Download:
pdf |
pdfJAPAN LESSONS-LEARNED PROJECT DIRECTORATE
JLD-ISG-2012-04
Guidance on Performing a Seismic Margin
Assessment in Response to the March 2012
Request for Information Letter
Interim Staff Guidance
Revision 0
November 16, 2012
ML12286A029
JAPAN LESSONS-LEARNED PROJECT DIRECTORATE
JLD-ISG-2012-04
Guidance on Performing a Seismic Margin
Assessment in Response to the March 2012
Request for Information Letter
Interim Staff Guidance
Revision 0
ADAMS Accession No.: ML2286A029
*Concurrence via e-mail
OFFICE
PM: NRR/JLD/PMB
LA: NRR/DORL*
QTE*
NAME
LRegner
ABaxter
JDougherty
DATE
11/05/2012
11/09/2012
11/02/2012
OFFICE
SLS: NRR/DE
DD: NRO/DDSEA
DD: NRR/JLD
NAME
KManoly
NChokshi*
RTaylor
DATE
11/15/2012
11/16/2012
11/16/2012
OFFICIAL RECORD COPY
November 16, 2012
ML12286A029
INTERIM STAFF GUIDANCE
JAPAN LESSONS-LEARNED PROJECT DIRECTORATE
GUIDANCE ON PERFORMING A SEISMIC MARGIN
ASSESSMENT IN RESPONSE TO THE MARCH 2012
REQUEST FOR INFORMATION LETTER
Purpose
The U.S. Nuclear Regulatory Commission (NRC) staff is providing this interim staff guidance
(ISG) as supplemental guidance to nuclear power reactor licensees on an acceptable method
for performing a seismic margin assessment (SMA) as referred to in the March 12, 2012, NRC
letter entitled, “Request for Information Pursuant to Title 10 of the Code of Federal
Regulations 50.54(f) Regarding Recommendation 2.1, 2.3, and 9.3, of the Near-Term Task
Force Review of Insights from the Fukushima Dai-ichi Accident,” hereafter called the “50.54(f)
letter.”
This document describes the enhancements to the NRC SMA method originally described in
NUREG/CR-4334, “An Approach to the Quantification of Seismic Margins in Nuclear Power
Plants,” which are needed to meet the objectives of the 50.54(f) letter. This ISG presents staff
positions on enhancements to the major elements of the NRC SMA and provides updated
references for the use of recent advances in both methods and guidance, including guidance in
the American Society of Mechanical Engineers/American Nuclear Society (ASME/ANS),
“Standard for Level 1/Large Early Release Frequency Probabilistic Risk Assessment for Nuclear
Power Plant Applications,” Standard ASME/ANS RA-Sa-2009, and the screening, prioritization,
and implementation document (SPID) currently under development by industry for NRC
endorsement.
This guidance, at this time, is only intended to be used for an enhanced NRC-method SMA
conducted in response to the 50.54(f) letter, and not for any other purposes. The NRC ISG
DC/COL-ISG-020, “Interim Staff Guidance on Implementation of a Probabilistic Risk
Assessment-Based Seismic Margin Analysis for New Reactors,” remains the NRC’s current
guidance for application of an SMA to new reactors licensing. A probabilistic risk assessment
(PRA)-based SMA is acceptable as long as the licensee follows the guidance in this ISG, the
SPID, and ISG DC/COL-ISG-020 (with the exception of elements that clearly and only relate to
new reactor licensing). The contents of this ISG have no implications for NRC ISG
DC/COL-ISG-020, the ASME/ANS probabilistic risk assessment (PRA) standard, or any other
document.
Licensees may propose other methods for satisfying the 50.54(f) letter. The NRC staff will
review such methods and determine their acceptability on a case-by-case basis.
Introduction
Following the events at the Fukushima Dai-ichi nuclear power plant in Japan on
March 11, 2011, the NRC established a senior-level agency task force referred to as the
Near-Term Task Force (NTTF). The agency tasked the NTTF with conducting a systematic and
methodical review of NRC regulations and processes and determining if the agency should
November 16, 2012
-1-
ML12286A029
make additional improvements to these programs in light of the events at Fukushima Dai-ichi.
As a result of this review, the NTTF developed a comprehensive set of recommendations,
documented in SECY-11-0093, “Near-Term Report and Recommendations for Agency Actions
Following the Events in Japan,” dated July 12, 2011. The NRC enhanced these
recommendations through interactions with stakeholders. SECY-11-0124, “Recommended
Actions To Be Taken without Delay from the Near-Term Task Force Report,” dated
September 9, 2011, and SECY-11-0137, “Prioritization of Recommended Actions To Be Taken
in Response to Fukushima Lessons Learned,” dated October 3, 2011, document the staff’s
efforts.
In accordance with the staff requirements memorandum (SRM) for SECY-11-0093, the NRC
staff reviewed the NTTF recommendations within the context of the NRC’s existing regulatory
framework and considered the various regulatory vehicles available to the agency to implement
the recommendations. SECY-11-0124 and SECY-11-0137 established the staff’s priorities for
the NTTF recommendations.
In March 2012, the NRC issued a 50.54(f) letter. Enclosure 1 of that letter, “Recommendation
2.1: Seismic,” described the actions related to seismic hazard and risk reassessments for
licensees to take in response to the letter. The SMA method is among the approaches
discussed in Enclosure 1, which may be appropriate for some plants depending on the outcome
of the hazard reassessment phase.
Enclosure 1 to the 50.54(f) letter states that, “[t]he SMA approach should be the NRC SMA
approach (e.g., NUREG/CR-4334, “An Approach to the Quantification of Seismic Margins in
Nuclear Power Plants,” issued in August 1985 (Agencywide Documents Access and
Management System (ADAMS) Accession No. ML090500182) as enhanced for full-scope
plants in NUREG-1407, “Procedural and Submittal Guidance for the Individual Plant
Examination of External Events (IPEEE) for Severe Accident Vulnerabilities”). The SMA
approach should include both core damage (accident prevention) and large early release
(accident mitigation).”
This document describes the enhancements to the NRC SMA method, originally described in
NUREG/CR-4334, that are needed to meet the objectives of the 50.54(f) letter. In addition, this
ISG presents staff positions on the major elements of the NRC SMA. This ISG also provides
updated references for the use of more recent advances in methods and guidance, including
guidance in the ASME/ANS standard and the SPID currently under development by industry for
NRC endorsement.
Three methods currently can be used to perform an SMA:
(1)
the PRA-based SMA method (as described in NRC ISG DC/COL-ISG-020)
(2)
the NRC SMA method (as described in NUREG/CR-4334, supplemented by
NUREG/CR-4482, “Recommendations to the Nuclear Regulatory Commission on Trial
Guidelines for Seismic Margin Reviews of Nuclear Power Plants,” and
NUREG/CR-5076, “An Approach to the Quantification of Seismic Margins in Nuclear
Power Plants: The Importance of BWR [boiling-water reactor] Plant Systems and
Functions to Seismic Margins”)
November 16, 2012
-2-
ML12286A029
(3)
the Electric Power Research Institute (EPRI) SMA method (as described in EPRI
NP-6041-SL Revision 1, “A Methodology for Assessment of Nuclear Power Plant
Seismic Margin” (EPRI, 1991)).
These three methods differ in two key areas: the initiators considered in the analysis; and, the
system logic model approach used.
This ISG addresses only the NRC SMA method. It does not address either the EPRI SMA
method or the PRA-based SMA method. The EPRI SMA method is not acceptable for satisfying
the objectives described in the 50.54(f) letter because of its use of success paths. In principle,
the full PRA-based SMA can be used as long as the licensee follows the guidance in this ISG,
the SPID, and ISG DC/COL-ISG-020 (with the exception of elements that clearly and only relate
to new reactor licensing). In addition, ISG DC/COL-ISG-020 does not address some specific
considerations pertinent to the 50.54(f) request (e.g. large early release frequency and mission
time).
The NRC staff will use this ISG guidance when reviewing the technical adequacy of enhanced
NRC-method SMAs submitted by licensees in accordance with the subject 50.54(f) letter. It
shall remain in effect until it has been superseded or withdrawn.
Public meetings on the proposed SMA guidance document were held to gain stakeholder input
before the guidance was formally issued for public comment. On September 10, 2012 (in the
Federal Register (FR), 77 FR 55510), the NRC requested public comments on draft
JLD-ISG-12-04. The comments received and staff responses are contained in “NRC
Responses to Public Comments,” for JLD-ISG-2012-04, which can be found at ADAMS
Accession No. ML12290A002.
Implementation
Except in those cases in which a licensee or construction permit holder proposes an acceptable
alternative method for complying with the 50.54(f) letter, the NRC staff will use the methods
described in this ISG to evaluate licensee submittals requested in the 50.54(f) letter.
Backfitting Discussion
Licensees may use the guidance in this document as one acceptable method for responding to
a portion of the information requested in the 50.54(f) letter. Accordingly, the NRC staff’s
issuance of this ISG is not considered backfitting, as defined in 10 CFR 50.109(a)(1).
Final Resolution
This ISG, or a portion thereof, may subsequently be incorporated into other guidance
documents, as appropriate.
Attachment
“Guidance on Performing a Seismic Margin Assessment in Response to the March 2012
Request for Information Letter.”
November 16, 2012
-3-
ML12286A029
References
American Society of Mechanical Engineers/American Nuclear Society, Standard ASME/ANS
RA-Sa-2009, “Standard for Level 1/Large Early Release Frequency Probabilistic Risk
Assessment for Nuclear Power Plant Applications,” 2009.
Electric Power Research Institute, “A Methodology for Assessment of Nuclear Power Plant
Seismic Margin,” EPRI Report NP-6041-SL, Revision 1, Palo Alto, California, 1991.
U.S. Nuclear Regulatory Commission, “An Approach to the Quantification of Seismic Margins in
Nuclear Power Plants,” NUREG/CR-4334, August 1985 (ADAMS Accession No.
ML090500182).
U.S. Nuclear Regulatory Commission, “Recommendations to the Nuclear Regulatory
Commission on Trial Guidelines for Seismic Margin Reviews of Nuclear Power Plants,”
NUREG/CR-4482, 1986 (ADAMS Accession No. ML12069A017).
U.S. Nuclear Regulatory Commission, “An Approach to the Quantification of Seismic Margins in
Nuclear Power Plants: The Importance of BWR Plant Systems and Functions to Seismic
Margins,” NUREG/CR-5076, 1988.
U.S. Nuclear Regulatory Commission, “Interim Staff Guidance on Implementation of a Seismic
Margin Analysis for New Reactors Based on Probabilistic Risk Assessment,” Interim Staff
Guidance DC/COL-ISG-020, March 15, 2010 (ADAMS Accession No. ML100491233).
U.S. Nuclear Regulatory Commission, “Recommendations for Enhancing Reactor Safety in the
21st Century, the Near-Term Task Force Review of Insights from the Fukushima Dai-ichi
Accident,” Commission Paper SECY-11-0093, July 12, 2011 (ADAMS Accession
No. ML11186A950).
U.S. Nuclear Regulatory Commission, “Staff Requirements – SECY-11-0093 – Near-Term
Report and Recommendations for Agency Actions following the Events in Japan,” Commission
Paper SRM-SECY-11-0093, August 19, 2011 (ADAMS Accession No. ML112310021).
U.S. Nuclear Regulatory Commission, “Recommended Actions to be Taken without Delay from
the Near-Term Task Force Report,” Commission Paper SECY-11-0124, September 9, 2011
(ADAMS Accession No. ML11245A158).
U.S. Nuclear Regulatory Commission, “Prioritization of Recommended Actions to be Taken in
Response to Fukushima Lessons Learned,” Commission Paper SECY-11-0137,
October 3, 2011 (ADAMS Accession No. ML11272A111).
U.S. Nuclear Regulatory Commission, “Staff Requirements – SECY-11-0124 – Recommended
Actions to be Taken without Delay from the Near-Term Task Force Report,” Commission Paper
SRM-SECY-11-0124, October 18, 2011 (ADAMS Accession No. ML112911571).
U.S. Nuclear Regulatory Commission, “Staff Requirements – SECY-11-0137 – Prioritization of
Recommended Actions to be Taken in Response to Fukushima Lessons Learned,” Commission
Paper SRM-SECY-11-0137, December 15, 2011 (ADAMS Accession No. ML113490055).
November 16, 2012
-4-
ML12286A029
U.S. Nuclear Regulatory Commission Letter to All Power Reactor Licensees et al., “Request for
Information Pursuant to Title 10 of the Code of Federal Regulations 50.54(f) Regarding
Recommendation 2.1, 2.3, and 9.3, of the Near-Term Task Force Review of Insights from the
Fukushima Dai-ichi Accident,” March 12, 2012 (ADAMS Accession No. ML12053A340).
U.S. Nuclear Regulatory Commission, “Draft Interim Staff Guidance on Performing a Seismic
Margin Assessment in Response to the March 2012 Request for Information Letter,”
JLD-ISG-12-04, September 4, 2012 (ADAMS Accession No. ML12222A327).
U.S. Nuclear Regulatory Commission, “NRC Responses to Public Comments: Japan Lessons
Learned Project Directorate Interim Staff Guidance JLD-ISG-2012-04: Performing a Seismic
Margin Assessment in Response to the March 2012 Request for Information Letter,” dated
November 16, 2012 (ADAMS Accession No. ML12290A002).
November 16, 2012
-5-
ML12286A029
Enhancements to the U.S. Nuclear Regulatory Commission Method
for Seismic Margin Assessment in Response to the March 2012
Request for Information Letter
1.0
Purpose
The U.S. Nuclear Regulatory Commission (NRC) staff is providing this interim staff guidance
(ISG) as supplemental guidance to nuclear power reactor licensees on an acceptable method
for performing a seismic margin assessment (SMA) as referred to in the March 12, 2012, NRC
letter entitled, “Request for Information Pursuant to Title 10 of the Code of Federal Regulations
50.54(f) Regarding Recommendation 2.1, 2.3, and 9.3, of the Near-Term Task Force Review of
Insights from the Fukushima Dai-ichi Accident,” hereafter called the “50.54(f) letter.”
This document describes the enhancements to the NRC SMA method, originally described in
NUREG/CR-4334, “An Approach to the Quantification of Seismic Margins in Nuclear Power
Plants,” which are needed to meet the objectives of the 50.54(f) letter. This ISG presents staff
positions on enhancements to the major elements of SMA and provides updated references for
the use of recent advances in methods and guidance, including guidance in the American
Society of Mechanical Engineers/American Nuclear Society (ASME/ANS), “Standard for
Level 1/Large Early Release Frequency Probabilistic Risk Assessment for Nuclear Power Plant
Applications,” Standard ASME/ANS RA-Sa-2009, (hereafter called the “ASME/ANS PRA
standard”) and the screening, prioritization, and implementation document (SPID) currently
under development by industry for NRC endorsement (see Section 4.1.2).
This guidance, at this time, is only intended to be used for an enhanced NRC method SMA
conducted in response to the 50.54(f) letter, and not for any other purposes. The NRC ISG
DC/COL-ISG-020, “Interim Staff Guidance on Implementation of a Probabilistic Risk
Assessment-Based Seismic Margin Analysis for New Reactors,” remains the NRC’s current
guidance for application of an SMA to new reactors licensing. A probabilistic risk assessment
(PRA)-based SMA is acceptable as long as the licensee follows the guidance in this ISG, the
SPID, and ISG DC/COL-ISG-020 (with the exception of elements that clearly and only relate to
new reactor licensing). The contents of this ISG have no implications for NRC ISG
DC/COL-ISG-020, the ASME/ANS PRA standard, or any other document.
Licensees may propose other methods for satisfying the 50.54(f) letter. The NRC staff will
review such methods and determine their acceptability on a case-by-case basis.
2.0
Key Terms and Concepts
This section defines key terms and concepts used in this ISG.
Accident Sequence–A representation in terms of an initiating event followed by a sequence of
failures or successes of events (such as system, function, or operator performance) that can
lead to undesired consequences, with a specified end state (e.g., core damage or early
release).
Attachment
-2Accident Sequence Analysis–The process to determine the combination of initiating events,
safety functions, and system failures and successes that may lead to core damage or large
early release.
Fragility–The conditional probability of the failure of a structure, system, or component (SSC) at
a given hazard input level. For seismic fragility, the input parameter could be peak ground
acceleration (PGA), peak spectral acceleration, floor spectral acceleration, or others. The
fragility calculation typically uses a double lognormal model with three parameters, which are
the median acceleration capacity (Am), the logarithmic standard deviation of the aleatory
(randomness) uncertainty in capacity (βR), and the logarithmic standard deviation of the
epistemic (modeling and data) uncertainty in the median capacity (βU). The aleatory and
epistemic uncertainty can be combined into a composite variability. The fragility using a
composite variability is referred to as the mean fragility.
Ground Motion Response Spectra (GMRS)–The site-specific spectra characterized by
horizontal and vertical response spectra determined as free-field motions on the ground surface
or as free-field outcrop motions on the uppermost in situ competent material using performancebased procedures in accordance with NRC Regulatory Guide 1.208, “A Performance-Based
Approach to Define the Site-Specific Earthquake Ground Motion.”
High Confidence of Low Probability of Failure (HCLPF) Capacity–A measure of seismic
ruggedness. HCLPF capacity is defined as the earthquake motion level at which there is a high
(95 percent) confidence of a low (at most 5 percent) probability of failure of a single SSC or of
an ensemble of them. It is formally defined (NUREG/CR-4334) using the lognormal fragility
model as Am exp [−1.65 (βR + βU)]. When the logarithmic standard deviation of composite
variability βC is used, the HCLPF capacity can be approximated as the ground motion level at
which the probability of failure is at most 1 percent. In this case, HCLPF capacity is expressed
as Am exp [−2.33 βC]. The conservative deterministic failure margin (CDFM) (methodology
described in EPRI Report NP-6041-SL Revision 1, “A Methodology for Assessment of Nuclear
Power Plant Seismic Margin” (EPRI, 1991), produces a HCLPF capacity estimate directly,
without developing the full fragility curve.
Large Early Release–The rapid, unmitigated release of airborne fission products from the
containment to the environment that occurs before the effective implementation of offsite
emergency response and protective actions such that there is a potential for early health effects.
Large Early Release Frequency (LERF)–The expected number of large early releases per unit
of time.
Min-Max Method–A method used to determine the HCLPF capacity of an accident sequence
from the HCLPF capacities of the contributing SSC failures, or the HCLPF capacity of the plant
as a whole from the HCLPF capacities of a group of seismic-initiated accident sequences. The
overall HCLPF capacity of two or more SSCs that contribute to a sequence using OR Boolean
logic is equal to the lowest individual HCLPF capacity of the constituents of the group. If AND
Boolean logic is used, the HCLPF capacity of the group is equal to the highest individual HCLPF
capacity of the constituents. When evaluating several accident sequences to determine the
“plant level HCLPF capacity,” the plant-level HCLPF capacity is equal to the lowest of the
sequence-level HCLPF capacities.
-3Review Level Earthquake (RLE)–A representation of an earthquake ground motion in the form
of a response spectrum (applied at a certain depth or location) used as the basis for the
analyses performed in a seismic margin assessment. The RLE is also often used as a “figure of
merit” for judgments based on the SMA as to the adequacy of the “seismic capacity” of an
individual SSC, of an accident sequence, or of the plant as a whole. Specifically, when
performing an SMA, an individual SSC’s seismic HCLPF capacity (or the capacity of an accident
sequence or of the plant as a whole) typically is compared to the RLE. If the HCLPF capacity is
greater than the RLE, the inference is that the capacity is “adequate” or that there is “adequate
seismic margin.” However, this latter judgment of adequacy depends on the application of the
SMA results as used by a decision maker. In the EPRI SMA methodology, and in some other
early SMA literature, the RLE is known as the “seismic margin earthquake,” but these are two
names for an essentially identical construct. For the purposes of addressing the 50.54(f) letter,
the RLE is the envelope of the safe-shutdown earthquake (SSE) and the GMRS, as discussed
in Section 4.3.1. The RLE should be applied at the GMRS location.
Safe Shutdown Earthquake Ground Motion (SSE)−The vibratory ground motion for which
certain SSCs are designed, pursuant to Appendix A to 10 CFR Part 100, to remain functional.
The SSE for the site is characterized by both horizontal and vertical free-field ground motion
response spectra at the free ground surface.
Seismic Equipment List (SEL)–The list of all SSCs that require evaluation in the seismic
fragilities task of a seismic margin assessment.
Soil Liquefaction–A fluid-induced loss of soil strength caused by seismic ground motion with two
typical failure modes: (1) flow failure where the shear strength of the soil drops below the level
needed to maintain stability; and (2) cyclic mobility failure (lateral spread). Either failure mode
can lead to excessive strains and displacements that could result in unacceptable performance
of supported SSCs.
3.0
Background, Overview, and Issues Related to the Seismic Margin
Assessment Method
3.1
Background on Seismic Margin Assessment
A panel of NRC-supported experts initially developed the SMA method between 1984 and 1985
to assess the capability of nuclear power plants to withstand earthquakes above their design
basis. The SMA method was first described in NUREG/CR-4334, which has been further
supplemented by NUREG/CR-4482, “Recommendations to the Nuclear Regulatory Commission
on Trial Guidelines for Seismic Margin Reviews of Nuclear Power Plants,” and
NUREG/CR-5076, “An Approach to the Quantification of Seismic Margins in Nuclear Power
Plants: The Importance of BWR Plant Systems and Functions to Seismic Margins.” Shortly
after the NRC’s development of the SMA method, the EPRI developed a related but different
SMA methodology, as described in EPRI Report NP-6041-SL (EPRI, 1991). These two
methods are commonly called the “NRC SMA method” and “EPRI SMA method,” respectively.
To assess the methods, researchers conducted a trial of the NRC SMA method at the Maine
Yankee Nuclear Power Plant (NUREG/CR-4826, “Seismic Margin Review of the Maine Yankee
Atomic Power Station”), a trial of the EPRI SMA method at the Catawba Nuclear Station (EPRI
Report NP-6359, “Seismic Margin Assessment of the Catawba Nuclear Station” (EPRI, 1988)),
and trials of both methods were conducted concurrently for the Hatch Nuclear Plant
-4(NUREG/CR-5632, “Seismic Margin Review of Plant Hatch Unit 1: System Analysis”). After
these three trial applications, the NRC endorsed using either the NRC or EPRI methods when
all of the licensed nuclear power plants undertook the IPEEE program in response to
Supplement 4 to NRC Generic Letter 88-20, “Individual Plant Examination of External Events
(IPEEE) for Severe Accident Vulnerabilities-10 CFR 50.54(f),” issued April 1991. However, in
the IPEEE staff guidance in NUREG-1407, “Procedural and Submittal Guidance for the
Individual Plant Examination of External Events (IPEEE) for Severe Accident Vulnerabilities,”
the NRC specified certain enhancements to the SMA methods if either was to be used to
respond to the IPEEE information request.
About two-thirds of the operating plants performed an EPRI SMA to satisfy the IPEEE
information request. Two performed an NRC SMA. The rest performed a seismic probabilistic
risk assessment (SPRA).
3.2
Comparison with Electric Power Research Institute Success Path-Based Seismic
Margin Assessment
The most important technical difference between the NRC SMA methodology and the EPRI
SMA methodology is the SMA systems analysis approach. The systems analysis develops the
seismic equipment list (SEL) that is the basis for the seismic fragilities part of the SMA
methodology. The NRC SMA method, as described in NUREG/CR-4334, uses a seismic PRA
fault-tree and event-tree approach to delineate accident sequences, although it is limited to only
a selected number of safety functions necessary to prevent core damage. The EPRI SMA
method uses a success-path approach in which two success paths are the basis for the
systems analysis. This approach defines and evaluates the HCLPF capacity of those SSCs
required to bring the plant to a stable condition (hot or cold shutdown) and to maintain that
condition for 72 hours.
Many of the other key systems-analysis assumptions are the same in both methods, including
assuming that the earthquake always causes unrecoverable loss of offsite power, that only
systems and components needed to accomplish certain core-damage-prevention functions are
within the scope, and that certain “screening tables” in the guidance reports can be used for
screening in or out major SSC categories. These screening tables were developed as the result
of expert judgments by the authors of the methodology guidance reports, based in turn on
earthquake experience data, test data, and various analyses that then existed in the literature.
Another key difference between the two methods is their methodological guidance for
developing the “HCLPF capacity”1 of an individual SSC. The EPRI method uses the CDFM
method for determining the HCLPF capacity, whereas the NRC method permits the CDFM
method, but prefers use of the fragility analysis (FA) method that is also known as the
“separation of variables” method. The CDFM method directly analyzes for the HCLPF capacity
of an individual SSC, whereas the FA method develops a fragility curve from which the HCLPF
capacity is extracted.
Part 5 of the ASME/ANS PRA standard, endorsed by the NRC in Regulatory Guide 1.200, “An
Approach for Determining the Technical Adequacy of Probabilistic Risk Assessment Results for
Risk-Informed Activities,” permits use of the CDFM method to derive the SPRA fragilities for
some applications (see note 1 to supporting requirement FR-F1 of the ASME/ANS PRA
1
See Section 2 of this document for a definition of the high confidence low probability of failure (HCLPF)
capacity.
-5standard). The fragility curve for an SPRA based on the CDFM method would be developed
from the HCLPF and assuming a generic composite β, as set forth in that standard.
3.3
Enhancements to the NRC and EPRI Seismic Margin Assessment Methods
Required in the IPEEE Program
When the NRC provided guidance in NUREG-1407 for performing the IPEEE, neither the NRC
SMA method nor the EPRI SMA method was judged adequate as originally developed.
NUREG-1407 specified certain enhancements if an SMA analysis was to be used in the IPEEE.
Today, based on experience gained over the two intervening decades, the NRC staff has
judged additional enhancements to be necessary if an SMA evaluation is to be used to satisfy
the 50.54(f) letter. These various enhancements, over and above what is in the original NRC or
EPRI guidance, are summarized in the following sections.
3.4
Features and Enhancements Necessary if an SMA is to be Used to Respond to the
NRC March 12, 2012, 50.54(f) Request for Information letter
A list of the high-level features and enhancements necessary if an SMA is to be used for the
purposes of responding to the 50.54(f) letter is presented below. Some of these topics are
similar to staff positions taken during the IPEEE program; others are additional enhancements.
Where appropriate, a staff position further describing the necessary enhancement is presented
in Section 4.
•
The SMA should use a systems-analysis approach that begins by following the NRC
SMA methodology, using event trees and fault trees, with enhancements; an EPRI SMA
approach using success-path systems logic is not acceptable.
•
The SMA should be a full-scope SMA, not a focused-scope or reduced-scope SMA
(as described in NUREG-1407; see Section 4.2.1 for additional details).
•
The systems model should be enhanced over what was contained in the original NRC
SMA guidance (in NUREG/CR-4334 and NUREG/CR-5076) and the NRC’s IPEEE
guidance (in NUREG-1407) (see Section 4.4 for more detail).
•
The scope should include certain containment functions and containment systems to
enable analysis of the plant-level HCLPF for large early release.
•
The “mission time” should extend to either 72 hours or when the plant reaches a stable
state, whichever is later (see Section 4.4.2).
•
The use of the so-called “min-max” method, if selected, should follow certain guidance
(see Section 4.6.2). The convolution method is the preferred method.
•
When developing sequence-level and plant-level HCLPF capacities, the analysis should
differentiate between those sequences that lead to core damage and those that lead to a
large early release.
•
The HCLPF capacities for sequences with non-seismic failures and human actions and
HCLPF capacities for sequences without them should be separately reported.
-6•
The SMA analysis should assume that the earthquake causes an unrecoverable loss of
offsite power. (The IPEEE specified the same assumption and is a standard assumption
in all SMAs.)
4.0
Staff Positions on Individual Technical Issues
4.1
Introduction
4.1.1
Organization of the staff positions
This section provides discussions and NRC staff positions on various technical issues. The
topics are broken into several subsections, as shown in the figure below. Additional guidance is
provided on the topics shown in the figure. Sections 5 and 6 of this document provide staff
positions on peer review and documentation.
-7-
4.1.2
The Screening, Prioritization, and Implementation Document
NRC staff currently is engaged with industry in developing guidance for some specific technical
elements needed to address the 50.54(f) letter. This applicable guidance will be provided in a
screening, prioritization, and implementation document (SPID), which has an expected
publication date of November 2012. The industry is developing the SPID with a significant level
of review and input from NRC staff; and it is anticipated that the SPID will be submitted to the
NRC by the Nuclear Energy Institute. Once the NRC staff completes its review of the SPID and
determines its adequacy, the NRC staff will endorse the SPID. To the extent appropriate, the
applicable SPID positions are incorporated into this ISG document, and Appendix A provides
some additional information on SPID positions.
-84.2
Seismic Margin Assessment Scope Issues
4.2.1
Introduction
Three methods currently can be used to perform an SMA: the PRA-based SMA method
(as described in ISG DC/COL-ISG-020), the NRC SMA method (as described in
NUREG/CR-4334, supplemented by NUREG/CR-4482 and NUREG/CR-5076), and the EPRI
SMA method (as described in EPRI NP-6041-SL Revision 1, “A Methodology for Assessment of
Nuclear Power Plant Seismic Margin” (EPRI, 1991)). The three methods differ in two key areas:
the initiators considered in the analysis and the system logic model approach used. As a result,
the three methods are appropriate for different applications and objectives.
The PRA-based SMA method and the NRC SMA method both use a fault-tree/event-tree
representation of the systems model; although the trees used in the NRC SMA method are
simplified compared to the PRA-based method. NUREG/CR-4334 limited the systems
considered in the NRC SMA method to those that support frontline functions, which is a
significant simplification over the PRA-based SMA method (and a seismic PRA). This ISG
enhances the systems considered to include those needed for the recirculation phase and cold
shutdown and systems that perform certain accident-mitigation and containment functions. The
EPRI method does not use an event-tree or logic-tree approach, but rather defines two success
paths that can be used to address a transient and a small loss-of-coolant-accident (LOCA).
While the PRA-based SMA method considers all potential initiators, both the NRC and the EPRI
SMA methods consider only seismically induced transients and small LOCAs.
This ISG addresses only the NRC SMA method. The EPRI SMA method does not achieve the
objectives described in the 50.54(f) letter because of its use of a success path approach (as
discussed above). In principle, the full PRA-based SMA can be used as long as the licensee
follows the guidance in this ISG, the SPID, and ISG DC/COL-ISG-020 (with the exception of
elements that clearly and only relate to new reactor licensing). In addition, ISG
DC/COL-ISG-020 does not address some specific considerations pertinent to the 50.54(f)
request.
In the IPEEE program, a tiered approach was taken and three “scope” levels were developed
for plants to use when taking into account estimates of site-specific hazard levels. The
definitions of these levels, termed “reduced scope,” “focused scope,” and “full scope,” are
provided in NUREG-1407. These IPEEE-based definitions are not used within this ISG, nor
within the recommended program, to address the 50.54(f) letter, except when specifically
referring to the IPEEE program or NUREG-1407. The use of the term “full scope” NRC SMA in
the context of this ISG and the 50.54(f) letter implies use of the NRC SMA with the
enhancements described in this document.
The NRC staff will determine the applicability of the SMA method to address the 50.54(f) letter
for each site during the screening and prioritization phase of the seismic reevaluation process.
The approaches described in this ISG are to be applied by all plants conducting a SMA in
response to the 50.54(f) letter. Plants for which an SMA is found to be appropriate may choose
to conduct a seismic PRA instead of an NRC SMA.
4.2.2
Addition of certain containment functions and containment systems to include
assessment of large early release
-9Technical Issue:
To understand the potential for large early radioactive releases, the scope
of the NRC SMA analysis as described in NUREG/CR-4334 needs to be
extended to evaluate accident sequences beyond the “early” “preventive”
safety functions in the original scope. Specifically, certain containment
functions and systems are to be included in the SMA’s scope.
Staff Position:
The SMA’s scope should be extended to include assessment of large
early release. Certain containment functions and containment systems
needed to address large early release are discussed in Section 4.4.1.
4.2.3
Differentiation between HCLPF capacities for core-damage sequences and for large
early release sequences
Technical Issue:
The 50.54(f) letter requires that information about sequences leading to
core damage and information about sequences leading to large
radioactive releases are both developed and reported.
Staff Position:
When analyzing for sequence-level and plant-level HCLPF capacities, the
SMA analysis should separately determine the HCLPF capacities for the
core-damage endpoint and for the large-early release endpoint.
4.2.4
Separate analysis of HCLPF capacities of sequences with and sequences without
non-seismic failures and human errors
Technical Issue:
In a typical seismic PRA, an important fraction of all of the accident
sequences involve a combination of failures caused by the earthquake
and other failures not related to the strong motion. Therefore, the SMA
analyst is required to separately determine the HCLPF for the accident
sequences containing only seismic failures and the HCLPF for the
accident sequences containing both seismic and non-seismic failures.
Neither non-seismic failures nor human errors are explicitly accounted for
using the “min-max” method in a traditional SMA when the HCLPF
capacity for an individual accident sequence is developed; nor are they
explicitly accounted for in aggregating to a “plant-level HCLPF capacity.”
The convolution method is one way to include non-seismic factors in the
quantification (see Section 4.6.1).
Staff Position:
4.2.5
When developing sequence-level and plant-level HCLPF capacities,
non-seismic failures and human errors should be included. The analysis
should separately determine the HCLPF capacities of sequences with
and sequences without non-seismic failures and human errors.
Relay chatter analysis and treatment of high-frequency response of certain SSCs
Technical Issue:
The analysis of the chatter of relays during earthquakes has long been a
technical concern. NUREG-1407 provided previous guidance for the
IPEEE program for full scope plants. The ASME-ANS PRA standard in
Part 5 and Part 10 also provides guidance. A related recent concern is
whether there are any other SSCs besides relays that are sensitive to
high frequency motions.
- 10 To address this long-standing question and to facilitate a more consistent
response to the 50.54(f) letter, industry has initiated a testing program of
potentially high frequency sensitive components that will serve as the
technical basis for new guidance. Industry representatives and NRC staff
are working jointly on the development of the testing program, which is to
be conducted in two phases.
The results of the testing program will be used to confirm the adequacy of
equipment response to high frequency input motions for plants for which
the GMRS exceeds the safe-shutdown earthquake (SSE) ground motion
in the high frequency range (>10hz). Plants that have exceedance only in
the high frequency range would screen out from performing an SMA or
SPRA, but would still have to address the performance of high frequency
components. The testing program will address the performance of
potentially high frequency sensitive components across the industry.
The testing program also will provide additional guidance and component
capacity information useful to plants undertaking further risk evaluations
because they have exceedances in both the high and low frequency
ranges. The results will be incorporated into the SPID as additional
guidance for addressing the 50.54(f) letter. Once the guidance in the
SPID is reviewed and endorsed by NRC staff, it can be used to address
the 50.54(f) letter.
Because Phase 2 of the program will finish after publication of the SPID,
a final report with additional guidance also will be issued. Once the
guidance in the final high frequency testing report is reviewed and
endorsed by NRC staff, it can be used to address the 50.54(f) letter as
well.
Staff Position:
This technical topic is covered in a separate high frequency testing
program. Once the NRC endorses the related guidance documents, they
can be used to address the 50.54(f) letter.
4.3
SMA Hazard, Ground Motion, and In-Structure Motion Issues
4.3.1
Selection of the Review Level Earthquake
Technical Issue:
The SMA methodology uses a review level earthquake (RLE)2 as the
ground motion level used in the analysis. The RLE is a representation of
an earthquake ground motion in the form of a response spectrum (applied
at a certain depth or location) that is used as the basis for the analyses
performed in a SMA.
While the same term was used in the IPEEE program, the method used
to determine the RLE for responding to the 50.54(f) letter is different. The
50.54(f) letter specifies that the licensee compare the site-specific GMRS3
with the plant’s SSE ground motion response spectrum and use the
2
3
In some documents, the term “seismic margin earthquake” or SME may be used.
Acceptable methods for development of the GMRS are described in the 50.54(f) letter.
- 11 envelope of these two spectra as the RLE. Although the method for
developing the RLE differs from earlier studies (such as the IPEEE), its
use within the SMA is similar.
Staff Position:
4.3.2
The RLE to be used in the SMA for a particular plant is the envelope of
the SSE and the GMRS over the entire frequency range. The method for
determining the RLE is specified in the 50.54(f) letter. The RLE should be
applied at the GMRS location.
Soil failures
Technical Issue:
Soil failure analyses include an evaluation for instability, excessive
settlement, and liquefaction. EPRI NP-6041 contains guidance on
performing these analyses. Fragility for seismically induced liquefaction
can be developed using the method described in Appendix G of
EPRI 1002988 report, “Seismic Fragility Application Guide” (EPRI, 2002).
In this method, the limit state may be defined in terms of the
consequences of liquefaction induced settlement on the site configuration
of safety-related SSCs, including site layout, umbilical between
structures, and buried pipes and concrete electrical ducts when adequate
justifications are provided. Additional guidance is found in the
ASME/ANS PRA standard.
Staff Position:
The assessment should include appropriate soil failure modes using EPRI
NP-6041 and EPRI 1002988. Additional guidance is found in the
ASME/ANS PRA standard. A more detailed evaluation of plant site
conditions using state-of-the-art approaches should be performed if soil
failure is deemed to have a significant potential.
4.3.3
Development of In-Structure Response Spectra
Technical Issue:
The assessment of structural response and the resulting in-structure
response spectra (ISRS) are important aspects of an SMA. Both the
overall amplitude and the shape of the response spectra used in the SMA
have a significant effect on the SMA results. Therefore, the ISRS must
be sufficiently accurate to provide confidence in the SMA results, in terms
of the core damage frequency, large early release frequency, and
dominant risk contributors identified.
At the same time, a significant amount of existing structural response
information in the form of structural models and ISRS (either from the
original design, IPEEE, or A-46 programs) is available for operating
reactors. Unfortunately, this information represents a wide range of
vintages, and the methods used to develop the information vary in their
consistency with current accepted practice. Use of this existing structural
response information, where appropriate, represents one avenue for
reducing the overall level of effort required; however, criteria to determine
the continued appropriateness of the models and information must be
applied to ensure that the objectives of the 50.54(f) letter are achieved.
- 12 In an effort to appropriately optimize the use of existing information, the
NRC and industry experts have been working to develop guidance on
several aspects of structural response that are included in the SPID and
this ISG. Appendix A discusses this guidance in detail. The topics
addressed are threefold:
•
Attributes of existing structural models needed for
appropriately addressing the 50.54(f) letter. As described in
Appendix A, existing structural “stick” models may be appropriate
if they have sufficient complexity and attributes to provide the
appropriate level of accuracy needed for the SMA (or PRA).
Appendix A provides specific criteria. Models not currently
meeting the criteria may be updated.
•
Scaling of ISRS. The scaling of ISRS is permitted, provided that
the spectral shapes of the original input motions and the new RLE
are similar and the use of scaling is documented and justified.
Scaling of dissimilar spectra is not permitted.
•
Use of fixed base models for soft rock conditions. The use of
fixed base models for structures founded on rock with shear wave
velocity greater than 3,500 feet/second is permitted, provided that
the general guidance in Appendix A and in the Staff Position
below is followed.
An experienced structural engineer should review the use of any existing
models or data, which also should be subject to peer review. Use of
existing information should be documented in the submission to the NRC
and should be adequately justified. The documentation and justification
provided should address any potential issues or deficiencies.
Staff Position:
Realistic ISRS should be calculated using the guidance on the use of
existing information and models provided in the SPID (with any exception
provided in the NRC endorsement of the SPID). Appendix A discusses
the SPID contents and guidance as it relates to development of the ISRS.
If an existing structural model is used, its attributes should be compared
to the SPID criteria and its applicability documented and justified. If ISRS
scaling is used, it should be consistent with current accepted practice and
SPID guidance on the use of scaling. The use of scaling should be
documented and justified. The SPID provides the technical basis
developed to support the use of fixed base models for structures founded
on rock with a shear wave velocity greater than 5,000 feet/second.
However, the use of fixed base models on softer rock may be justified in
some cases. The use of fixed base models for Vs>3,500 feet/second, if
used, should be justified.
An experienced structural engineer should review the use of any existing
models or data, which also should be subject to peer review. Use of
existing information should be documented in the submission to the NRC
and should be adequately justified. The documentation and justification
provided should address any potential issues or deficiencies.
- 13 4.3.4
Median seismic responses of systems and components
Technical Issue:
The system and component seismic responses should be
median-centered and based on current state-of-the-art or models and
assessment methods that meet the staff position in Section 4.3.3.
Staff Position:
Realistic equipment response should be calculated using ASME/ANS
PRA Standard Part 10 (High Level Requirement HLR-SM-C) or using the
staff position in Section 4.3.3.
4.4
SMA Systems Analysis Issues
4.4.1
Enhancements to the PRA-type systems SMA model beyond those in the original
guidance
Technical Issue:
The original NRC guidance for the systems-analysis aspect of an SMA
analysis is described in NUREG/CR-4334, supplemented by
NUREG/CR-4482 and NUREG/CR-5076. The NRC staff provided certain
enhancements in the IPEEE guidance (NUREG-1407). For an SMA
performed to address the 50.54(f) letter, the scope must be extended to
identify sequences involving a potential large early radioactivity release.
Certain features and enhancements are considered necessary to
accomplish this expanded scope. The staff position describes these
features and enhancements.
Staff Position:
The staff position is as follows:
Initiating events: The postulated seismic-caused initiating events should
include unrecoverable loss of offsite power, small LOCAs, and certain
transients, such as safety-relief-valve initiators. Seismic-initiated large
LOCAs need not be included.
Safety functions: The original SMA methodology requires consideration
of only the “Group A” safety functions as explained in NUREG/CR-4334
for pressurized-water reactors (PWRs) and NUREG/CR-5076 for
boiling-water reactors (BWRs). For both PWRs and BWRs, these include
subcriticality and early emergency core cooling system injection until
stabilization of temperature and pressure. These should be
supplemented by systems necessary to achieve core cooling and
long-term heat removal for times beyond the “early” period. Depending
on the specific design, this can mean including systems and functions
through the recirculation phase for a PWR, or through switchover to
suppression pool cooling for a BWR, and then establishment of residual
heat removal and other functions needed to bring the plant to a stable
state.
Scope of containment analysis: As identified in Section 4.2.2, the
systems-analysis scope should include certain containment functions and
containment systems to address large early releases. Examples include
containment penetrations and containment isolation systems,
- 14 containment pressure suppression and overpressure-protection systems,
containment heat removal systems (early and late), and hydrogen control
systems. These are needed so that the SMA can differentiate between
those rare seismic-initiated core damage accident sequences that lead to
large early releases and the larger number of sequences that do not.
Assessment of containment structural failure is outside the scope of an
SMA, as set forth in the original guidance.
Successes: The systems model should retain “successes” of those SSCs
for which detailed fragility calculations are not required because of their
strong seismic capacity. This should be done to facilitate review of the
accident sequences. For these SSCs, it is acceptable to use an
estimated or generic HCLPF capacity. Section 4.5.2 provides further
guidance.
4.4.2
“Mission time” for the accident analysis: until a stable state is reached
Technical Issue:
The seismic risk assessment required under the 50.54(f) letter seeks to
understand risk contributors for accident sequences that involve the
potential for core damage, or that involve the potential for a large release
of radioactivity. To understand the latter type of sequence, the SMA
analysis should study sequences for as long after the earthquake as is
necessary for the reactor to reach a stable state. That stable state might
be a “safe” state, or a state involving extensive damage to the core, or a
state involving a large radioactivity release, or somewhere in between.
The scope of this type of analysis is greater than that for an SMA
performed under the IPEEE guidance (NUREG-1407), which calls for a
72-hour mission time as the appropriate scope.
Staff Position:
For each potential accident sequence, the mission time for the safety
systems and functions that the SMA analysis evaluates should extend
either to 72 hours or to the time required to achieve a stable state,
whichever is longer.
4.4.3
Plant system and accident sequence analysis: Selection of the Seismic Equipment List
Technical Issue:
The SMA systems analysis involves, in part, the selection of an SEL.
This is a standard aspect of both SMA and seismic PRA. In addressing
the 50.54(f) letter, the SEL should include equipment needed to asses
both core damage and large early release. Some past guidance has
differed on the inclusion of large early release.
Staff Position:
The SEL should include the systems necessary to achieve cold shutdown
and the appropriate containment systems, as set forth in Section 4.4.1
above. The starting point for constructing the SEL is the set of SSCs
included in the internal events PRA model, to which a number of SSCs
with earthquake-specific issues must be added, such as passive
components not present in the internal events model whose seismic
failure could be important to core damage or large early release.
- 15 4.5
SMA Fragility and Capacity Issues
4.5.1
Plant walkdown methodology
Technical Issue:
The seismic walkdown is an important activity in any SMA. The
walkdown ensures that the seismic fragilities or margins are realistic and
plant-specific, and finds any as-designed, as-built, and as-operated
seismic vulnerabilities in the plant. It should be done in sufficient detail
and documentation so that the subsequent screening or margin
evaluation is traceable. EPRI NP-6041 and NUREG/CR-4334 provide
licensee guidance on walkdown.
Technical guidance on the walkdown is found in EPRI NP-6041 and in
Part 10 of ASME/ANS PRA Standard, which refers back to EPRI NP6041.
Staff Position:
4.5.2
The seismic walkdown should be conducted in conformance with either
Part 10 of ASME/ANS PRA Standard or EPRI NP-6041.
Screening approach and level for SSCs
Technical Issue:
The SEL SSCs generally can be broken into three categories or “bins”
based on their seismic capacity as compared to the RLE (and in-structure
response spectra) at the site of interest. These categories are “inherently
rugged” SSCs, “high seismic capacity” SSCs, and other SSCs (i.e., which
cannot be shown to have high seismic capacity such that they can be
screened out), as shown schematically in the figure below. SSCs that fall
into each of these bins are addressed differently within an SMA both in
the approach used to determine their capacity and their treatment in the
modeling process. By identifying components that are inherently rugged
or have high seismic capacity, attention and resources can be focused on
determining the capacity of SSCs with a greater likelihood of being risk
significant.
“rugged“ threshold
screening level
Increasing SSC seismic capacity
- 16 -
Inherently rugged SSCs are those extremely robust components believed
to have a seismic capacity beyond any realistic earthquake loading levels
(e.g., manual valves). These components generally are mutually agreed
upon by the technical community and, by convention, they are not
included in SMA or SPRA models. EPRI NP-6041-SL provides guidance
on inherently rugged components.
Other SSCs may be less rugged but still have sufficient capacity so that
their failures would be unlikely to contribute significantly to the seismic
core damage in a seismic SMA. These components, noted as high
seismic capacity SSCs, should still be incorporated into the model,
although detailed fragility calculations are not warranted. The screening
Tables 2-3 and 2-4 in EPRI NP-6041 and the capacity information in
EPRI TR-103959, “Methodology for Developing Seismic Fragilities,” and
EPRI 1002988 are helpful for assessing if a particular SSC has a seismic
capacity (HCLPF) higher than the screening level (which is generically
higher than the RLE.) Other sources of information, such as design
drawings and IPEEE analyses also may be useful. For the components
that can be classified as having high seismic capacity, detailed margin
evaluations need not be performed (except for anchorage calculations),
but the SSCs must be included in the systems model (both the event
trees and the fault trees) to assist the analyst in understanding the most
risk-significant accident sequences. The HCLPF capacities assigned are
set equal to the screening level. The design review and walkdown should
confirm the validity of this assignment.
- 17 Seismic fragilities must be calculated for all SSCs that are neither
inherently rugged nor shown to have high seismic capacity. These SSCs
are included in the model with their actual capacities determined using
the methods described in Sections 4.5.3 and 4.5.4 of this document.
In the past, different screening approaches have been applied in SMA.
This experience has shown that the ability to obtain risk insights from an
SMA can be significantly curtailed if the screening level is set too low. An
example was the IPEEE program, in which a number of plants found that
the “surrogate elements” (a stand-in for screened components)
dominated the accident sequences. This outcome greatly limited the
ability to identify the actual most risk significant SSCs.
These past lessons indicate that additional guidance must be provided to
ensure that the objectives of the 50.54(f) letter are met. As a result, the
NRC and industry experts have developed guidance on the appropriate
screening approaches and levels to be used for inclusion in the SPID.
That guidance is summarized below.
The risk information needed includes core damage frequency, large early
release frequency, and the identification of the dominant risk contributors,
as indicated in the March 12, 2012, letter. To gain the necessary risk
insights from an SMA, the screening level must be set well above the
RLE. Based on the results of the analyses performed to support the
guidance in the SPID, either of two criteria may be used for the initial
screening of SSCs.
The screening level may be set as either:
•
A screening level consistent with an HCLPF capacity that is
2.5 times the RLE, or
•
A screening level equivalent to the HCLPF that leads to a
frequency of failure on the order of 5x10-7/yr using a mean point
estimate approach, an assumed composite βc, and the site
hazard.
A selected screening level based on one of the above criteria can be
used in reviewing previous IPEEE, A-46, or design basis calculations to
judge if explicit fragility or HCLPF calculations are needed for each SSC.
The NUREG/CR-4334 and EPRI-NP-6041 screening tables can be used
to identify and assign conservative HCLPF values to the high seismic
capacity SSCs using a screening level that is higher than the RLE as
defined above. The use of these tables must include satisfying caveats
associated with the tables as well as anchorage evaluations, as
appropriate. This enhancement is from earlier guidance, and the IPEEE
and NUREG/CR-4334, which allowed the screening level to be set at the
RLE.
- 18 Once the SMA analysis has been performed, a check must be conducted
to ensure that none of the components identified as among the dominant
contributors to either the HCLPF capacity for core damage or the HCLPF
capacity for large early release are high seismic capacity SSCs. If any
high seismic capacity SSCs are identified as among the dominant
contributors, then the actual HCLPF capacities of the components must
be developed and the analysis rerun.
A check should also be performed to ensure that each of the non-seismic
failures associated with high seismic capacity SSCs are retained in the
model.
Staff Position:
When identification of high-seismic capacity components is performed,
the basis for identifying them, including supporting documents, should be
fully described. EPRI NP-6041-SL Revision 1 and NUREG/CR-4334
guidance may be used, provided the following enhancements are applied.
(Use of screening tables of NUREG/CR-4334 and EPRI NP-6041 must
include satisfying caveats associated with the tables and should include
anchorage evaluations, as appropriate.)
•
The components identified as having a high seismic capacity as
compared to the screening level should be assigned capacities
equal to the screening level unless a detailed fragility calculation
has been performed. These components should be retained in
the system model for accident sequence analysis.
•
The screening level may be set as either:
•
o
A screening level consistent with an HCLPF capacity that
is 2.5 times the RLE, or
o
A screening level equivalent to the HCLPF that leads to a
frequency of failure on the order of 5x10-7/yr using a mean
point estimate approach, an assumed composite
variability, and the site hazard.
Once the SMA analysis has been performed, a check must be
conducted to ensure that none of the following conditions exist:
o
A high seismic capacity SSC that was assigned a seismic
capacity equal to the screening level is identified as a
dominant contributor to HCLPF for core damage.
o
A high seismic capacity SSC that was assigned a seismic
capacity equal to the screening level is identified as a
dominant contributor to HCLPF of large early release.
If any of the above conditions exist, the screening level should be
reevaluated and adjusted and actual HCLPF capacities of the
components should be analyzed using the methods described in
- 19 Sections 4.5.3 or 4.5.4, and incorporated into the model to the extent
necessary, and the analysis should be rerun.
4.5.3
Fragility analysis (FA) method for evaluation of the HCLPF capacity of an SSC
Technical Issue:
The FA method is described in a number of references
(e.g., NUREG/CR-2300, “PRA Procedures Guide: A Guide to the
Performance of Probabilistic Risk Assessments for Nuclear Power
Plants,” NUREG/CR-4334, NUREG/CR-4482, and EPRI TR-103959).
Typically, the seismic fragility of a component is characterized by a
double lognormal model whose parameters are Am, βR, and βU. Am is the
median capacity. βR is the logarithmic standard deviation of the capacity
and represents the variability due to the randomness of the earthquake
characteristics for the same acceleration and to the structural response
parameters that relate to these characteristics. βU is the logarithmic
standard deviation of the median capacity and represents the
uncertainties in models and model parameters. The “seismic margin” is
defined in terms of the HCLPF capacity and is calculated using the
equation
HCLPF = Am exp[-1.64*( βR + βU)]
For some applications, it may be sufficient to develop a mean fragility
curve characterized by a lognormal probability distribution with
parameters of Am and βc, where βc = (β2R + β2U)½ is the logarithmic
standard deviation of composite variability. The HCLPF capacity is taken
as the 1 percent conditional-probability-of-failure value, i.e.,
HCLPF = Am exp(-2.33 βc).
For every component in the plant system model, the FA method
evaluates the family of fragility curves or a single composite fragility curve
from which the HCLPF capacity (known as seismic margin) is estimated.
As an alternate approach, the CDFM method (described below) can be
used with a generic βc to develop the fragility of an SSC. The use of
generic βc, along with recommended values, is found in the SPID (upon
SPID endorsement by the staff).
Staff Position:
4.5.4
Fragility analysis for SSCs should be performed either in accordance with
Part 5 of the ASME/ANS PRA standard or using the CDFM method with a
generic βc following the guidance in the SPID and Sections 4.5.4 and
4.6.1 below.
CDFM method for evaluation of the HCLPF Capacity of an SSC
Technical Issue:
The FA method of estimating the HCLPF capacity of SSCs, although
universally applicable, does require the median factors of safety for
different variables affecting the response and capacity to be estimated as
well as their logarithmic standard deviations. “Seismic margin” for any
SSC is defined in terms of the HCLPF capacity. The HCLPF capacity of
- 20 an SSC can be calculated directly using the CDFM method.
EPRI NP-6041-SL (EPRI, 1991) describes the CDFM method and
provides several examples. In this procedure, the values of different
variables that figure into the HCLPF capacity evaluation are judiciously
selected at median values (best estimate or structural model and
conservative estimate of median damping) or some conservative values
(e.g., code specified minimum material strength or 95 percent
exceedance actual strength if test data are available; 84 percent
non-exceedance ground response spectrum). The HCLPF calculation of
capacity follows the deterministic procedures used in the seismic design
and qualification of SSCs.
Staff Position:
If the HCLPF capacity is evaluated using the CDFM method, the analysis
should be performed in accordance with the EPRI NP-6041-SL, with any
necessary adjustments as discussed in EPRI TR-103959 (EPRI, 1994)
and EPRI 1019200 (EPRI, 2009).
4.6
SMA Integration Issues
4.6.1
Sequence-level and plant-level HCLPF capacity: plant margin evaluation using the
convolution method
Technical Issue:
The convolution method for evaluating plant margin is preferred over the
“min-max” method described in Section 4.6.2. In the convolution method,
accident sequences are evaluated by combining input fragility curves
according to the Boolean expression for each sequence. Seismic and
random or human failure probabilities are calculated and combined
(convolved) for discrete intervals of ground acceleration and then
integrated over the range of interest. It is also important to keep the
“success” events (i.e., the “PRA event” when a component does not fail)
in the calculation. The result is the family of fragility curves for each
accident sequence. By combining these accident sequences that result
in core damage or large early release, the plant level fragility curves are
obtained. The “plant-level seismic margin” is then evaluated as the
HCLPF capacity, defined as equal to the seismic ground acceleration at
which the probability of failure is equal to 5 percent with a confidence
level of 95 percent. This method of developing the plant level fragility
curves will retain the information for the analyst and the NRC to develop
accident sequence frequencies, CDF and LERF calculations using the
site-specific seismic hazard as needed. The method is described in
Appendix 5-A of the ASME/ANS PRA standard. If a single composite
fragility curve is input for each SSC in the accident sequences, the
resulting plant level fragility curve also will be a single curve and the plant
seismic margin will be the HCLPF capacity, defined in this case as the
ground acceleration at which the probability of failure is equal to 1
percent.
To simplify both the seismic PRA and SMA analyses, a hybrid method
suggested in EPRI Report TR-103959 (EPRI, 1993) and Kennedy (1999)
could be used. The main feature of this method is the development of a
seismic fragility starting with the HCLPF capacity. First, the HCLPF
- 21 capacity of the component is estimated using the CDFM method. Next,
the logarithmic standard deviation βc is estimated using judgment and
following the guidance given in Kennedy (1999). For structures and
major passive mechanical components mounted on ground or at low
elevations within structures, βc typically ranges from 0.3 to 0.5. For active
components mounted at high elevations in structures the typical βc range
is 0.4 to 0.6. When specific information is not available, values of βc as
provided in the SPID are recommended (upon SPID endorsement by the
staff). The median capacity is calculated using the equation
Am = HCLPF * exp (2.33 βc)
and an approximate fragility curve for the component is thereby obtained.
Using this composite fragility curve for each component in the system
model, the plant level fragility curve is obtained following the convolution
approach described above. Reed and Kennedy (1994) further
recommend that this approximate fragility curve be used for each
component in the systems analysis to identify the dominant contributors
to the seismic risk (e.g., core damage frequency). For a few components
that dominate the seismic risk, more accurate fragility parameter values
should be obtained and a new quantification done to obtain a more
accurate mean core damage frequency, and to confirm that the dominant
contributors have not changed.
Staff Position:
4.6.2
The convolution method for evaluating plant margin is preferred over the
“min-max” method described in Section 4.6.2. If the convolution method
is used, the analyst should perform a margin evaluation of accident
sequences using the composite fragility curves for SSCs, mean values of
random unavailabilities, and operator error rates. The analysis should
also take into account the success terms and include the SSCs that have
been assigned HCLPF capacities equal to screening level (and larger
than RLE). For the important accident sequences (i.e., direct core
damage, large early release, and low seismic margins), the full family of
fragility curves (Am, βR, and βU) and probability distributions on random
failure rates and operator error rates should be used in the convolution
procedure to obtain the accident sequence fragility curves and plant level
fragility curves. The seismic margin of the plant is then evaluated as the
plant-level HCLPF capacity, defined as the ground acceleration value
corresponding to the 95 percent confidence of not more than 5 percent
probability of failure.
Sequence-level and plant-level HCLPF capacity: guidance on using the “min-max”
method
Technical Issue:
The min-max method (see the definition in Section 2) is a way that an
SMA analysis can derive an approximate sequence-level HCLPF capacity
from the HCLPF capacities of individual SSCs that comprise the accident
sequence, or derive an approximate plant-level HCLPF capacity from the
most important (lowest) sequence-level HCLPF capacities that emerge
from the SMA analysis.
- 22 However, the min-max method sometimes is only a rough approximation
and it gives an HCLPF capacity that can be either higher or lower than
the “true” value derived from a full seismic PRA that does not make this
approximation. For the case of a single “accident sequence,” the
discrepancy in the reported sequence-level HCLPF capacity arises when
two (or more) HCLPF values for individual SSCs are close numerically.
The distortion can be either conservative (too low) or non-conservative
(too high), depending both on the AND-OR structure of the sequence’s
Boolean logic and on whether the actual (unknown) fragility curves for
these SSCs have steep or shallow shapes (small or large βc values.) The
plant-level HCLPF capacity can be similarly distorted.
The existence of this potential distortion requires the SMA analyst to be
alert to the problem. In cases like those mentioned, the SMA analyst
should derive the sequence-level HCLPF capacity using the convolution
approach.
The distortion is not important, however, in two cases: if a single SSC
“dominates” the HCLPF capacity of an accident sequence or if a single
accident sequence “dominates” the overall seismic risk profile. The
analyst must justify if the min-max approach is used.
Staff Position:
If the min-max approach is used, it should be done in accordance with
NUREG/CR-4334 and a justification should be provided that this
approach provides reasonable estimates for the sequences under
consideration and at the plant level. The convolution approach
(discussed in Section 4.6.1) is the preferred approach.
5.0
Staff Positions on Documentation
5.1
Documentation Content
Technical Issue:
To address the request in the 50.54(f) letter, the SMA should be
conducted and documented in accordance with the provisions in this ISG
and the 50.54(f) letter. Documentation to be submitted to the NRC
includes a number of elements, as shown below.
Staff Position:
For plants that perform an SMA, the following information is requested (in
each of the following elements a description of how the applicable ISG
positions are met should be included. Any alternate approach should be
clearly identified along with its technical basis):
(1)
Describe how the RLE was developed and used in the SMA and
the location at which the RLE was applied (i.e., the control point
elevation)
(2)
The definition and justification of the response spectrum shape
used for the FA of SSCs, accident sequences, and the plant, if it
differed from the RLE
- 23 -
5.2
(3)
A summary of the plant system models, including event trees and
fault trees and how they were developed
(4)
A description of the methodologies used to quantify the seismic
margins of high confidence of low probability of failure (HCLPF)
capacities of SSCs, together with key assumptions. This should
include details of response analysis, generation of ISRS, and
other details
(5)
A detailed list of the SSC seismic margin values with reference to
the method of seismic qualification, the dominant failure modes,
and how the margin analysis is principally supported
(e.g., analysis, test data, experience data)
(6)
For each analyzed SSC, the parameter values defining the
seismic margin (e.g., the HCLPF capacity and any other
parameter values such as the median acceleration capacity and
the logarithmic standard deviation or “beta” values) and the
technical bases for the values
(7)
The general bases for screening SSCs, including screening levels
and lists of SSCs considered inherently rugged and a list of SSCs
considered as high capacity SSCs
(8)
Identification of the methods used to calculate sequence-level and
plant-level HCLPFs
(9)
Risk-significant sequences, dominant cut-sets, and associated
Booleans for both core damage and large early release
(10)
Sequence-level and plant level HCLPF capacities for both core
damage and large early release
(11)
A discussion of sensitivity to random failures and operator errors
(12)
A discussion of the treatment of uncertainties
(13)
A discussion of how the dominant sequences are identified
(14)
A description of the process used to ensure that the SMA is
technically adequate, including a description of the approach to
peer review, the dates and findings of peer reviews, and a
description of how peer review findings were closed out
(15)
A list of the identified plant-specific vulnerabilities and actions
planned or taken
Separate Reporting of HCLPF Capacities of Dominant Sequences for Core
Damage and for Large Early Release
Technical Issue:
See the issue discussed in Section 4.2.3. This is the documentation
requirement.
- 24 Staff Position:
5.3
When reporting sequence-level and plant-level HCLPF capacities, the
SMA analysis should separately report HCLPF capacities for the
core-damage endpoint and the large-release endpoint.
Separate Reporting of HLCPF Capacities of Sequences with and Sequences
without Non-Seismic Failures and Human Errors
Technical Issue:
See the issue discussed in Section 4.2.4. This is the documentation
requirement.
Staff Position:
When reporting sequence-level and plant-level HCLPF capacities, the
SMA analysis should separately report HCLPF capacities for sequences
with and sequences without non-seismic failures and human errors.
5.4
Information Retained for Audit
Technical Issue:
Some additional information, beyond that submitted to the NRC in
response to the 50.54(f) letter, should be retained for NRC audit. This is
for both reviewing the response to the 50.54(f) letter, and for any future
uses of the program analyses and results.
Staff Position:
The information retained for NRC audit should include (but is not limited
to):
•
•
•
•
applicable event trees and fault trees
current versions of the system notebooks (if applicable)
walkdown checklists and reports
evaluation results
In general, all documents essential for a practitioner in the field to
understand and trace what was done in the SMA should be retained. In
addition, the way in which the validity of these documents has been
ensured should be documented.
6.0
Staff Positions on Peer Review Attributes, Activities, and Documentation
Technical Issue:
The peer review is a key element of the SMA process that increases
confidence and assurance that the results of the assessment are reliable
and provide the information necessary for regulatory decisions.
Appropriate documentation of the peer review process is also important.
Staff Position:4
Peer review should include the following attributes:
•
4
The peer review process should be consistent with the comments
provided in the NRC staff letter for NEI 12-13, “External Hazards
The staff positions below are consistent with the peer review process set forth in the ASME/ANS PRA
standard, as endorsed in Regulatory Guide 1.200.
- 25 PRA Peer Review Process Guidelines,”5 as well as Regulatory
Guide 1.200.
•
An in-process peer review, as described in the draft ISG on
NEI 12-13 “External Hazards PRA Peer Review Process
Guidelines,” is preferred over a one-time late stage review, but the
latter is also acceptable.
•
Peer reviewers on various technical elements should have the
opportunity to interact with each other when performing the
reviews, and on critical items (e.g., results of screening or the SEL
development), the peer review should be conducted as a team.
•
Particular attention should be paid to justifications for use of
models or methods that are not consistent with current practice
(e.g., the use of the original structural models, the site response
assessment using limited data). In cases in which the SPID is
used, reviews should be based on the adherence to the approach
and intent of the SPID guidance and the adequacy of the model to
address the 50.54(f) letter.
•
The peer review process includes a review of the following SMA
activities:
o
selection of the SSCs included on the SEL
o
review a sample of the documentation from the seismic
walkdowns
o
seismic response analyses
o
seismic HCLPF capacity assessments for individual SSCs
o
sequence-level and plant-level HCLPF quantification
o
final report
The peer review team should be assembled based on the following
considerations:
5
•
The peer review team should have combined experience in the
areas of systems engineering, seismic capability engineering, and
seismic PRAs or seismic margin methodologies.
•
The reviewer(s) focusing on the seismic fragility work should have
successfully completed the Seismic Qualification Utility Group
(SQUG) Walkdown Screening and Seismic Evaluation Training
Course or equivalent, or shall have demonstrated experience in
seismic walkdowns.
The NRC letter with comments on NEI 12-13 can be found at ADAMS Accession No. ML12321A280.
- 26 •
One of the peer reviewers should be designated as the overall
team leader. The peer review team leader is responsible for the
entire peer review process, including completion of the final peer
review documentation. The team leader is expected to provide
oversight related to both the process and technical aspects of the
peer review. The team leader also should pay attention to
potential issues that could occur at the interface between various
activities.
•
Reviewers should be independent of those who are doing the
work.
The peer review process should be clearly documented in the report
submitted to the NRC. Documentation in the report should include
the following:
7.0
•
The names and qualifications of the team members.
•
A description of the peer review process.
•
A discussion of the key findings and a discussion on how the
findings were addressed.
•
Information on the disposition of comments.
•
The review of the final report.
•
The conclusions of the peer review.
•
The peer review report should be documented in a separate
report.
Citations
American Society of Mechanical Engineers/American Nuclear Society, “Standard for Level
1/Large Early Release Frequency Probabilistic Risk Assessment for Nuclear Power Plant
Applications,” Standard ASME/ANS RA-Sa-2009, 2009.
Electric Power Research Institute, “Seismic Margin Assessment of the Catawba Nuclear
Station,” EPRI Report NP-6359, Palo Alto, California, 1988.
Electric Power Research Institute, “A Methodology for Assessment of Nuclear Power Plant
Seismic Margin,” EPRI Report NP-6041-SL, Revision 1, Palo Alto, California, 1991.
Electric Power Research Institute, “Methodology for Developing Seismic Fragilities,” EPRI
Report TR-103959, Palo Alto, California, 1994.
Electric Power Research Institute, “Seismic Fragility Application Guide,” EPRI Report 1002988,
Final Report, Palo Alto, California, December 2002.
- 27 Electric Power Research Institute, “Seismic Probabilistic Risk Assessment Implementation
Guide,” EPRI Report 1002989, Palo Alto, California, 2003.
Electric Power Research Institute, “Seismic Fragility Application Guide,” EPRI 1019200,
Palo Alto, California, 2009.
Kennedy, R.P., “Overview of Methods for Seismic PRA and Margins Methods Including Recent
Innovations,” in Proceedings of the OECD/Nuclear Energy Agency Workshop on Seismic Risk,
Tokyo, Japan; August 10-12, 1999 (available from the OECD Nuclear Energy Agency, Le Seine
St.-Germain, 12 boulevard des Iles, F-92130 Issy-les-Moulineaux, France).
U.S. Nuclear Regulatory Commission, “Seismic Margin Review of Plant Hatch Unit 1: System
Analysis,” NUREG/CR-5632, 1990.
U.S. Nuclear Regulatory Commission, “PRA Procedures Guide: A Guide to the Performance of
Probabilistic Risk Assessments for Nuclear Power Plants,” NUREG/CR-2300, December 1982.
U.S. Nuclear Regulatory Commission, “An Approach to the Quantification of Seismic Margins in
Nuclear Power Plants,” NUREG/CR-4334, August 1985 (Agencywide Documents Access and
Management System (ADAMS) Accession No. ML090500182).
U.S. Nuclear Regulatory Commission, “Recommendations to the Nuclear Regulatory
Commission on Trial Guidelines for Seismic Margin Reviews of Nuclear Power Plants,”
NUREG/CR-4482, 1986 (ADAMS Accession No. ML12069A017).
U.S. Nuclear Regulatory Commission, “Seismic Margin Review of the Maine Yankee Atomic
Power Station,” NUREG/CR-4826, in 3 vols. 1987.
U.S. Nuclear Regulatory Commission, “An Approach to the Quantification of Seismic Margins in
Nuclear Power Plants: The Importance of BWR Plant Systems and Functions to Seismic
Margins,” NUREG/CR-5076, 1988.
U.S. Nuclear Regulatory Commission, “Individual Plant Examination of External Events (IPEEE)
for Severe Accident Vulnerabilities- 10 CFR 50.54(f),” Generic Letter 88-20, Supplement No. 4,
April 1991. Available at http://www.nrc.gov/reading-rm/doc-collections/gen-comm/genletters/1988/gl88020s4.html.
U.S. Nuclear Regulatory Commission, “Procedural and Submittal Guidance for the Individual
Plant Examination of External Events (IPEEE) for Severe Accident Vulnerabilities,”
NUREG-1407, June 1991 (ADAMS Accession No. ML063550238).
U.S. Nuclear Regulatory Commission, “A Performance-Based Approach to Define the
Site-Specific Earthquake Ground Motion,” Regulatory Guide 1.208, 2007 (ADAMS Accession
No. ML070310619).
U.S. Nuclear Regulatory Commission, “An Approach for Determining the Technical Adequacy of
Probabilistic Risk Assessment Results for Risk-Informed Activities,” Regulatory Guide 1.200
Revision 2, March 2009 (ADAMS Accession No. ML090410014).
- 28 U.S. Nuclear Regulatory Commission, “Interim Staff Guidance on Implementation of a
Probabilistic Risk Assessment-Based Seismic Margin Analysis for New Reactors,” Interim Staff
Guidance DC/COL-ISG-020, March 15, 2010 (ADAMS Accession No. ML100491233).
U.S. Nuclear Regulatory Commission, “Recommendations for Enhancing Reactor Safety in the
21st Century, the Near-Term Task Force Review of Insights from the Fukushima Dai-ichi
Accident,” Commission Paper SECY-11-0093, July 12, 2011 (ADAMS Accession
No. ML11186A950).
U.S. Nuclear Regulatory Commission Letter to All Power Reactor Licensees et al., “Request for
Information Pursuant to Title 10 of the Code of Federal Regulations 50.54(f) Regarding
Recommendation 2.1, 2.3, and 9.3, of the Near-Term Task Force Review of Insights from the
Fukushima Dai-ichi Accident,” March 12, 2012 (ADAMS Accession No. ML12053A340).
U.S. Nuclear Regulatory Commission, “Nuclear Regulatory Commission (NRC) Comments on
NEI 12-13, ‘External Hazards PRA Peer Review Process Guidelines’” NRC Staff letter, dated
November 16, 2012, ADAMS Accession No. ML12321A280.
APPENDIX A
POSITIONS FROM THE SCREENING,
PRIORITIZATION AND IMPLEMENTATION DOCUMENT
Appendix A
APPENDIX A: POSITIONS FROM THE SCREENING,
PRIORITIZATION AND IMPLEMENTATION DOCUMENT
This interim staff guidance (ISG) incorporates the guidance from the screening, prioritization,
and implementation document (SPID) on six technical elements. The SPID guidance is further
discussed below for three technical topics incorporated into this ISG. In addition, guidance on
the topics of potentially high frequency sensitive equipment and SSC screening methods and
levels is provided in Sections 4.2.5 and 4.5.2, respectively.
The topics covered below are:
•
•
•
use of existing structural models
scaling of in-structure response spectra
use of fixed-based structural models for soft rock
Use of Existing Structural Models
The development of in-structure response spectra (ISRS) is required for both seismic margin
assessment (SMA) and seismic probabilistic risk assessment (SPRA). Using existing structural
models, where appropriate, will facilitate the timely completion of the SPRA/SMA effort within
the desired accuracy required as part of the response to the 50.54(f) letter. Industry and the
NRC have agreed that in some cases existing structural models (i.e., those used for design
basis or in USI-A-46/IPEEE studies) could be used in structural dynamic analyses performed to
support SPRAs or SMAs required as part of the response to the 50.54(f) letter. However, not all
models have the appropriate attributes, or are of sufficient complexity, to adequately capture the
structural response.
Therefore, an experienced structural engineer(s) (and a peer reviewer) must perform a review of
each of the existing models to determine the adequacy of the models for dynamic analysis for
application in risk assessments conducted for addressing Recommendation 2.1. If necessary,
the existing structural models can be enhanced to bring it to an acceptable level. Industry and
the NRC agreed to a set of criteria (provided below) to determine if an existing model can be
used directly, or if it must be enhanced or replaced.
Each licensee will need to demonstrate and document that its models are adequate for
addressing the 50.54(f) letter and meet the criteria in the SPID (if a new model is not developed
using current practice). Any potential structural issues including the adequacy of the model
should be addressed and justified in the documentation. The model itself, the modeling
process, and the documentation should be subject to peer review, which also will be
documented.
The criteria against which structural engineer(s) and peer reviewer(s) should review the existing
models are listed below.
(1)
The structural models should be capable of capturing the overall structural responses for
both the horizontal and vertical components of ground motion.
(2)
If there is significant coupling between the horizontal and the vertical responses, one
combined structural model should be used for analyzing all three directions of the
A-1
Appendix A
earthquake. See ASCE 4-98 Section 3.1.1.1 “Models for Horizontal and Vertical
Motions.”
(3)
Structural mass (total structural, major components, and appropriate portion of live load)
should be lumped so that the total mass, as well as the center of gravity, is preserved.
Rotational inertia should be included if it affects the response in the frequency range of
interest. See ASCE 4-98 Section 3.1.4.1, “Discretization of Mass” Part (b) 1.
(4)
The number of nodal or dynamic degrees of freedom should be sufficient to represent
significant structural modes. All modes up to structural natural frequencies of about
20 Hz in all directions should be included (vertical floor slab flexibility generally will not
be considered because it is expected to have frequencies above 15 Hz). This will
ensure that the seismic responses and ISRS developed in the 1 to 10 Hz frequency
range are reasonably accurate. See ASCE 4-98 Section 3.1.4.1, “Discretization of
Mass,” Part (b) 2.
(5)
Torsional effects resulting from eccentricities between the center of mass and the center
of rigidity should be included. The center of mass and the center of rigidity may not be
coincident at all levels, and the torsional rigidity should be computed. See ASCE 4-98
Section 3.1.8.1.3, “Requirements for Lumped-mass Stick Models [LMSM],” Parts (b)
and (c). Alternatively, a multiple LMSM may be used if the stiffness elements are
located at the centers of rigidity of the respective groups of element and the individual
models are properly interconnected.
(6)
The analyst should determine if one stick model sufficiently represents the structure. For
example, two stick models could be appropriate for the analysis of internal and external
structures of the containment founded on a common mat.
(7)
The structural analyst should review whether in-plane floor flexibility (and subsequent
amplified seismic response) has been captured appropriately for developing accurate
seismic response up to the 15 Hz frequency. Experience has shown that for nuclear
structures with floor diaphragms that have length to width ratios greater than about 1.5,
the in-plane diaphragm flexibility may need to be included in the LMSM. As with all
these recommendations, alternate approaches can be used when justified.
Scaling of In-Structure Response Spectra
The NRC staff and industry have agreed that scaling approaches can be used in developing
ISRS for those cases in which the new site-specific hazard spectral shape is approximately
similar to the spectral shape previously used to generate the ISRS. The use of scaling will
reduce the effort involved in performing detailed soil structure interaction analyses for the new
hazard response spectrum, facilitating the timely completion of the SMA and SPRA efforts for
those plants that are screened-in.
Guidance on scaling is provided in industry documents such as EPRI report NP-6041-SL
Revision 1 and EPRI report 103959. Scaling of ISRS is an accepted approach that has been
used in previous SMA and SPRAs, including the recent Surry pilot SPRA. An example
approach for the scaling of “non-similar” shapes conducted for the Surry pilot SPRA project will
be described in the SPID. Unfortunately, hard and fast rules as to what is “close enough” are
hard to come by. The NRC staff and industry have agreed that it is not possible at this time to
provide more than general guidance (with examples of what clearly is and what clearly is not
A-2
Appendix A
acceptable). The SPID provides examples of pairs of spectra that are and are not sufficiently
similar to justify the use of scaling.
The acceptability of scaling of responses will be based on the following:
•
previously developed ISRS
•
shapes of the previous input response spectrum or review level earthquake (RLE)
ground motion
•
shapes of the new RLE ground motion, and the structural natural frequencies, mode
shapes, and participation factors
Licensees will need to demonstrate and document that ISRS scaling is appropriate for the site
and each applicable structure in their submission to the NRC. Any potential structural issues
with the use of scaling should be addressed and justified in the documentation. The use of
scaling and the documentation should be subject to peer review, which also will be documented.
Scaling of rock or soil sites, where the shape of the new hazard spectrum is not highly similar to
the previous spectrum, is not recommended without justification that demonstrates the validity of
the scaling approach.
Use of a Fixed-Base Structural Model for Soft Rock Conditions
Some existing structural models and ISRS were developed using an earlier definition of rock
(Vs ≥ 5,000 ft/sec), which is now considered a soft rock. Based on recent analyses discussed in
the SPID, for purposes of the seismic reevaluation process, this earlier definition of rock can be
used for the development of the ISRS because past analyses and experience has shown that
the amplified response spectra in the 1-10 Hz rock-founded structures fare approximately the
same from a fixed based model and a model that uses soil-structure interaction analysis.
Therefore, for most rock-founded structures, it is conservative to use fixed base dynamic
analyses even when the shear wave velocities are not as high as the current definition of rock
(of Vs ≥ 9200 ft/sec). The exception may be for structures with high frequency first modes. It is
also acceptable for some nuclear power plants located on rock with a Vs ≥ 3,500 ft/sec to use a
fixed base model, depending on properties of the site and the structure.
For the purpose of addressing the seismic reevaluation process, fixed based models can be
used in the dynamic analyses of rock-founded structures with Vs ≥ 3,500 ft/sec, provided that
justification is provided. Sites founded on rock with Vs between 3,500 and 5,000 ft/sec should
pay special attention to the potential issues discussed in the SPID.
Additional Reference
American Society of Civil Engineers 4-98, “Seismic Analysis of Safety-Related Nuclear
Structures, American Society of Civil Engineers,” Standard ASCE 4-98, 1998.
A-3
| File Type | application/pdf |
| File Modified | 0000-00-00 |
| File Created | 0000-00-00 |