Rates for Evaluating PWSCC of
Thick-Wall Alloy 600 Material”, Proceedings of 11th International Conference on Environmental
Degradation of Materials in Nuclear Power Systems – Water Reactors”, ANS, 2003, pp. 166-179.

[7] White G. A., Nordmann N. S., Hickling J. and Harrington C. D. “Development of a Crack Growth
Rate Disposition Curves for Primary Water Stress Corrosion Cracking (PWSCC) of Alloy 82, 182
and 132 Weldments”, Proceedings of 12th International Conference on Environmental Degradation
of Materials in Nuclear Power Systems – Water Reactors”, TMS, 2005, pp. 511-531.

316
 NEA/CSNI/R(2009)2

317
NEA/CSNI/R(2009)2

318
 NEA/CSNI/R(2009)2

319
NEA/CSNI/R(2009)2

320
 NEA/CSNI/R(2009)2

321
NEA/CSNI/R(2009)2

322
 NEA/CSNI/R(2009)2

323
NEA/CSNI/R(2009)2

324
 NEA/CSNI/R(2009)2

325
NEA/CSNI/R(2009)2

326
 NEA/CSNI/R(2009)2

327
NEA/CSNI/R(2009)2

328
 NEA/CSNI/R(2009)2

329
NEA/CSNI/R(2009)2

330
 NEA/CSNI/R(2009)2

Activities of the PFM Subcommittee of the Japan Welding Engineering Society

on Probabilistic Fracture Mechanics

Noriyoshi Maeda, JNES, Japan

Shinobu Yoshimura, University of Tokyo, Japan
Yasuhiro Kanto, University of Ibaraki, Japan

1. Overview

With an early focus on usefulness of Probabilistic Fracture Mechanics (PFM), the Japan Welding
Engineering Society has been conducting research studies on PFM-related technologies since 1987.
Particularly, the PFM Subcommittee established within the Nuclear Engineering Research Committee of
the society has been conducting since 1996 “A research study on the applicability of PFM to the
assessment of reliability of structural components of nuclear power plants.” So far the subcommittee has
presented the study results at domestic and international conferences and in journals of related fields and
drawn an international attention as being advanced research group concerning PFM studies.

The objectives of this report are to discuss the representative PFM studies in Japan, analysis codes
developed and cases of application of these codes, as well as to propose how the further studies in the
future should be conducted.

2. Probabilistic fracture mechanics (PFM) and probabilistic risk assessment (PRA)

2.1 Fracture mechanics and PFM

In 2000, Japan government has introduced the code “Rules on Fitness-for-Service for Nuclear Power
Plants” of Japan Society of Mechanical Engineering (hereafter called “Fitness-for-Service Code”), based
on which cracks in nuclear power plant components found by the periodic inspections are required to be
evaluated using fracture mechanics and subsequent treatment and measures are specified. The code was
established based on the ASME Code Section XI and by incorporating research results in Japan including
fatigue crack growth rates under nuclear power plant environment and stress corrosion crack growth rates
of various materials, as well as by reflecting operation experience of power plants.

PFM can be regarded as a new form of evolution of fracture mechanics. In the structural evaluation
process of actual components, probabilistic variations and uncertainties exist in location, size and growth
rate of cracks, strength of material, and magnitudes and frequency of load applied. Conventionally, all of
such variations and uncertainties have been considered within the framework of the safety factor.
However, if the variation and uncertainty of each individual factor are conservatively estimated and simply
added together, the final evaluation result will end up being excessively conservative. To avoid this
difficulty, such factors are treated as random variables, and by precise evaluation of the correlation of such
random variables, the frequency (probability) of occurrence of the failure is estimated in a realistic system,
rather than considering whether a failure can occur under a deterministic condition. This approach is
called Probabilistic Fracture Mechanics (PFM).

331
NEA/CSNI/R(2009)2

2.2 Standard process of PFM analysis

The Standard process of PFM analysis is as follows. A failure analysis model is first built up, which is
capable to track the entire process of the failure (failure scenario) including generation of a crack, growth
of the crack and the failure or leakage of coolant caused by the crack in the component. Next, a random
variable space is configured taking various factors into account such as variations in the sizes of initial
cracks, variations in the loads applied during operation and their frequencies, variations in the growth rates
of cracks, probability of overlooking cracks during Pre-Service Inspection (PSI) or In-Service Inspection
(ISI), variations in the fracture toughness of materials, and uncertainty in decision making whether coolant
leaks out of through wall cracks.

When all of these preparations are done, the next step is to perform an actual simulation where one set of
variables related to a crack such as crack depth and crack length is selected (such a set is called a sample)
while considering the variations of each variable (probabilistic distribution), and then an analysis following
the failure scenario is performed to calculate the timing and causes of the failure. Such an analysis is
performed for a large number - maximum of about 100 million - of samples selected, and for each time
interval, the ratio of number of samples that lead to the failure to the total number of samples is calculated.
The calculated ratio gives a probability of failure at the time. In other words, this is an approach to check
every possibility thoroughly, taking variations and uncertainties of all relevant factors into account. Such
an approach where the samples randomly selected and analyzed by the computer to calculate the
probabilities is called the Monte Carlo simulation.

2.3 Relation between PFM and PRA

The product of the amount of damage caused by a single accident and the its frequency of occurrence
(probability of the occurrence) is called a risk. In modern societies, when we conduct a certain act for an
attempt to obtain some benefit, at the same time we generate a possibility (or a risk) of suffering damage.
Therefore, in our day to day conduct, it is quite important to accurately perceive probable risks and take
measures to avoid or reduce as much risks as possible, rather than to be preoccupied with an idea of only
increasing benefits.

In the case of nuclear power plants, Core Damage Frequency (CDF) is considered to be the most
fundamental risk. An approach to assess such risk in a quantitative manner is called Probabilistic Safety
Assessment (PSA) or Probabilistic Risk Assessment (PRA). In the approach of PSA or PRA, all events
that may possibly occur at a nuclear power plant are identified, and the frequency of occurrence of each
event as well as its effects on the core damage are quantitatively assessed, and then the safety level of the
plant is assessed based on the CDF values. The frequency of occurrence of an accident in the plant is
calculated based on the frequency of occurrence of the earliest event called “initiating event” and the
probability of losing the functions of systems for the mitigation of the effect of the accident. Generally, the
frequency of occurrence of an initiating event and the probability of loss of its function of the mitigation
system are derived from the evaluation of statistical data obtained from the accidents occurred in the past.
Nonetheless, the frequency of occurrence of an initiating event or the probability of equipment failure,
which has hardly or never occurred so far cannot be calculated from the statistical data.

On the contrary, PFM is capable to theoretically assess the probabilities of failures of nuclear power plant
structural components based on the realistic data such as material property changes caused by degradation due
to ageing, plant operation history and periodic inspection data. The result of PFM analysis is also quite
important to enhance the accuracy of PSA assessment. Additionally, even if the failures occurred to main
components are small enough to affect the CDF value; such failures may lead to a secondary accident or plant
shutdowns. For this reason, it is desirable to appropriately assess unscheduled plant shutdowns as a certain type

332
 NEA/CSNI/R(2009)2

of risks attributable to such failures, and make a routine effort to appropriately maintain the plant to reduce such
risks.

3. PFM analytical codes

In Japan, the following practical PFM codes have been developed by research institutes and other related
organisations, and gradually becoming available.

3.1 PASCAL (PFM analysis of structural components in ageing LWR)

(1) PASCAL ver. 2

The Japan Atomic Energy Agency has been developing a probabilistic fracture mechanics analysis code named
PASCAL (PFM Analysis of Structural Components in Ageing LWR) as part of the research on the integrity of
structural components of LWR. This code analyses the failure probability caused by the load exerted at the
event of reactor transient such as Pressurised Thermal Shock (PTS). This code first assesses the level of the
embrittlement of a reactor pressure vessel (RPV) using an embrittlement prediction equation developed in
Japan. Next, this code judges if the crack starts growth based on the comparison of its fracture parameter to the
fracture toughness at each time interval during the transient and judges if the crack causes failure by growing to
the size of wall thickness. As a ratio of the number of cracks that led to failures to the total number of cracks
sampled, conditional probability of start of propagation or a conditional probability of failure is calculated as
shown in Figure 1. Furthermore, PASCAL makes the analysis work easier with its Graphical User Interface
(GUI), through which the types of transient events such as LOCA, the chemical components of the steel that
affect the level of neutron embrittlement, the introduction of non-destructive inspection can be input. Based on
the development up to 2001, PASCAL Ver. 1 having functions including automated, stratified Monte Carlo
simulation, fracture evaluation with elastic-plastic fracture mechanics criterion, crack growth analysis for a
semi-elliptical crack and evaluation of recovery of fracture toughness by thermal annealing was released. Since
2002 various functional improvements were made, and a standard analysis method was proposed on the basis of
the sensitivity analysis results. And then PASCAL Ver. 2 with the GUI that incorporated the standard method
was developed.

Figure 1: Process of PASCAL calculation

333
NEA/CSNI/R(2009)2

(2) PASCAL-SC, EQ, EC

PFM analysis codes, PASCAL-SC, EQ, EC have been developed as a PASCAL series in order to assess
the growth of stress corrosion cracking, fatigue cracks and wall thinning in piping, respectively. These
codes are capable to assess various effects of seismic loads on these cracks and wall thinning. Preparation
of input data, execution of calculation and graphic representation of the calculation results can be done
through the GUI.

3.2 REAL-P (Creep fatigue failure probability calculation program)

Creep fatigue failure probability calculation program “REAL-P Ver.1.0” (hereinafter referred to as
REAL-P) was developed by the Japan Nuclear Cycle Development Institute and released in 2003. This
code estimates the failure probability of nuclear components by simulating the generation and growth of
cracks in the weld joints subjected to the internal pressure and thermal transients focusing on the effect of
creep fatigue and ductile fracture.

REAL-P is one of the programs of MSS-REAL (Material Strength and Structural Reliability Evaluation)
system including REAL-A and REAL-D program for the evaluation of material strength and is developed
to cope with the necessity to incorporate the probability theory into the System Based Code, which is
currently being developed. The System Based Code has an original structure of codes for high temperature
structural design which aims at the commercialisation of fast breeder reactors by improving existing codes
fundamentally. The goals of the System Based Code is to enhance the reliability, safety and economic
efficiency simultaneously by finding the optimum combination of the factors such as postulated events,
materials, fabrication, installation, pre-service inspection, operation and maintenance while taking into a
variety of technical options into account. REAL-P is a program in the System Based Code scheme to
calculate the probability of failure of the components in service, considering various technical options that
are incorporated into probability theory and derives a combination of probabilities of the generation of
cracks due to creep fatigue and the growth of initial cracks up to the failure. The calculated result is treated
as an index to develop the System Based Code system. Additionally, REAL-P can be applied not only to
the System Based Code but to the assessment of crack (defect) that can cause troubles of reactor
components by calculating the behaviour of growth of cracks during the operation and the probability of
failure due to the growth of initial cracks.

3.3 PEPPER (Probabilistic evaluation program for pipe aiming economical and reliable design)

PEPPER is a non-linear probabilistic fracture mechanics analysis code developed by Tepco Systems
Corporation for the purpose of assessment of reliability of piping. PEPPER is capable to evaluate fatigue,
SCC, creep and wall thinning as the degradation modes. To provide fracture mechanics parameters,
PEPPER has functions of calculation of the stress intensity factor, J-integral, and creep J-integral. The
basic sampling method is the classical Monte Carlo sampling method; however, to improve the accuracy
and to reduce the assessment time, PEPPER adopts the stratified Monte Carlo sampling method for crack
sizes and the weighted Monte Carlo sampling method for seismic loads whose frequency of occurrence is
very low. Other random variables are (1) applied load, (2) crack growth rate, (3) material strength, (4)
through wall (leakage) condition, (5) defect detection probability, and (6) measurement error of defect
geometry.

In addition to the behaviour of growth of actual cracks, PEPPER is capable to perform a hypothetical
assessment of defects based on the Fitness-for-Service Code or the “Rules on Pipe Wall Thinning
Management” of JSME, and therefore PEPPER can assess the probability of failure resulting from the
difference between the actual phenomena and controlling conditions imposed by the applicable standards.
Consequently, not only the probability of failure due to a hypothetical defect that has been the subject of

334
 NEA/CSNI/R(2009)2

classical PFM codes but also the reliability of piping under a controlled condition can be assessed. Thus,
PEPPER can propose a methodology of rational piping control based on the reliability.

By combining with GENPRO, a pre-code that assesses the probability of generation of cracks due to a
creep-fatigue damage, the probability of generation of cracks for each assessment year can be assessed.
(Measured probability of generation of defects can be also input, if available) Therefore, the probability of
failure of piping can be assessed, reflecting damages in the actual components.

Additionally, PEPPER-M developed in 2007 can assess the SCC behaviour including (1) crack initiation,
(2) crack growth, (3) crack coalescence, (4) plastic collapse of components having multiple cracks and
make it possible to assess the reliability of components having defects caused by material degradation
dominant in LWR.

3.4 PLEAS-PP (Plant life evaluation simulation system - PFM analysis program for piping system)

This is a PFM code that Kawasaki Heavy Industries, Ltd. developed uniquely to assess the piping integrity.
This code features its fast calculation process by utilising the table form and the weighted sampling method
to reduce volume of calculations, as well as the capability to perform an assessment compatible with the
Fitness-for-Service Code.

4. Cases of application of PFM codes

In addition to the PFM codes development, researches on probabilistic failures of components is being
conducted using PFM code mentioned above, PRAISE code or other PFM codes developed by research
organisations.

4.1 Assessment of failure probabilities of PWR RPV subjected to PTS

The probabilities of initiation, growth and arrest of growth of cracks in the PWR RPV were calculated
using PASCAL code. The RPV is postulated to be embrittled by neutron irradiation and subjected to PTS
transient. Through the sensitivity analyses, the most standard procedure of PFM analysis for RPV was
prepared as a guideline and round robin problems were given to check the PFM analytical code and the
analysis procedure.

With the objective of confirming the effects of non-destructive inspections conducted during the ISI, the
conditional probabilities of failure for the transients having typical frequencies were calculated by
changing the level of inspection (ratio of welds to be inspected to total welds). The calculation showed
that the conditional probabilities of failure decrease as the level of non-destructive inspection increases. It
was also realized that the conditional probabilities of failure were significantly reduced by introducing
inspections having higher accuracy. Based on these results, a study on the optimisation is about to start.

The conditional probabilities of crack initiation (CPI) under various transients were calculated using
PASCAL code as a function of neutron fluence and the results were shown in Figure 2. On the other hand,
the temperature margin ∆T for the failure (∆T=TKImax-TKIC=KImax, TKImax : the temperature where stress
intensity factor of the crack has maximum value at the transient, TKIC=KImax : the temperature where fracture
toughness of the RPV steel has equal value with the stress intensity factor of the crack) was
deterministically calculated as a function of neutron fluence. It was perceived that the relation between
CPI and ∆T could be expressed by a common curve irrespective of the type of transients.

335
NEA/CSNI/R(2009)2

Figure 2: Effect of neutron fluence on the CPI under various transients

4.2 Defects in PWR RPV exempted from assessment by the Fitness-for-Service Code

The Fitness-for-Service Code defines allowable cracks exempted from assessment (ACEA) which can be
supposed not to affect the integrity of components during the remnant service life. The depth to wall
thickness ratio of ACEA is defined as a function of depth to length ratio. The function is deduced by the
requirement that the stress intensity factor of a crack is sufficiently smaller than the fracture toughness of
the component material during the normal plant operation, and notable growth of cracks due to fatigue
does not appear during the remnant life.

In order to understand the meaning of the ACEA from a probabilistic point of view, a RPV of PWR under
PTS transients having thickness t and one surface ACEA which is planar and semi-elliptical whose depth is
a and length, 2c was analysed by PASCAL code. As the result of the analysis, cracks having equal failure
probability were depicted as a relation of a/t and a/c, as shown in Figure 3. It is obvious that this curve did
not coincide with the curve provided in the Fitness-for-Service Code.

336
 NEA/CSNI/R(2009)2

Figure 3: Comparison of defects exempted from assessment

4.3 Assessment of probabilities of failure of LWR piping due to ageing

As a method to properly assess the safety of aged plants, the PFM analysis is frequently used because it
can assess the integrity of components by taking factors such as the loads applied to components and their
variations, and distribution of defect sizes and probability of existence of defects.

Focusing on fatigue and stress corrosion cracking as major degradation modes, scenarios of crack
initiation, growth, leakage and fracture of piping caused by cracks were prepared based on the domestic
and overseas research results, and a standard PFM analysis model was developed. For these scenarios,
using PEPPER code, various sensitivity analyses that included probabilistic variations in dominant
quantities and the probabilities of detection by inspection of cracks were conducted. Based on the results
of the analyses, a PFM analysis procedure, which was regarded as to be the most standard for aged LWR
piping was developed as a guideline.

4.4 Effects of variations in crack geometry in the application of the Fitness-for-Service Code

In the practice of the plant maintenance management in accordance with the Fitness-for-Service Code, if
cracks detected by the inspection are evaluated and judged to be insignificant for the safety, the operation
is allowed to continue without repairing. A probabilistic validation of this process reveals that not only the
detectability of cracks but also the accuracy to measure the crack sizes are quite important. Using PEPPER
code, the sensitivity analyses conducted on the piping systems whose main failure mode was fatigue
showed that the effect of the variations in the accuracy to measure the crack sizes has significant effect in
small size (diameter) piping. Especially, in the case the thermal expansion was the dominant load, it was

337
NEA/CSNI/R(2009)2

shown that the probability of failure increased by taking the variations in the accuracy to measure the crack
sizes into account, and the probability would not significantly decrease, even if the safety factor was raised.
When SCC is dominant as a failure mode, this tendency became even more conspicuous.

4.5 Probabilistic assessment of the Leak-Before-Break (LBB) behaviour of LWR piping

In order to conduct a probabilistic assessment for the Leak-Before-Break (LBB) behaviour of LWR piping,
failure scenarios and an analysis model were identified, and input data was selected based on the domestic
and overseas research results. Additionally, in order to evaluate the applicability of LBB, an LBB
identification index was newly proposed. The sensitivity analyses using PEPPER code revealed that the
bigger the pipe diameter, the larger the applicability would be, and also the leak detection capability
significantly affected the probability of failure.

4.6 Seismic assessment of LWR piping

In order to assess the uncertainties of seismic loads quantitatively used in the assessment of structural
design and integrity of major structural components of nuclear power plants, a PFM analytical model that
takes into account the magnitude of earthquakes, frequency of earthquake occurrence (seismic hazard) and
variations in loads applied to the structural components during earthquakes was developed, and analyses
were conducted with PEPPER code. In the case of an excessive seismic load that exceeded the flow stress
of the material, a fracture mechanical approach was not necessary, therefore analyses with the condition
that the generated stress would not exceed the flow stress were conducted. The sensitivity analyses
revealed that loads other than the seismic load make the probability of failure of small-size piping high,
therefore such failure was hardly affected by earthquakes. On the other hand, the large-size piping, whose
probability of failure due to loads other than the seismic load was low, showed a significant increase in the
probability of failure when the seismic load was taken into accounts.

4.7 Probabilistic assessment of wall thinning behavior of LWR piping

Based on the difference between the calculated value using published prediction formula and actual value
by measurements for the wall thinning rates of piping, a probability density function for the wall thinning
rates was developed, and then a model to analyse the probability was developed. The sensitivity analyses
using PEPPER code by changing pipe diameters, operating stresses and inspection intervals revealed that
the probability of failure of small-size piping was high, and that it would be important, in order to reduce
the probability of failure, to increase the inspection frequency in accordance with the wall thinning rates, as
well as to improve the accuracy to predict the wall thinning rates by accumulating data by actual
measurements.

4.8 Plant maintenance optimisation based on risk, cost and profitability

If an inspection with high performance of crack detection is conducted with a high frequency, even small
cracks would be detected and failures would be decreased, however, the inspection cost would increase. In
the reverse case, unscheduled shutdowns due to accidents would increase, and thereby the electric power
sales would decrease and the costs for repair would increase. For this reason, an optimisation model for
the plant maintenance cost was developed, focusing on the profitability which is defined as the difference
between the sales of electricity and total cost including maintenance cost and operation cost. Results of
sensitivity analysis using PRAISE code revealed an inspection frequency which would give the maximum
profit. Sensitivity analyses including a change of piping material to that of having less susceptible to crack
initiation and smaller growth rate are planned.

338
 NEA/CSNI/R(2009)2

4.9 Assessment of probability of failure of heat transfer tube of steam generator of PWR

A PFM analysis model for heat transfer tubes of steam generators (SG) of PWR was developed assuming a
circumferential stress corrosion crack initiated at the boundary of enlarged section of a heat transfer tube as a
major degradation mode. Probability analysis models for a failure of single heat transfer tube and simultaneous
failures of multiple heat transfer tubes were developed. Based on these models, sensitivity analyses concerning
various maintenance strategies such as an improvement in defect detectability of non-destructive inspection,
extension of inspection intervals, sampling inspection and introduction of allowable cracks exempted from
assessment (ACEA) by the Fitness-for-Service Code were conducted using Dr. Mainte code developed by
Nuclear Fuel Industry, Ltd.. Evaluation for Inconel 600 revealed that the growth rates of SCC would
significantly affect the probability of failure. It was also suggested that an improvement in defect detectability
would be desirable to implement ACEA concept into the operation effectively.

4.10 Optimisation of maintenance of SG based on risk, cost and profitability

Based on the PFM analysis model introduced in 4.9, a model for the optimisation of SG maintenance
strategies to avoid unscheduled plant shutdowns was developed from the financial viewpoints including
costs for maintenance and counter measures for accident, and profitability of the plant. For the evaluation
of the profitability, the financial engineering techniques such as net present value and real option were
used. For Inconel 600, calculation using Dr. Mainte code revealed that the improvement in defect
detectability of inspection gave large influence to costs and profitability. It was also suggested that the
implementation of ACEA concept into the operation coupled with the use of materials which had better
SCC resistance would reduce maintenance cost effectively while securing the profit.

4.11 Improvement of PFM analysis model and analysis technique

In the practical PFM analyses, the Monte Carlo simulation is used as a basic technique for the calculation
of probabilities. To make the calculation faster, techniques such as weighted sampling method and
stratified sampling methods are adopted. In order to make the calculation still more faster, a PFM code
that implemented a parallel processing was developed. Additionally, PFM analyses normally proceed on
the assumption that cracks exist on the surface of the components. However, in reality, some portion of the
initial cracks may be buried under the surface of components. For this reason, a new PFM analysis model
that focused on the subsurface cracks was developed.

5. Implementation of risk-informed assessment in the future

To assess the risk in nuclear power plant, experts in nuclear technologies consider how to achieve safety
levels using the index Core Damage Frequency (CDF). On the other hand, leakages of coolant caused by
occasionally-occurring failures of piping or other components and subsequent unscheduled plant
shutdowns are dismissed as ordinary troubles that could occur in any components, as long as they do not
bring significant consequences in the CDF indices, and thus these failures tend to be treated lightly. CDF
approach may be a rational decision making from scientific point of view of safety. Nevertheless, in
reality, to continuously reassure the public by earnestly addressing such troubles that seem to be
technically minor will probably lead to build the confidence among the public for such an enormous
system technology that deals with radiation as a nuclear power plant. Therefore, in addition to achieving
“the safety levels” of the plant by the CDF assessment, the day-to-day solving of troubles would be an
important management objective for the enhancement of “public reassurance.” In other words, it would
be effective to categorise risks that endanger the safety and reassurance into two category; one includes
risks that endanger the safety from a scientific viewpoint (called type 1 risks), and the other includes risks
that endanger the public reassurance (called type 2 risks). Based on the realization of such types of risk,

339
NEA/CSNI/R(2009)2

the following two approaches would be taken when transferring from the deterministic assessment to the
risk-informed assessment.

A. To improve economic efficiency while maintaining the present safety and reliability levels.

B. To pursue the economic efficiency while making efforts to further improve the present safety and
reliability levels.

Here, two indices, i.e., safety and reliability are discussed. However, the influences to the safety of the
maintenance strategy needs to be distinguished from that of reliability and economic efficiency. In the
PSA, structural components such as piping (called passive components) tend to be treated lightly as
compared with dynamic components such as pumps (called active components). However, referring the
fact that many of the causes of long-term plant shutdowns are often attributable to leakages from passive
components and rupture of them, it is surmised that failures of major structural components would give far
greater influence to reliability and economic efficiency than to the safety. It is therefore desirable to
perform assessments while paying attention to such difference of influence, as well as considering the type
1 and type 2 risks at the same time as shown in Figure 4.

In Japan, The following activities are regarded as to be necessary in order to utilise the risk information for
realistic issues.

1) Development of assessment methodologies (Qualitative methodologies such as risk mapping,

Quantitative methodologies such as Probabilistic Fracture Mechanics.).

2) Development of database for assessment (Material data, inspection data, operation data, regulations).

3) Training of assessment experts.

4) Establishment of regulations and standards for risk-informed assessments.

Item (4) – regulations and standards is a challenging issue. Currently, Japan Nuclear Safety Commission is
in a process of developing the safety objectives, and therefore one of the possible solutions would be a
top-down approach that is to break down the commission’s objectives into elements, based on which the
regulations and standards for the risk-informed assessments would be build. The other solution would be a
bottom-up approach, just like the Fitness for Service Code, where the private sector develops the
regulations and standards for the risk-informed assessments and proposes the adoption to the government.

The implementation of the risk-informed assessments is indispensable to give nuclear power the status of
one of basic energies in the 21st century society where the public safety and reassurance will be required to
the utmost.

340
 NEA/CSNI/R(2009)2

Cost and financial data

Operation history data

Structural and material data

Ageing degradation Maintenance data (Inspection, Repair, Fitness for Service Code)
/damage model
Fatigue fracture PFM analysis Core damage scenario
Embrittlement/Ductile fracture
Plastic collapse
Stress corrosion cracking (SCC)
Erosion Probability of component failure
(Reliability assessment) PSA analysis

Type 2 risk
Unscheduled Core damage frequency
Cost/Profitability shutdown
(Economic assessment) (Safety assessment)

Guideline for social acceptability Type 1 risk

of implementation of risk-based Dialogue-based
maintenance
multidimensional visibility

Decision making support for maintenance optimization (Plant owners)

Figure 4: Comprehensive risk reduction based on type 1 and type 2 risks

341
NEA/CSNI/R(2009)2

342
 NEA/CSNI/R(2009)2

343
NEA/CSNI/R(2009)2

344
 NEA/CSNI/R(2009)2

345
NEA/CSNI/R(2009)2

346
 NEA/CSNI/R(2009)2

347
NEA/CSNI/R(2009)2

348
 NEA/CSNI/R(2009)2

349
NEA/CSNI/R(2009)2

350
 NEA/CSNI/R(2009)2

351
NEA/CSNI/R(2009)2

352
 NEA/CSNI/R(2009)2

353
NEA/CSNI/R(2009)2

354
 NEA/CSNI/R(2009)2

NURBIM Fatigue Benchmarking of SRRA Code for Risk-Informed ISI

Emil Milanov (Westinghouse)

Abstract

This paper describes the benchmark study of Westinghouse Structural Reliability and Risk Assessment
(SRRA) Model and Software (WCAP-14572, Revision 1-NP-A, Supplement 1) against the NURBIM
fatigue case. NURBIM is a project sponsored by the European Commission focusing on the definition of
best practice methodologies for performing risk analyses and establishing a set of criteria to accept
risk-informed-in-service-inspection (RI-ISI) as a valid tool for managing plant safety. The paper provides
step-by-step description of the methodology and results of the NURBIM fatigue benchmarking and
sensitivity studies of the Westinghouse SRRA models and associated software.

The NURBIM fatigue benchmark study was aimed to review five Structural Reliability Models (SRMs)
and the associated software for stress corrosion cracking (SCC) and fatigue. Three baseline cases have
been considered in the SCC part representing different pipe sizes (small, medium and large) and three
baseline cases in the fatigue part. Depending on the capabilities of the codes some of the input parameters
for the calculations were treated as fixed values, some as non-fixed (random). For the non-fixed
parameters, generally three different distributions are used - low, base and high values of uncertainty. For
the Fatigue Base Cases and Sensitivity Studies, the inputs for the SRRA model come directly from the
NURBIM Report for the low, base and high cases for small, medium and large pipe sizes, respectively.
The ISI parameter for the SRRA models is converted to a value compatible with NURBIM fatigue
benchmark study and used consequently as SRRA input. The SRRA input files were generated for the base
cases of three pipe sizes, with two different sets of uncertainties for each pipe size. The uncertainties were
on initial depth and flow stress for the NURBIM only set and on eight additional parameters for all the
NURBIM and SRRA uncertainties.

After the execution of the required SRRA programs, results are generated for base-case leak probabilities
at 0 and 40 years. These results are comparable to those from other participants in the NURBIM Fatigue
Study Report. The SRRA calculated base-case leak probabilities with all (NURBIM and SRRA)
uncertainties are about a factor of two higher at 0 years and about a factor of 10 higher at 40 years. For the
sensitivity studies the SRRA calculated changes in leak probabilities are also similar with those from other
SRMs in the NURBIM Report for the changes in flaw depth, aspect ratio, load level, number of cycles and
fatigue crack growth coefficient.

355
NEA/CSNI/R(2009)2

356
 NEA/CSNI/R(2009)2

357
NEA/CSNI/R(2009)2

358
 NEA/CSNI/R(2009)2

359
NEA/CSNI/R(2009)2

360
 NEA/CSNI/R(2009)2

361
NEA/CSNI/R(2009)2

362
 NEA/CSNI/R(2009)2

363
NEA/CSNI/R(2009)2

364
 NEA/CSNI/R(2009)2

365
NEA/CSNI/R(2009)2

366
 NEA/CSNI/R(2009)2

CANTIA-Probabilistic Code for CANDU Tube Inspection Assessment

S. T. Revankar and Brian Wolf, Purdue University, School of Nuclear Engineering, USA
J. Riznic, Canadian Nuclear Safety Commission, Canada

Abstract

The regulatory requirements for steam generator tubing structural integrity in Canada are stated in Clause
14 of CAN/ CSA – N285.4-94, Standard on Periodic Inspection of Nuclear Power Plant Components. A
flaw indication where the predicted wall loss exceeds the 40% of a wall thickness criterion, a
fitness-for-service assessment is required to demonstrate tube integrity. Canadian licensees are using
probabilistic modelling on regular basis to disposition flaws and to demonstrate compliance with these
requirements. In parallel with licenses activities, the CNSC developed a computer based methodology for
probabilistic assessment of inspection strategies for steam generator tubes. The program CANTIA was
created to fulfil this objective. CANTIA uses Monte Carlo techniques to determine approximate
probabilities of steam generator tube failures under accident conditions and primary to secondary leak rates
under normal and accident conditions at future points in time. The program also determines approximate
future flaw distributions and non destructive examination results from the input data. The probabilities of
failure and leak rates and the future flaw distributions can be influenced by performing inspections of the
steam generator tubes at some future points in time, and removing defective tubes from the population.
The effect of different inspection and maintenance strategies can therefore be determined as a direct effect
on the probability of tube failure and the primary to secondary leak rate. The paper presents preliminary
trial runs to demonstrate the importance of probability of detection (POD) and initiation of new cracks on
rupture and leak rate probabilities.

D:\Riznicj\CSNI WS_Risk-Informed Piping Integrity Management_2008\Revankar Et Al_Abstratc Rv2.doc

367
NEA/CSNI/R(2009)2

368
 NEA/CSNI/R(2009)2

369
NEA/CSNI/R(2009)2

370
 NEA/CSNI/R(2009)2

371
NEA/CSNI/R(2009)2

372
 NEA/CSNI/R(2009)2

373
NEA/CSNI/R(2009)2

374
 NEA/CSNI/R(2009)2

375
NEA/CSNI/R(2009)2

376
 NEA/CSNI/R(2009)2

377
NEA/CSNI/R(2009)2

378
 NEA/CSNI/R(2009)2

379
NEA/CSNI/R(2009)2

380
 NEA/CSNI/R(2009)2

381
NEA/CSNI/R(2009)2

382
 NEA/CSNI/R(2009)2

Probabilistic Modelling of Primary Water Stress Corrosion Cracking

(PWSCC) in Piping Dissimilar Metal Butt Welds

Bruce Bishop, Westinghouse Electric Company, LLC, USA

This paper describes the development, application and general risk results of a probabilistic fracture
mechanics (PFM) model used in a non-proprietary report MRP-116NP1 by the Electric Power Research
Institute (EPRI) that was submitted to the U.S. Nuclear Regulatory Commission (USNRC) for information.

1. Summary and conclusions of MRP-116NP

The risk evaluations documented in MRP-116NP1 have been intended to cover all the Alloy 82/182 butt
weld locations in operating pressurized water reactor (PWR) plants in the USA. The three goals of this
work are the following:

• Quantify the probability of leakage from both axial and circumferential flaws,
• Assess the impact of the calculated change in core damage risk per USNRC regulatory guide (RG)
1.1742 guidelines,
• Develop a recommendation as to the adequacy of the current American Society of Mechanical
Engineers (ASME) Section XI inspection requirements for these regions.

The methodology used for the evaluations summarised in MRP-116NP1 is consistent with that used in two
previous submittals developed by Westinghouse for the Pressurised Water Reactor Owners Group
(PWROG) that have been reviewed and approved by the USNRC. The first PFM model is for fatigue and
stress corrosion crack growth of an existing circumferential flaw in a stainless or carbon steel piping butt
weld described in WCAP-14572, Revision 1-NP-A, Supplement 13. The second PFM model is for flaw
size and time to initiate and grow an axial flaw in Alloy 600 base metal due to PWSCC in described
WCAP-149014. Key effects treated in these reports include crack initiation, crack growth, conditions for
failure of the pipe, and the consequences of a range of piping leaks on the conditional core damage
frequency (CDF).

Because there was no existing model for time to initiation of Alloy 82/182 weld material due to PWSCC,
the model for Alloy 600 base metal is used without any terms for material microstructure or the effects of
material processing. A new factor for weld material is also used so the model would match the observed
PWSCC induced flaws that were discovered in the vessel outlet nozzle weld of the V. C. Summer Nuclear
Plant after 17 years of operation. These flaws and operating conditions are described in a 2001 paper5
presented at the ASME Pressure Vessel and Piping (PVP) Conference. Note that cracking in the weld
metal only was considered in this study. There are a few cases where the safe end material is Alloy 600,
and these were not treated by this evaluation.

Since the existing PFM models for reactor vessel head penetrations already included crack growth due to
PWSCC, the new correlation for weld metal (Alloy 82, 132, and 182) developed by the Materials
Reliability Program (MRP) PWSCC Crack Growth Expert Panel6 are used directly. Only the input
parameters for the original Alloy 600 model need to be updated. The existing PFM models for fatigue

383
NEA/CSNI/R(2009)2

crack growth (FCG) for piping risk-informed in-service inspection (RI-ISI) in WCAP-14572 Supplement
13 and NUREG-16617 are used directly for the PFM models for butt welds. The USNRC Report on nickel
alloy welds8 indicates that FCG of Alloy 82/182 weld material is best characterised as a factor on the Alloy
600 CGR. Analyses of this data are used to develop the mean value and uncertainty for the factor on weld
material FCG. A benchmarking process is used to address the validity of these enhancements to the
existing PFM models. The conclusion is that the enhanced PFM models used to calculate the leak
probabilities for axial and circumferential flaws in Alloy 82/182 butt welds in piping nozzles have been
benchmarked with existing inspection failure (small-leak) data, such as that of Bamford and Hall9 and
independently verified to produce accurate results.

The benchmarked PFM model for PWSCC in Alloy 82/182 butt welds is applied for both axial and
circumferential flaws to a number of nozzle welds in typical PWR plants. The small leak (through-wall
flaw) probability calculated at 40 years of plant life with a 10-year inspection interval, ranges from
5.00E-05 to 1.69E-02 for axial flaws and from 2.70E-08 to 8.03E-03 for a circumferential flaw.
Probabilities for larger leaks that would be used for risk analyses of postulated small, medium or large loss
of coolant accident (SLOCA, MLOCA, or LLOCA) would have a lower value. A comparison of axial
versus circumferential leak probabilities shows that the axial probability is consistently higher than the
circumferential probability. The axial flaw length would be limited in extent to the interface with material
not susceptible to PWSCC, which is consistent with service experience. This evaluation considered this
length to be equal to the thickness of the weld, which is a conservative assumption.

Sensitivity studies are also performed on the effects of key input items to the PFM models, such as the
accuracy and frequency of in-service inspections. These studies indicate that there are several items that
would contribute to the conservatisms in the calculated leak probabilities. Two that result in the greatest
overall contribution are the treatment of weld residual stress and credit for leak detection. Residual
stresses are treated as a maximum value that is constant through the wall, with no relaxation as the flaw
grows through the wall. The capability for leak detection at all plants is required to be at least 1 gallon per
minute (gpm or ~4 liter/minute). However, the actual leak detection capability is now significantly better
than this value, typically by an order of magnitude. The results presented in MRP-116NP1 do not consider
this factor in determining the leak probability of either axial or circumferential flaws for any potential
break size or consequence. In addition, no consideration of plant mitigation actions in addressing leaks
was included in the determination of the conditional core damage probabilities.

The results of the risk assessment showed that the change in total plant risk was well within the RG 1.1742
guidelines for “insignificant change” when considering either CDF or large early release frequency (LERF)
for a 40-year plant life. Calculated CDF point estimate values for a 40-year life ranged from 2.81E-08 to
6.32E-07. These values for total plant change in risk were determined by combining the worst-case leak
probability for each location with a generic CCDP value of 3.00E-03 and combining the contributions from
the individual nozzles. The calculation for a plant-specific application would be lower in all instances.
The consequences for LERF would be lower than those for core damage, based on actual plant data.

While the plant specific input and results are proprietary and not included in this paper, some general
trends and critical conclusions from the PFM and risk evaluations are as follows:

• The fabrication history of the weld is a key contributor to leak probability due to its effects on
residual stress per MRP-10610 and MRP-11411 and time to initiation.
• Changes in inspection frequency or improvements in capability or accuracy have only a small
benefit.
• Risk results do not justify any required changes in the current 10-year ASME Code Section XI
inspection interval, as long as all Alloy 182/82 locations are included.

384
 NEA/CSNI/R(2009)2

A review of the critical nozzle leak and risk assessment results would suggest that the reactor pressure
vessel outlet nozzle is the most critical when considering potential leaks, plant reliability, and plant safety.
The pressuriser surge line, pressuriser spray, decay heat, and pressuriser safety and relief nozzle would
follow in order of concern.

2. References

1. Non-Proprietary Version, Materials Reliability Program: Probabilistic Risk Assessment of Alloy

82/182 Piping Butt Welds (MRP-116NP), EPRI, Palo Alto, CA, 2004. 1009806.

2. U.S. Nuclear Regulatory Commission, An Approach For Using Probabilistic Risk Assessment In
Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis. Regulatory Guide 1.174,
Revision 1, November 2002.

3. Westinghouse Structural Reliability and Risk Assessment (SRRA) Model for Piping Risk-Informed
In-service Inspection: February 1999. WCAP-14572, Rev. 1-NP-A, Supplement 1.

4. Background and Methodology for Evaluation of Reactor Vessel Closure Head Penetration Integrity
for the Westinghouse Owners Group: July 1997. WCAP-14901, Revision 0.

5. W. H. Bamford et al, Development of the Technical Basis for Plant Startup for the V. C. Summer
Nuclear Plant. Proceedings of the ASME 2001 Pressure Vessels and Piping Conference, Atlanta, GA
(2001).

6. Minutes of EPRI-MRP PWSCC Crack Growth Expert Panel Meeting, October 3, 2003–Gaithersburg,
Maryland, EPRI Letter PWR-MRP 2003-38 By John Hickling, October 20, 2003.

7. NUREG-1661, Technical Elements of Risk-Informed In-service Inspection Programs for Piping,

U.S. NRC, Office of Nuclear Regulator Research, January 1999.

8. U.S. Nuclear Regulatory Commission, Effects of Alloy Chemistry, Cold Work, and Water
Chemistry on Corrosion Fatigue and Stress Corrosion Cracking of Nickel Alloy and Welds: April
2001. NUREG/CR-6721 (ANL-01/07). O.K. Chopra, W.K. Soppet and W.J. Shack, Argonne
National Laboratory.

9. W. H. Bamford and J. F. Hall, A Review of Alloy 600 Cracking in Operating Nuclear Plants:
Historical Experience and Future Trends, in Proceedings, Eleventh International Conference on
Environmental Degradation of Materials in Nuclear Reactors-Water Reactors, American Nuclear
Society, 2003.

10. Materials Reliability Program: Welding Residual and Operating Stresses in PWR Alloy 182 Butt
Welds (MRP-106), EPRI, Palo Alto, CA: 2004. 1009378.

11. Materials Reliability Program: Evaluation of the Effect of Weld Repairs on Dissimilar Metal Butt
Welds (MRP-114), EPRI, Palo Alto, CA: 2004. 1009559.

385
NEA/CSNI/R(2009)2

386
 NEA/CSNI/R(2009)2

387
NEA/CSNI/R(2009)2

388
 NEA/CSNI/R(2009)2

389
NEA/CSNI/R(2009)2

390
 NEA/CSNI/R(2009)2

391
NEA/CSNI/R(2009)2

392
 NEA/CSNI/R(2009)2

393
NEA/CSNI/R(2009)2

394
 NEA/CSNI/R(2009)2

395
NEA/CSNI/R(2009)2

396
 NEA/CSNI/R(2009)2

Risk-Informed Leak-Before-Break Reconciliation with Modern Codes

and Standards for Plant Life Extension

M.J. Kozluk and X. Duan (Reactor Engineering Department, AECL)

Abstract

This paper presents and discusses the pivotal issues associated with the application of risk-informed
leak-before-break to disposition gaps that may exist between legacy plant designs and evolving design
codes and standards for the purpose of life extension. The paper uses the example of a retrofit requirement
that design Structures, Systems, and Components (SSCs) important to safety be designed to accommodate
the dynamic effects associated with the postulate rupture (i.e., pipe whip and jet impingement) of
high-energy piping systems.

The paper argues that the consequences arising from the postulated failures must be assessed in order that
the risk can be understood and managed appropriately. The paper describes how the elements of
leak-before-break could be used to argue that the probability of failure is sufficiently low that the
Postulated Initiating Event (PIE) could be excluded from the design basis and the role of in-service
inspections could be used to provide ongoing assurance that the probability of the PIE remains sufficiently
low throughout the remaining life of the plant.

For the purpose of discussion, the IAEA Nuclear Safety Guide NS-G-1.11 “Protection against Internal
Hazards other than Fires and Explosions in the Design of Nuclear Power Plant” is used as the modern
standard to be retrofitted to a Canadian multi unit CANDU nuclear generating station as part of a
systematic review of safety for the purposes of extending the life of the reactor units.


“CANDU” is a registered trade-mark of Atomic Energy of Canada Limited.

397
NEA/CSNI/R(2009)2

398
 NEA/CSNI/R(2009)2

399
NEA/CSNI/R(2009)2

400
 NEA/CSNI/R(2009)2

401
NEA/CSNI/R(2009)2

402
 NEA/CSNI/R(2009)2

403
NEA/CSNI/R(2009)2

404
 NEA/CSNI/R(2009)2

405
NEA/CSNI/R(2009)2

406
 NEA/CSNI/R(2009)2

407
NEA/CSNI/R(2009)2

408
 NEA/CSNI/R(2009)2

409
NEA/CSNI/R(2009)2

410
 NEA/CSNI/R(2009)2

411
NEA/CSNI/R(2009)2

412
 NEA/CSNI/R(2009)2

LIST OF PARTICIPANTS

413
NEA/CSNI/R(2009)2

414
 NEA/CSNI/R(2009)2

LIST OF PARTICIPANTS

BELGIUM
Mr. Emil MILANOV Tel: +32 6 728 8236
Principal Engineer Fax: +32 6 728 8336
Equipment Design and Ageing Management Eml: [email protected]
Westinghouse Electric Belgium S.A.
43 rue de l'Industrie
1400 Nivelles

Mr. Olivier NICOLAS Tel: +32 2 773 8434

Tractebel Engineering Eml: [email protected]
7 Avenue Ariane
B-1200 Brussels

CANADA
Mr. Keith S. DINNIE Tel: +1 416 592 8328
Director, Risk Informed Methods Fax: +1 416 592 4930
AMEC-NSS Ltd Eml: [email protected]
700 University Avenue - 4th Floor
Toronto
Ontario M5G 1X6

Mr. Michael J. KOZLUK Tel: +1 905 823 9060 ext 4326

Principal Engineering, Component Integrity Fax: +1 905 491 4248
Atomic Energy of Canada Limited (AECL) Eml: [email protected]
Mail stop: SP1-F2
2251 Speakman Drive
Mississauga, Ontario L5K 1B2

Mr. Eric NADEAU Tel: +1 905 823 9060 Ext 4632

Specialist Engineer Fax: +1 905 491 4248
Atomic Energy of Canada Limited (AECL) Eml: [email protected]
2251 Speakman Drive
Mississauga, Ontario L5K 1B2

Dr. Jovica RIZNIC Tel: +1 613 943 0132

Canadian Nuclear Safety Commission (CNSC) Fax: +1 613 943 1292
Engineering Assessment Division Eml: [email protected]
280 Slater Street
P.O.Box 1046, Station B
Ottawa, Ontario K1P 5S9

415
NEA/CSNI/R(2009)2

Mr. Duan XINJIAN Tel: +1 905 823 9060 Ext 6646

Senior Engineer Fax: +1 905 491 4248
Atomic Energy of Canada Limited (AECL) Eml: [email protected]
2251 Speakman Drive
Mississauga, Ontario L5K 1B2

Dr. Zhaojing ZENG Tel: +1 613 944 4035

Specialist Fax: +1 613 995 5086
Operational Engineering Assessment Division Eml: [email protected]
Canadian Nuclear Safety Commission (CNSC)
PO Box 1046 Station B
280 Slater Street

CZECH REPUBLIC
Ms. Hana DLOUHÄ Tel: +420 2 2162 4758
Nuclear Safety Inspector Fax: +420 2 2162 4398
State Office for Nuclear Safety Eml: [email protected]
Senovážné nám
Praha 1, 110 00

Mr. Jan DOTREL Tel: +420 2 6617 2614

Nuclear Research Institute REZ plc Fax: +420 2 2094 0519
Division of Integrity and Technical Engineering Eml: [email protected]
Husinec-Rez 130
250 68 REZ

Dr. Ladislav HORACEK Tel: +420 2 6617 3164

Nuclear Research Institute REZ plc Fax: +420 2 2094 0519
Division of Integrity and Technical Engineering Eml: [email protected]
Husinec-Rez 130
250 68 REZ

FINLAND
Dr. Kaisa SIMOLA Tel: +358 20 722 6725
Senior Research Scientist Fax: +358 20 722 6027
VTT Technical Research Centre Eml: [email protected]
P.O. Box 1000
FI-02044 VTT

FRANCE
Mr. Olivier ALLAIN Tel: +33 3 80 29 40 35
Safety Inspector Fax: +33 3 80 29 40 88
ASN/DEP French Nuclear Safety Authority Eml: [email protected]
15-17 avenue Jean Bertin
BP 16610
21 066 Dijon Cedex

416
 NEA/CSNI/R(2009)2

Dr. Stéphane CHAPULIOT Tel: +33 1 69 08 95 82

Chef de Laboratoire Fax: +33 1 69 08 87 84
CEA-Saclay Eml: [email protected]
DEN-DM2S-SEMT-LISN
91191 Gif sur Yvette

Mr. Claude FAIDY Tel: +33 4 72 82 72 79

Consultant Engineer Fax: +33 4 72 82 76 99
Edf-Septen Eml: [email protected]
12-14 Avenue Dutrievoz
69628 Villeurbanne Cedex

Mr. Stéphane MARIE Tel: +33 1 69 08 92 57

Engineer – CEA Fax: +33 1 69 08 87 84
DEN/DANS/DM25/SEMT/LISN Eml: [email protected]
91191 Gif sur Yvette

Ms. Rachel VAUCHER Tel: +33 3 80 29 40 33

Project Manager Fax: +33 3 80 29 40 88
ASN/DEP French Nuclear Safety Authority Eml: [email protected]
15-17, avenue Jean Bertin
BP 16610
21066 Dijon Cedex

GERMANY
Mr. Thomas SCHIMPFKE Tel: +49 221 2068 743
Gesellschaft für Anlagen- und Reaktorsicherheit mb Fax: +49 221 2068 834
Schwertnergasse 1 Eml: [email protected]
D-50667 Köln

Mr. Helmut SCHULZ Tel: +49 2207 7423

Consultant Fax: +49 1511 4948 764
Calenbergerweg 4 Eml: [email protected]
D-51515 Kuerten

Dr. Georg WACKENHUT Tel: +49 711 685 63044

Group Leader Fax: +49 711 685 63053
MPA, University Stuttgart Eml: [email protected]
Pfaffenwaldring 32
D-70569 Stuttgart

JAPAN
Mr. Naoki HIROKAWA Tel: +81 3 6364 2422
Chief Engineer Fax: +81 3 5245 8401
Tepco Systems Co. Eml: [email protected]
Shibusawa City Place Eitai
2-37-28 Eitai Koto-Ku
Tokyo, 135-0034

417
NEA/CSNI/R(2009)2

Mr. Tamio KORIYAMA Tel: +81 3 4511 1552

Senior Staff Fax: +81 3 4511 1598
Probalistic Safety Assessment Group Safety Analysis Eml: [email protected]
Japan Nuclear Energy Safety Organisation (JNES)
Kamiya-cho MT Bldg 4-3-20
Toramomon, Minato-ku

Mr. Noriyoshi MAEDA Tel: +81 3 4511 1711

Senior Staff Fax: +81 3 4511 1898
Japan Nuclear Energy Safety Organisation (JNES) Eml: [email protected]
Tokyu-Reit Toranomon Bldg. 3-17-1
Toranomon, Minato-ku
Tokyo 105-0001

Mr. Shunsuke OGIYA Tel: +81 3 4511 1961

Safety Information Research Division Fax: +81 3 4511 1998
Japan Nuclear Energy Safety Organization (JNES) Eml: [email protected]
Tokyu Reit Toranomon Bldg. 3-17-1
Toranomon, Minato-ku
Tokyo 105-0001

Dr. Kunio ONIZAWA Tel: +81 292 82 6039

Leader, Reactor Component Reliability Res. Group Fax: +81 292 82 5406
Japan Atomic Energy Agency Eml: [email protected]
Shirakata 2-4
Tokai-mura, Naka-gun, Ibaraki 319-1195

SPAIN
Mr. Carlos CASTELAO LÓPEZ Tel: +34 91 346 02 71
Consejo de Seguridad Nuclear - CSN Fax: +34 91 346 05 88
c/Justo Dorado 11 Eml: [email protected]
28040 Madrid

Mr. Jose M. CONDE Tel: +34 91 346 02 53

Head, R & D Office Fax: +34 91 346 05 88
Consejo de Seguridad Nuclear - CSN Eml: [email protected]
c/Justo Dorado 11
28040 Madrid

Dr. Carlos CUETO-FELGUEROSO Tel: +34 91 659 87 24

Tecnatom, SA Fax: +34 91 659 86 77
Av. Montes de Oca 1 Eml: [email protected]
San Sebastian de los Reyes
Madrid

Mr. Marcelo FERNANDEZ-BOLAÑOS Tel: +34 91 346 01 72

Mechanical & Structural Engineering Branch Head Fax: +34 91 346 02 16
Consejo de Seguridad Nuclear - CSN Eml: [email protected]
C/Justo Dorado 11
28040 Madrid

418
 NEA/CSNI/R(2009)2

Dr. Jose Maria FIGUERAS CLAVIJO Tel: +34 91 346 02 04

Technical Advisor for Structural Engineering Fax: +34 91 346 05 88
Consejo de Seguridad Nuclear - CSN Eml: [email protected]
C/Justo Dorado 11
28040 Madrid

Ms. Belen GARCIA GARCIA Tel: +34 97 781 88 78

Equipos y Materiales Eml: [email protected]
Dirección de Servicios Técnicos
Asociación Nuclear Ascó Vandellós, A.I.E
Ctra. N-340, Km. 1123
43892 L'hospitalet de l'Infant

Mr. Carlos GAVILAN MORENO Tel: +34 96 189 43 00 (Ext. 362)

Mechanical and Inservice Inspection Manager Fax: +34 96 219 64 77
Maintenance - Cofrentes Nuclear Power Plant Eml: [email protected]

Mr. Eduardo GUTIÉRREZ FÉRNANDEZ Tel: +34 91 767 58 54

Senior Engineer Fax: +34 91 768 60 64
Nuclear Generation Division Eml: [email protected]
Iberdrola Ingenieria y Construcción
Avda. Manoteras, 20, Build. D
28050 Madrid

Mr. Alvaro JUNGHANNS Tel: +34 91 346 06 65

Mechanical & Structural Engineering Technician Fax: +34 91 346 02 16
Consejo de Seguridad Nuclear – CSN Eml: [email protected]
C/Justo Dorado 11
28040 Madrid

Mr. Roberto LACALLE Tel: +34 63 582 22 44

Structural Engineering Eml: [email protected]
Nuclenor/INESCO
Hernan Cortés 26
Santander
Cantabria

Mr. José Angel MARTINEZ Tel: +34 91 346 02 85

Engineering Department Fax: +34 91 346 02 16
Consejo de Seguridad Nuclear - CSN Eml: [email protected]
C/Justo Dorado 11
28040 Madrid

Ms. Cristina MARTIN-SERRANO LEDESMA Tel: +34 91 767 58 54

Senior Engineer, Nuclear Generation Division Fax: +34 91 768 60 64
Iberdrola Ingerieria y Construcción Eml: [email protected]
Avda. Manoteras, 20, Build. D
28050 Madrid

419
NEA/CSNI/R(2009)2

Mr. Carlos MENDOZA GOMEZ Tel: +34 91 346 02 21

Mechanical & Structural Engineering Technician Fax: +34 91 346 02 16
Consejo de Seguridad Nuclear – CSN Eml: [email protected]
C/Justo Dorado, 11
28040 Madrid

Mr. Porfirio PÉREZ TEJEDOR Tel: +34 62 894 97 29

Inservice Inspection Manager Fax: +34 96 219 64 77
Maintenance - Cofrentes Nuclear Power Plant Eml: [email protected]
Iberdrola

Mr. Aranzazu RISQUEZ BAILON Tel: +34 91 767 58 54

Engineer, PSA Specialist Fax: +34 91 768 60 64
Iberdrola Ingenieria y Construcción Eml: [email protected]
Avda. Manoteras, 20, Build. D
28050 Madrid

SWEDEN
Dr. Bjorn BRICKSTAD Tel: +46 8 698 8427
SKI - Statens kärnkraftinspektion Fax: +46 8 661 9086
Swedish Nuclear Power Inspectorate Eml: [email protected]
Klarabergsviadukten
SE-106 58 Stockholm

Mrs. Karen GOTT Tel: +46 8 698 8454

Materials & Chemistry Fax: +46 8 661 9086
Department of Reactor Technology & Structural Inte Eml: [email protected]
Swedish Nuclear Power Inspectorate (SKI)
SE-106 58 Stockholm

Mr. Ake JONSSON Tel: +46 491 786 868

ISI Coordinator Fax: +46 491 786 090
OKG AB Eml: [email protected]
572 83 Oskarshamn

Mr Anders LEJON Tel: +46 340 66 85 79

RI-ISI Responsible Fax: +46 340 66 01 86
Quality & Inspection Eml: [email protected]
Ringhals AB
430 22 Väröbacka

Mr. Jan LOTMAN Tel: +46 173 81407

RI-ISI Applications Eml: [email protected]
Forsmarks Kraftgrupp AB
S-74203, Osthammar

Mr. Anders RICHNAU Tel: +46 340 66 71 70

Specialist Fax: +46 340 66 01 86
Quality & Inspection Eml: [email protected]
Ringhals AB
430 22 Väröbacka

420
 NEA/CSNI/R(2009)2

Mr. Erik WENNERSTRÖM Tel: +46 10 505 36 32

Technical Coordinator Eml: [email protected]
AF-TÜV Nord AB
Östra Sandgatan 14
Box 5014
250 05 Helsingborg

SWITZERLAND
Mr. Reinhard FUCHS Tel: +41 56 267 74 12
Section Head Technical Support Fax: +41 56 267 77 90
Kernkraftwerk Leibstadt AG Eml: [email protected]
5325 Leibstadt

Dr. Susanne F. SCHULZ Tel: +41 56 310 39 20

Physicist, Dr.rer.nat. Fax: +41 56 310 38 54
Swiss Federal Nuclear Safety Inspectorate (HSK) Eml: [email protected]
Section Mechanical and Civil Engineering
CH-5232 Villigen-HSK

UNITED KINGDOM
Mr. Adam TOFT Tel: +44 1925 25 3122
Nuclear Safety Consultant Fax: +44 1925 25 2285
Serco Technical and Assurance Services Eml: [email protected]
Walton House, Birchwood Park
Risley, Warrington
Cheshire, WA3 6GA, England

UNITED STATES OF AMERICA

Mr. Scott CHESWORTH Tel: +1 408 978 8200
Manager, Risk-Based Programs Fax: +1 408 978 8964
Structural Intergrity Asociates, Inc Eml: [email protected]
3315 Almaden Expressway
Suite No. 24
San Jose, CA 95118

Mr. Nathaniel COFIE Tel: +1 408 978 8200

Senior Associate Fax: +1 408 978 8964
Structural Integrity Associates Eml: [email protected]
3315 Almaden Expressway
Suite No. 24
San Jose CA 95118

Ms. Samantha CRANE Tel: +1 301 415 6380

Materials Engineer Fax: +1 301 415 5074
US Nuclear Regulatory Commission Eml: [email protected]
Two White Flint North
M/S T-10 M05
11545 Rockville Pike
Rockville MD 20852

421
NEA/CSNI/R(2009)2

Mr. Richard HAESSLER Tel: +1 412 374 5308

Principal Engineer Fax: +1 412 374 5099
Westinghouse Electric Co Eml: [email protected]
4350 Northern Pike
Monroeville PA, 15146

Mr. Bengt LYDELL Tel: +1 520 883 4335

SIGMA-Phase Inc Fax: +1 520 883 4335
16917 S. Orchid Flower Trl Eml: [email protected]
Vail AZ 85641-2701

Mr. Wallace NORRIS Tel: +1 301 415 6796

U.S. Nuclear Regulatory Commission Fax: +1 301 415 5074
Two White Flint North Eml: [email protected]
M/S T-10 E10
11545 Rockville Pike
Rockville, MD 20852-2738

Mr. Patrick O'REGAN Tel: +1 508 497 5045

4 Alexander Road Fax: +1 810 816 7052
Hopkinton, MA 01748 Eml: [email protected]

Prof. Shripad REVANKAR Tel: +1 765 496 1782

Associate Professor Fax: +1 765 494 9570
Purdue University Eml: [email protected]
School of Nuclear Engineering
West Lafayette, IN 47907

Dr. Fredric SIMONEN Tel: +1 509 375 1703

Laboratory Fellow Eml: [email protected]
Pacific Northwest National Laboratory
PO Box 999
Richland 99354, Washington

International Organisations

International Atomic Energy Agency, Vienna

Dr. Jiri MANDULA Tel: +431 2600 22788
Nuclear Engineer Fax: +431 2600 29598
IAEA Eml: [email protected]
Wagramer strasse 5
P.O. Box 100, A-1400 Vienna

OECD Nuclear Energy Agency, Issy-les-Moulineaux

Mr. Alejandro HUERTA Tel: +33 1 45 24 10 57
OECD-NEA Fax: +33 1 45 24 11 29
Nuclear Safety Division Eml: [email protected]
Le Seine St-Germain
12 bd des Iles
F-92130 Issy-Les-Moulineaux

422
 NEA/CSNI/R(2009)2

European Commission, Joint Research Centre, Petten

Dr. Arne ERIKSSON Tel: +31 224 56 53 83
Scientific Officer Fax: +31 224 56 56 41
Joint Research Centre(JRC) -EC Eml: [email protected]
P.O. Box 2
1755 ZG Petten

Dr. Luca GANDOSSI Tel: +31 224 56 52 50

Scientific Officer Fax: +31 224 56 56 41
Joint Research Centre (JRC) - Institute for Energy Eml: [email protected]
Westerduinweg 3
1755 LE Petten

423