SPID Seismic Evaluation Guidance

Seismic Evaluation
Guidance
Screening, Prioritization and
Implementation Details (SPID) for the
Resolution of Fukushima Near-Term Task
Force Recommendation 2.1: Seismic
NOT

1025287
Report, November 2012
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF
WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC.
(EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S)
BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
(A)
MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)
WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B)
ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER
(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY ITS
TRADE NAME, TRADEMARK, MANUFACTURER, OR OTHERWISE, DOES NOT NECESSARILY
CONSTITUTE OR IMPLY ITS ENDORSEMENT, RECOMMENDATION, OR FAVORING BY EPRI.
THE TECHNICAL CONTENTS OF THIS DOCUMENT WERE NOT PREPARED IN ACCORDANCE WITH
THE EPRI NUCLEAR QUALITY ASSURANCE PROGRAM MANUAL THAT FULFILLS THE
REQUIREMENTS OF 10 CFR 50, APPENDIX B AND 10 CFR PART 21, ANSI N45.2-1977
AND/OR THE INTENT OF ISO-9001 (1994). USE OF THE CONTENTS OF THIS DOCUMENT IN
NUCLEAR SAFETY OR NUCLEAR QUALITY APPLICATIONS REQUIRES ADDITIONAL ACTIONS BY
USER PURSUANT TO THEIR INTERNAL PROCEDURES.
NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or
e-mail [email protected].
Electric Power Research Institute, EPRI, and TOGETHERSHAPING THE FUTURE OF
ELECTRICITY are registered service marks of the Electric Power Research Institute, Inc.
Copyright 2012 Electric Power Research Institute, Inc. All rights reserved.
This report describes research sponsored by EPRI. The following

organizations prepared this report:
Acknowledgments
Simpson Gumpertz & Heger Inc.

4000 MacArthur Blvd. Suite 710
Newport Beach, CA 92660
Principal Investigator
G. Hardy
K. Merz
RPK Structural Mechanics Consulting, Inc.
28625 Mountain Meadow Road
Escondido, CA 92026
R. Kennedy
Dominion Resources, Inc.,
5000 Dominion Blvd
Glen Allen, VA 23060
D. Bhargava
Nuclear Energy Institute
1776 I Street Northwest
Washington, DC
K. Keithline
Lettis Consultants International, Inc.
4155 Darley Ave, Suite A
Boulder, CO 80305
R. McGuire

Seismic Evaluation Guidance:
Screening, Prioritization and
Implementation Details (SPID)
for the Resolution of Fukushima
Near-Term Task Force
Recommendation 2.1: Seismic

Pacific Engineering and Analysis

856 Seaview Drive
El Cerrito, CA 94530
W. Silva
R. Darragh
g iii h
Southern Nuclear Operating Company, Inc.

42 Inverness Center Parkway
Birmingham, AL 35242
D. Moore
South Carolina Electric & Gas Company
P.O. Box 88
Jenkinsville, SC 29065
R. Whorton
ERIN Engineering & Research, Inc.
2001 N. Main Street, Suite 510
Walnut Creek, CA 94596
D. True
Electric Power Research Institute
3420 Hillview Avenue
Palo Alto, CA 94304-1338
Principal Investigators
J. Hamel
R. Kassawara
S. Lewis
J. Richards
EPRI gratefully acknowledges the following individuals and their
companies for their support of this report.
Gregory Krueger, Exelon Corporation
Andrea Maioli, Westinghouse Electric Company
Caroline McAndrews, Southern California Edison
Vincent Anderson, ERIN Engineering & Research, Inc.
EPRI also gratefully acknowledges the support of the following
members of the NRC staff and contractors for their significant
efforts in the development of this report through public meetings and
public comments.
Nilesh Chokshi
Annie Kammerer
Robert Budnitz
Christopher Gratton
Clifford Munson
Jon Ake
M.K. Ravindra
Product
Description
Following the accident at the Fukushima Daiichi nuclear power

plant resulting from the March 11, 2011, Great Tohoku Earthquake
and subsequent tsunami, the NRC Commission established a Near
Term Task Force (NTTF) to conduct a systematic review of NRC
processes and regulations and to determine if the agency should
make additional improvements to its regulatory system. The NTTF
developed a set of recommendations intended to clarify and
strengthen the regulatory framework for protection against natural
phenomena. Subsequently, the NRC issued a 50.54(f) letter that
requests information to assure that these recommendations are
addressed by all U.S. nuclear power plants. This report provides
guidance for conducting seismic evaluations as requested in
Enclosure 1 of the 50.54(f) letter [1]. This 50.54(f) letter requests
that licensees and holders of construction permits under 10 CFR
Part 50 reevaluate the seismic hazards at their sites against presentday NRC requirements and guidance. Based upon this information,
the NRC staff will determine whether additional regulatory actions
are necessary.
Objectives
The objective of the work reported in this document is to provide
guidance on the performance of plant seismic evaluations, and in
particular those intended to satisfy the requirements of NTTF
Recommendation 2.1: Seismic.
Approach
The approach taken was to formulate guidance for the seismic
evaluations through a series of expert meetings, supplemented by
analytical research to evaluate selected criteria. Previous seismic
evaluations are described and applied, to the extent applicable.
Screening methods are described for evaluating newly calculated
seismic hazards against previous site-specific seismic evaluations, as
well as to determine the structures, systems, and components (SSCs)
that are appropriate to be modeled in a seismic probabilistic risk
assessment (SPRA).
A number of public meeting were also held with the NRC during
development of the guidance to discuss evaluation criteria and to
ensure the guidance met the requirements of NTTF
Recommendation 2.1: Seismic.
g v h
Results and Findings

This report outlines a process and provides guidance for investigating
the significance of new estimates of seismic hazard and, where
necessary, performing further seismic evaluations. This guidance is
pr
gn for use in responding
sp
ng to the U.S. Nuclear
primarilyy designed
Recommendation 2.1: Seismic evaluations. The guidance includes a
screening process for evaluating updated site-specific seismic hazard
and ground motion response spectrum (GMRS) estimates against
the plant safe shutdown earthquake (SSE) and High Confidence of
Low Probability of Failure (HCLPF) capacities. It also provides a
selected seismic risk evaluation criteria as well as spent fuel pool
evaluation criteria.
Applications, Value and Use
The guidance in this report is intended primarily for use by all U.S.
nuclear power plants to meet the requirements of NTTF
Recommendation 2.1: Seismic. The primary value in this guidance is
that it has been reviewed with the NRC and can be applied by all
plants to provide a uniform and acceptable industry response to the
NRC. Furthermore, the guidance related to seismic evaluations is of
value for any seismic risk assessment.
Keywords
Earthquakes
Fukushima
Seismic hazard
Fragilities
SPRA
g vi h
Table of
Contents
Section 1:
Purpose and Approach ............ 1-1
1.1
Background on Seismic Risk
Evaluations in the U.S. ................... 1-1
1.1.1
Individual Plant Examination
of External Events Seismic ............ 1-2
1.1.2
Generic Issue 199 .............. 1-3
1.2
NRC NTTF Recommendations ........ 1-4
1.3
Approach to Responding to
Information Request for NTTF
Recommendation 2.1 ........................ 1-5
Section 2:
Seismic Hazard Development ...... 2-1
2.1
Introduction and Background ..... 2-1
2.2
Seismic Source
Characterization .......................... 2-2
2.3
Ground Motion Attenuation ....... 2-3
2.4
Site Seismic Response ........... 2-4
2.4.1
Site Response for Sites with
Limited Data ............................ 2-5
2.4.2
Horizons and SSE Control
Point
2-5
2.5
Hazard Calculations and
Documentation ............................. 2-8
2.5.1
PSHA and Hazard Calculations ... 2-8
2.5.2
Seismic Hazard Data
Deliverables ............................ 2-9
2.5.3
Seismic Hazard Data at
Control Points and Base-Rock ............ 2-9
Section 3:
GMRS Comparisons and Screening
of Plants ........................... 3-1
3.1
Background on Screening ......... 3-1
3.2
SSE Screening Task (SSE-toGMRS Comparison) .......................... 3-1
3.2.1
Special Screening
Considerations .......................... 3-2
There are two special screening
considerations: ......................... 3-2
3.3
IPEEE Screening Task ............ 3-5
3.3.1
IPEEE Adequacy ................. 3-6
3.3.2
Development of HCLPF Spectrum .. 3-8
3.3.3
Comparison of IPEEE HCLPF
Spectrum to GMRS ........................ 3-8
3.4
Treatment of High-Frequency
Exceedances ............................... 3-9
3.4.1
Scope of High-Frequency
Sensitive Components ................... 3-10
3.4.2
Phase 1 Testing ............... 3-14
3.4.3
Phase 2 Testing ............... 3-19
Section 4:
Seismic Hazard and Screening
Report .............................. 4-1
Section 5:
Prioritization (Schedule) ....... 5-1
Section 6:
Seismic Risk Evaluation ......... 6-1
6.1
Background on SPRA and SMA ...... 6-1
6.1.1
SPRA Methods and Procedures .... 6-1
6.1.2
NRC SMA Methods and
Procedures .............................. 6-5
6.2
Criteria for Selection of Risk
Evaluation Method (SPRA vs. SMA) .......... 6-6
6.3
Key Elements of Seismic
Structural and SSI Response ............... 6-7
6.3.1
Structure Modeling ............. 6-7
6.3.2
Seismic Response Scaling ....... 6-9
6.3.3
Fixed-Based Analysis of
Structures for Sites Previously
Defined as Rock ...................... 6-11
6.4
Key Elements of
Fragility/Capacity for the Resolution
of NTTF Recommendation 2.1 ............. 6-11
6.4.1
Hybrid Approach for
Fragilities ............................ 6-11
6.4.2
High-Frequency Capacities ..... 6-13
6.4.3
Capacity-based Selection of
SSCs for Performing Fragility Analyses . 6-15
6.5
Key Elements of SPRA/SMA Scope
and Plant Logic Modeling ................. 6-18
6.5.1
Evaluation of LERF ............ 6-18
6.6
Comparison to ASME/ANS SPRA
Standard and RG1.200 ..................... 6-21
6.6.1
Background .................... 6-21
6.6.2
Comparison of 2.1 Seismic
Approach to the SPRA Standard .......... 6-22
6.7
Peer Review .................... 6-65
6.8
SPRA Documentation ............. 6-66
Section 7:
Spent Fuel Pool Integrity
Evaluation .......................... 7-1
7.1
Scope of the Seismic
Evaluation for the SFP .................... 7-1
7.2
Evaluation Process for the SFP .. 7-3
7.2.1
Evaluation of Penetrations
above Top of Fuel ......................... 7-4
7.2.2
Evaluation of Penetrations
below Top of Fuel ......................... 7-5
7.2.3
Evaluation of Potential for
Siphoning Inventory ....................... 7-6
g viii h
7.3
Guidance for Additional
Evaluations ............................... 7-7
7.3.1
Drain-down and Evaporative
Losses
7-7
7.3.2
Assessment of the Potential
for Sloshing .............................. 7-8
Section 8:
References ...................... 8-1
Appendix A: Control Point Discussion from

Standard Review Plan ................ A-1
Appendix B: Development of Site-Specific
Amplification Factors ............... B-1
B1.0
Introduction .................... B-1
B2.0
Description of Sites Requiring
Response Analysis and Basis for
Alternative Models ........................ B-2
B2.1
Background on the Treatment of
Uncertainties ............................. B-3
B3.0
Development of Base-Case
Profiles and Assessment of Epistemic
Uncertainty in Profiles and Dynamic
Material Properties ....................... B-4
B-3.1 Development Process for Base-Case
Shear-Wave Velocity Profiles .............. B-5
B-3.2 Capturing Epistemic Uncertainty in
Velocity Profiles ......................... B-6
B-3.2.1 Epistemic Uncertainty in Final
Hazard Calculations ....................... B-8
B-3.3 Nonlinear Dynamic Material
Properties B-9
B-3.4 Densities .......................... B-11
B-4.0 Representation of Aleatory
Variability in Site Response ............. B-11
B-4.1 Randomization of Shear-Wave
Velocities B-11
B-4.2 Aleatory Variability of Dynamic
Material Properties ...................... B-12
B-5.0 Development of Input Motions ....... B-12
B-5.1 Simple Seismological Model to
Develop Control Motions .................. B-13
B-5.1.1 Magnitude ........................ B-14
B-5.1.2 Attenuation (Q(f)) Model ......... B-14
B-5.1.3 Kappa ............................ B-15
B-5.1.3.1 Development of Base Case Kappa
Models
B-15
B-5.1.3.2 Representation of Epistemic
Uncertainty in Kappa ..................... B-17
B-5.1.4 Source Model ..................... B-17

B-5.1.4.1 Development of Input Motions ... B-18
B-6.0 Development of Probabilistic Hazard
Curves
B-19
B-7.0 Hazard-Consistent, StrainCompatible Material Properties (HCSCP) ... B-20
Appendix C: Sensitivity Studies to Develop
Criteria for Analyzing RockFounded Structures as Fixed-Base
Models .............................. C-1
C1.0
Containment Structure ........... C-1
C2.0
Main Steam Valve House
Structure C-3
Appendix D: Sensitivity of Computed Annual
Probability of Failure PF to
Assumed Logarithmic Standard
Deviation Used in Hybrid Method
with Capacities Defined by 1%
Failure Probability Capacity C1% .... D-1
D1.0
Introduction .................... D-1
D2.0
Simplified Seismic Risk
Equation
D-2
D3.0
Sensitivity of Failure
Probability PF to ........................ D-3
D4.0
References ...................... D-4
g x h
List of
Figures
Figure 1-1 Recommended Approach to Respond

to Information Request 2.1 ................ 1-7
Figure 2-1 Steps to Obtain Site-Specific Seismic Hazard ....... 2-1
Figure 2-2 Soil Site Example .......................... 2-7
Figure 2-3 Rock Site Example ......................... 2-7
Figure 3-1 Example Comparison of GMRS to SSE (5%
Damping) ................................... 3-2
Figure 3-2 Example Comparison of GMRS to SSE and LHT
(5% Damping) ................................ 3-3
Figure 3-3 Screening
Example Narrow Exceedances at 2 Hz
and 6 Hz (5% Damping) .......................... 3-5

Figure 3-4 Example Comparison of GMRS to IHS (5%
Damping) ................................... 3-9
Figure 3-5 Example High Frequency Ground Motion Response
Spectrum ................................... 3-15
Figure 3-6 Random Multi-Frequency Test Input Motions ..... 3-17
Figure 3-7 Filtered Random Multi-Frequency Test Input
Motions .................................... 3-17
Figure 6-1 Example Seismic Hazard Curve ................ 6-2
Figure 6-2 Example Seismic Fragility Curve................ 6-3
Figure 6-3 Overview of the SPRA Methodology ............. 6-4
Figure 6-4 Example for Selection of SPRA vs. SMA .......... 6-7
Figure 6-5 Example of Ground Response Spectra that are
Similar ..................................... 6-10
Figure 6-6 Example of Ground Response Spectra that are not
Similar ..................................... 6-10
Figure 6-7 Potentially High-Frequency Sensitive Component

Screening ................................... 6-14
Figure 6-8 Capacity-based Criteria for Fragility Analyses ...... 6-16
Figure 7-1 Basic Process for Evaluation of Potential Failures for
SFP Penetrations ............................... 7-4
Figure 7-2 Basic Process for Evaluation of Potential Siphoning
of SFP Inventory ............................... 7-7
Figure B-1 Logic tree illustrating the process for capturing
uncertainty in the development of site-specific amplification
functions. ................................... B-29
Figure B-2 Template Shear Wave Velocity Profiles for Soils,
Soft Rock, and Firm Rock. Rock Profiles Include Shallow
Weathered Zone. Indicated velocities are for VS30. ........ B-30
Figure B-3 Illustration of how available information is used to
develop a mean base-case profile. ................... B-31
Figure B-4 Illustration of the range of velocity ............. B-32
Figure B-5 Method used to account for epistemic uncertainty in
site specific shear wave velocity profiling where limited
information is available. ......................... B-33
Figure B-6 Illustration of Upper Range and Lower Range BaseCase profiles (10th and 90th percentiles) developed to
represent the epistemic uncertainty in the Mean Base-Case
for firm rock conditions. ......................... B-34
Figure B-7 Illustration of the development of Upper Range and
Lower Range profiles to accommodate epistemic
uncertainties for the hypothetical example shown in Figure
B-3. ...................................... B-35
g xii h
Figure B-8 Generic G/GMAX and hysteretic damping curves

for cohesionless soil [18]. Note that damping will be limited
to 15% for this application. ....................... B-36
Figure B-9 Comparison of median amplification functions (5%damped PSa) derived using the EPRI (1993) [18] (see
Figure B-8) and Peninsular Range [40]) G/GMAX and
hysteretic damping curves. ........................ B-37
Figure B-10 Comparison of median amplification functions
(5%-damped PSa) derived using the EPRI (1993) [18] (see
Figure B-8) and Peninsular Range [40] G/GMAX and
hysteretic damping curves. ........................ B-38
Figure B-11 Generic G/GMAX and hysteretic damping curves
developed for firm rock in the EPRI (1993) study [12] (from
Dr. Robert Pyke). ............................. B-39
Figure B-12 Illustration of effect of various factors in the simple
seismological model on Fourier spectral shape. .......... B-40
Figure B-13 Illustration of effect of various factors in the simple
seismological model on response spectral shape. .......... B-41
Figure B-14 Comparison of amplification functions (5%-damped
PSa) computed for magnitudes of M 5.5, 6.5, and 7.5, using
the single-corner source model and the 400 m/sec VS30
stiff-soil template profile (Figure B-2) with the EPRI (1993)
[18] G/GMAX and hysteretic damping curves (Figure B-8). . B-42
Figure B-15 Comparison of amplification functions (5%-damped
PSa) computed for magnitudes of M 5.5, 6.5, and 7.5. ..... B-43
Figure B-16 Comparison of amplification functions (5% damped
PSa) computed using the Single- and Double-Corner source
models (Tables B-4 and B-6) for the 400 m/sec VS30 stiffsoil template profile (Figure B-2) with the EPRI (1993) [18]
G/GMAX and hysteretic damping curves (Figure B-8). .... B-44
Figure B-17 Comparison of amplification functions (5% damped

PSa) computed using the Single- and Double-Corner source
models (Tables B-4 and B-6) for the 400 m/sec VS30 stiffsoil template profile (Figure B-2) with the EPRI (1993) [18]
G/GMAX and hysteretic damping curves (Figure B-8). .... B-45
g xiv h
Table 3-1 EPRI 1015109 Potentially High Frequency Sensitive
List of
Tables
Items ...................................... 3-12

Table 3-2 AP1000 Potentially High Frequency Sensitive Items .. 3-13
Table 3-3 High Frequency Confirmation Component Types ... 3-14
Table 3-4 Phase 1 Test Samples....................... 3-14
Table 6-1 Partial List of SPRA Technical References ......... 6-5
Table 6-2 Recommended C, R, U, and C50%/C1% Values to
Use in Hybrid Method for Various Types of SSCs ........ 6-13
Table 6-3 Consideration of LERF Contributors in SPRA ..... 6-19
Table 6-4 Comparison of SPID Guidance to ASME/ANS
PRA Standard Supporting Requirements: Element SHA ... 6-24
PRA Standard Supporting Requirements: Element SFR .... 6-41
PRA Standard Supporting Requirements: Element SPR .... 6-54
List of
Acronyms
AEF
annual exceedance frequency
ANS
American Nuclear Society
AOV
air-operated valve
ASME
American Society of Mechanical Engineers
BWR
boiling water reactor
CAV
Cumulative Absolute Velocity
CDF
core damage frequency
CDFM
Conservative Deterministic Failure Margin
CENA
Central and Eastern North America
CEUS
Central and Eastern United States
COLA
Combined Operating License Application
COV
coefficient of variation
ESP
early site permit
FSAR
Final Safety Analysis Report
FRMF
Filtered Random Multi-Frequency
GI
Generic Issue
GIP
Generic Issues Program
GL
Generic Letter
GMPE
ground motion prediction equations
GMRS
ground motion response spectra
HCLPF
High Confidence of Low Probability of

Failure
HCSCP
Hazard-Consistent, Strain-Compatible
Material Properties
IEEE
Institute of Electrical and Electronics

Engineers
IHS
IPEEE HCLPF Spectrum
IPEEE
Individual Pant Examination of Individual

Events
ISG
Interim Safety Guide
ISRS
in-structure response spectra
LERF
large-safety release frequency
LHT
Low Hazard Threshold
LMSM
Lumped-mass stick models
MCC
motor control center
MCCB
Molded case circuit breakers
MOV
motor-operated valve
MSVH
Main Steam Valve House
NGA
Next Generation Attenuation
NPP
nuclear power plant
NRC
Nuclear Regulatory Commission
NTTF
Near Term Task Force
PGA
Peak Ground Acceleration
PSD
power spectral density
PSHA
probabilistic seismic hazard analysis
RLE
review level earthquake
RLME
Repeated Large Magnitude Earthquake
RMF
Random Multi-Frequency
RVT
Random Variation Theory
SAMG
Severe Accident Management Guidance
SCDF
seismic core damage frequency
SER
Safety Evaluation Report
SFP
spent fuel pool
SMA
seismic margin assessment
SPID
Screening, Prioritization, and

Implementation Details
SPRA
seismic probabilistic risk assessment
SQUG
Seismic Qualification Utilities Group
SSC
structures, systems, and components
SSE
safe-shutdown earthquake
SSHAC
Senior Seismic Hazard Analysis Committee
SSI
soil-structure interaction
TS
Time Series
UHRS
uniform hazard response spectrum
UHS
uniform hazard spectrum
USNRC
United States Nuclear Regulatory Commission
WNA
Western North America
WUS
Western United States
ZPA
zero period acceleration
g xviii h
Section 1: Purpose and

Approach
Following the accident at the Fukushima Daiichi nuclear power plant resulting
from the March 11, 2011 Great Tohoku Earthquake and subsequent tsunami,
the United States Nuclear Regulatory Commission (NRC) established the Near
Term Task Force (NTTF) in response to Commission direction. The NTTF
issued a report
ep that made a series of recommendations, some of which were to be
.
NRC issued a
50.54(f) letter that requests information to ensure that these recommendations
are addressed by all U.S. nuclear power plants (NPPs). The principal purpose of
this report is to provide guidance for responding to the request for information in
the 50.54(f) Letter, Enclosure 1, Recommendation 2.1: Seismic [1].
Although the guidance in this document is specifically directed at supporting
responses to the 50.54(f) letter, much of the guidance is appropriate for elements
of any seismic risk evaluation.
Section 1 of this report provides the background on two past seismic programs
(IPEEE and GI 199) that are particularly relevant to the 2.1 seismic assessment,
and summarizes both the NTTF recommendations and the technical approach
intended to support the response to the 2.1 seismic requests. Section 2
characterizes the seismic hazard elements of the response to the information
requests. Section 3 contains the ground motion response spectra (GMRS)
screening criteria associated with the resolution of the 2.1 seismic issue. Section 4
describes the elements of the recommended seismic hazard and screening report
to be submitted to the NRC. Section 5 describes the schedule prioritization for
completion of the seismic risk part of the 2.1 seismic program. Section 6 contains
the seismic risk evaluation methods for those plants required to conduct these
assessments. Finally, Section 7 documents an approach to the evaluation of the
seismic integrity of spent fuel pool integrity assessment.
1.1
Background on Seismic Risk Evaluations in the

U.S.
The risk posed by seismic events to plants operating in the United States was
previously assessed in the mid-1990s as part of the response to the request for an
Individual Plant Examination of External Events [2]. Further efforts to
understand seismic risks, particularly in light of increased estimates of seismic
hazard for some sites, led to the initiation of the Generic Issue 199 program [6].
g 1-1 h
An understanding of these two programs provides valuable background for the

discussion of seismic evaluations related to the current 50.54(f) letter.
1.1.1
Individual Plant Examination of External
Events Seismic
On June 28, 1991, the NRC issued Supplement 4 to Generic Letter (GL) 88-20,
"Individual Plant Examination of External Events (IPEEE) for Severe Accident
Vulnerabilities," [2]. This supplement to GL 88-20, referred to as the IPEEE
program, requested that each licensee identify and report to the NRC all plantspecific vulnerabilities to severe accidents caused by external events. The IPEEE
program included the following four supporting objectives:
1. Develop an appreciation of severe accident behavior.
2. Understand the most likely severe accident sequences that could occur at the
licensee's plant under full-power operating conditions.
3. Gain a qualitative understanding of the overall likelihood of core damage and
fission product releases.
4. Reduce, if necessary, the overall likelihood of core damage and radioactive
material releases by modifying, where appropriate, hardware and procedures
that would help prevent or mitigate severe accidents.
The following external events were to be considered in the IPEEE: seismic
events; internal fires; high winds; floods; and other external initiating events,
including accidents related to transportation or nearby facilities and plant-unique
hazards. The IPEEE program represents the last comprehensive seismic
risk/margin assessment for the U.S. fleet of NPPs and, as such, represents a
valuable resource for future seismic risk assessments.
EPRI conducted a research project to study the insights gained from the seismic
portion of the IPEEE program [3]. The scope of that EPRI study was to review
the vast amounts of both NRC and licensee documentation from the IPEEE
program and to summarize the resulting seismic IPEEE insights, including the
following:
Results from the Seismic IPEEE submittals
Plant improvements/modifications as a result of the Seismic IPEEE

Program
NRC responses to the Seismic IPEEE submittals
The seismic IPEEE review results for 110 units are summarized in the EPRI
Report [3]. Out of the 75 submittals reviewed, 28 submittals (41 units) used
seismic probabilistic risk assessment (PRA) methodology; 42 submittals (62
units) performed seismic margin assessments (SMAs) using a methodology
developed by EPRI [39]; three submittals (three units) performed SMAs using
an NRC developed methodology; and two submittals (four units) used sitespecific seismic programs for IPEEE submittals.
g 1-2 h
In addition to the EPRI review of seismic IPEEE insights, the NRC conducted
a parallel study. NUREG-1742, "Perspectives Gained from the Individual Plant
Examination of External Events (IPEEE) Program," issued April 2002 [4],
provides insights gained by the NRC from the seismic part of the IPEEE
program. Almost all licensees reported in their IPEEE submittals that no plant
vulnerabilities were identified with respect to seismic risk (the use of the term
"vulnerability" varied widely among the IPEEE submittals). However, most
licensees did report at least some seismic "anomalies," "outliers," or other
concerns. In the few submittals that did identify a seismic vulnerability, the
findings were comparable to those identified as outliers or anomalies in other
IPEEE submittals. Seventy percent of the plants proposed improvements as a
result of their seismic IPEEE analyses.
1.1.2
Generic Issue 199
In support of early site permits (ESPs) and combined operating license

applications (COLAs) for new reactors, the NRC staff reviewed updates to the
seismic source and ground motion models provided by applicants. These seismic
updates included new EPRI models to estimate earthquake ground motion and
updated models for earthquake sources in the Central and Eastern United States
(CEUS), such as those around Charleston, South Carolina, and New Madrid,
Missouri. These reviews produced some higher seismic hazard estimates than
previously calculated. This raised a concern about an increased likelihood of
exceeding the safe-shutdown earthquake (SSE) at operating facilities in the
CEUS. The NRC staff determined that, based on the evaluations of the IPEEE
program, seismic designs of operating plants in the CEUS do not pose an
imminent safety concern. At the same time, the NRC staff also recognized that
because the probability of exceeding the SSE at some currently operating sites in
the CEUS is higher than previously understood, further study was warranted. As
a result, the NRC staff concluded on May 26, 2005 [5] that the issue of increased
seismic hazard estimates in the CEUS should be examined under the Generic
Issues Program (GIP).
Generic Issue (GI)-199 was established on June 9, 2005 [6]. The initial screening
analysis for GI-199 suggested that estimates of the seismic hazard for some
currently operating plants in the CEUS have increased. The NRC staff
completed the initial screening analysis of GI-199 and held a public meeting in
February 2008 [7], concluding that GI-199 should proceed to the safety/risk
assessment stage of the GIP.
Subsequently, during the safety/risk assessment stage of the GIP, the NRC staff
reviewed and evaluated the new information received with the ESP/COLA
submittals, along with NRC staff estimates of seismic hazard produced using the
2008 U.S. Geological Survey seismic hazard model. The NRC staff compared
the new seismic hazard data with the earlier seismic hazard evaluations
conducted as part of the IPEEE program. NRC staff completed the safety/risk
assessment stage of GI-199 on September 2, 2010 [8], concluding that GI-199
should transition to the regulatory assessment stage of the GIP. The safety/risk
assessment also concluded that (1) an immediate safety concern did not exist, and
g 1-3 h
(2) adequate protection of public health and safety was not challenged as a result
of the new information. NRC staff presented this conclusion at a public meeting
held on October 6, 2010 (ADAMS Accession No. ML102950263). Information
Notice 2010-018, "Generic Issue 199, Implications of Updated Probabilistic
Seismic Hazard Estimates in Central and Eastern United States on Existing
Plants, dated September 2, 2010 [9], summarizes the results of the GI-199
safety/risk assessment.
For the GI-199 safety/risk assessment, the NRC staff evaluated the potential risk
significance of the updated seismic hazards using the risk information from the
IPEEE program to calculate new seismic core damage frequency (SCDF)
estimates. The changes in SCDF estimate calculated through the safety/risk
assessment performed for some plants lie in the range of 10-4 per year to 10-5 per
year, which meet the numerical risk criteria for an issue to continue to the
regulatory assessment stage of the GIP. However, as described in NUREG-1742
[4], there are limitations associated with utilizing the inherently qualitative
insights from the IPEEE submittals in a quantitative assessment. In particular,
the NRC stafff s assessment did not provide insight into which structures,
systems, and components (SSCs) are important to seismic risk. Such knowledge
is necessary for NRC staff to determine, in light of the new understanding of
seismic hazards, whether additional regulatory action is warranted. The GI 199
issue has been subsumed into Fukushima NTTF recommendation 2.1 as
described in subsequent sections.
1.2
NRC NTTF Recommendations
The NRC issued an information request on March 12, 2012 related to the
Fukushima NTTF recommendations 2.1, 2.3, and 9.3 [1]. The requested seismic
information associated with Recommendation 2.1 is stated to reflect:
Information related to the updated seismic hazards at operating NPPs
Information based on a seismic risk evaluation (SMA or seismic probabilistic

risk assessment (SPRA)), as applicable
Information that would be obtained from an evaluation of the spent fuel pool
(SFP)
The basic seismic information requested by the NRC is similar to that developed
for GI-199 as presented in the draft GL for GI-199 [10]. The NRC has
identified an acceptable process for responding to the 2.1 seismic requests, which
is documented in Attachment 1 to the March 12, 2012 10CFR 50.54(f) letter
[1]. The NRC asks each addressee to provide information about the current
hazard and potential risk posed by seismic events using a progressive
screening/evaluation approach. Depending on the comparison between the reevaluated seismic hazard and the current design basis, the result is either no
further risk evaluation or the performance of a seismic risk assessment. Risk
assessment approaches
acceptable
to the staff, depending on the new hazard
pp
pt
estimates,
-type of SMA that was described in
NUREG-1407 [11] for IPEEEs, with enhancements.
g 1-4 h
1.3 Approach to Responding to Information Request for

NTTF Recommendation 2.1
The approach described in this report has been developed by EPRI, working
with experts from within the nuclear industry, with the intent of identifying
reasonable measures that can be employed to reduce the resources that might be
required to complete an effective seismic evaluation. More specifically, the
approach was designed to constitute a specific path to developing a response to
qu for information made in connection with NTTF Recommendation
the request
acceptable approach for the seismic elements of Recommendation 2.1
(documented in Attachment 1 to Seismic Enclosure 1 of the March 12, 2012
Request for Information [1]). In general, the approach described in this report is
intended to conform to the structure and philosophy of the nine steps suggested
by the NRC and outlined in that attachment. Key elements of the approach are
designed to streamline several of these nine steps (summarized below) while still
yielding an appropriate characterization of the impact of any change in hazard for
the plant being evaluated. Figure 1-1 illustrates the process for employing this
approach; it is based on a progressive screening approach and is broken down
into four major task areas:
Seismic Hazard and Site Response Characterization
GMRS Comparisons and Plant Screening
Prioritization of Risk Assessments
Seismic Risk Evaluation
The following paragraphs provide a brief discussion about each individual step in
Figure 1-1. The subsequent sections of this guide contain the detailed
descriptions of the methods and the documentation associated with this
approach.
Step 1. Develop site-specific control point elevation hazard curves over a range
of spectral frequencies and annual exceedance frequencies determined from a
probabilistic seismic hazard analysis (PSHA).
Step 2. Provide the new seismic hazard curves, the GMRS, and the SSE in
graphical and tabular format. Provide soil profiles used in the site response
analysis, as well as the resulting soil amplification functions.
Step 3. Utilize a screening process to eliminate certain plants from further
review. If the SSE is greater than or equal to the GMRS at all frequencies
between 1 and 10 Hz, then addressees may terminate the evaluation (Step 4)
after providing a confirmation, if necessary, that SSCs which may be affected by
high-frequency ground motion, will maintain their functions important to safety.
A similar screening review based on the IPEEE High Confidence of Low
Probability of Failure (HCLPF) Spectrum comparison to the GMRS can also be
conducted. Diamonds 3a thru 3f outline the overall screening process, and
Section 3 provides additional guidance.
g 1-5 h
Step 4. This step demonstrates termination of the process for resolution of

NTTF Recommendation 2.1 for plants whose SSE is greater than the calculated
GMRS.
Step 5. Based on criteria described in Section 6.2, perform a SPRA (steps 6a
and 7a) or a SMA (steps 6b and 7b). Step 5 also describes the prioritization
process for determining completion schedules for the seismic risk assessments.
Step 6a. If a SPRA is performed, it needs to be technically adequate for
regulatory decision making and to include an evaluation of containment
performance and integrity. This guide is intended to provide an acceptable
approach for determining the technical adequacy of a SPRA used to respond to
this information request.
Step 6b. If a SMA is performed, it should use a composite spectrum review level
earthquake (RLE), defined as the maximum of the GMRS and SSE at each
spectral frequency. The SMA should also include an evaluation of containment
performance and integrity. The American Society of Mechanical
Engineers/American Nuclear Society (ASME/ANS) RA-Sa-2009 [12] provides
an acceptable approach for determining the technical adequacy of a SMA used to
respond to this information request. In addition, the NRC is generating an
Interim Safety Guide (ISG) on the NRC SMA approach that will be acceptable
for this 2.1 application [15].
Step 7a. Document and submit the results of the SPRA to the NRC for review.
The "Requested Information" Section in the main body of Enclosure 1 [1]
identifies the specific information that is requested. In addition, addressees are
requested to submit an evaluation of the SFP integrity.
Step 7b. Document and submit the results of the SMA to the NRC for review.
The "Requested Information" Section in the main body of Enclosure 1 [1]
identifies the specific information that is requested. In addition, addressees
should submit an evaluation of the SFP integrity.
Step 8. Submit plans for actions that evaluate seismic risk contributors. NRC
staff, EPRI, industry, and other stakeholders will continue to interact to develop
acceptance criteria in order to identify potential vulnerabilities.
Step 9. The information provided in Steps 6 through 8 will be evaluated in
Phase 2 to consider any additional regulatory actions.
g 1-6 h
Figure 1-1
Recommended Approach to Respond to Information Request 2.1
g 1-7 h
Section 2: Seismic Hazard

Development
2.1
Introduction and Background
Seismic hazard analysis and the calculation of up-to-date seismic response spectra
is the first step to informed evaluations on priorities to mitigate seismic risk. To
determine if a reevaluation of seismic risk for a nuclear power plant is
appropriate, the comparison
of the up-to-date
seismic response
spectra with the
approp
pa
p
po
sp
account for both relative and absolute differences between up-to-date
p
seismic
response spectra and the exi
the seismic design spectra.
The first major part of the Seismic Enclosure 1 of the March 12, 2012 Request
for Information [1] is to calculate seismic hazard at existing plant sites by first
calculating uniform hazard response spectra (UHRS), using up-to-date models
representing seismic sources, ground motion equations, and site amplification.
From the UHRS results, GMRS are calculated. Figure 2-1 depicts (for
illustrative purposes only) the three basic elements of the seismic hazard analysis
(seismic source characterization, ground motion attenuation, and site
amplification), which will be described in more detail in the sections below.
Figure 2-1
Steps to Obtain Site-Specific Seismic Hazard
g 2-1 h
2.2
Seismic Source Characterization
Seismic Sources for the CEUS For the region designated the CEUS (United
States east of the Rocky Mountains), a regional study was jointly conducted by
USNRC, EPRI, and DOE during the period 2009-2011 to develop a
comprehensive representation of seismic sources for nuclear plant seismic
evaluation purposes. The results were published in 2012 [14] and provide an
acceptable source characterization model to use for seismic hazard studies [23, p.
115]. This study was conducted as a Senior Seismic Hazard Analysis Committee
(SSHAC) Level 3 study [13, 23], meaning that a detailed step-by-step process
was used to evaluate data and interpretations on earthquake occurrences, their
potential locations and sizes, and the rates with which they might occur, and that
process was documented and reviewed in a structured way. This ensured that all
credible data and interpretations were appropriately considered. Specifically,
detailed workshops were held that addressed the fundamental technical bases
upon which models of seismic sources could be developed, and alternative
models, with their technical bases, were defined. This applied to the geometries
of seismic sources, as well as to the parameters of the sources (earthquake
magnitude
distributions, rates of activity,
and
gn
ty maximum magnitudes,
gn
parameters were quantitatively weighted to express the credibility of each
alternative. A Technical Integration team conducted these analyses and
documented the derivation of weights so that a logic-tree approach (alternatives
with weights) could be used to characterize the interpretations and their
uncertainties. This set of interpretations forms the basis for characterizing the
distribution of future earthquake occurrences in the CEUS. Because of the large
regional study area of the CEUS Seismic Source Characterization project,
detailed evaluations of geology, topography, and other data in the vicinity of
NPPs was not undertaken.
Seismic sources were defined in the CEUS Seismic Source Characterization
project in two categories. First were Repeated Large Magnitude Earthquake
(RLME) sources, which represent sources where there is evidence of repeated,
large-magnitude earthquakes. The two major RLME sources in the CEUS are
the New Madrid seismic zone and the Charleston seismic zone. However, the
CEUS Seismic Source Characterization project identified additional RLME
sources on the basis of paleo-earthquake and other evidence.
The second category of seismic sources were background sources, which are large
regions within which earthquakes are modeled as occurring according to an
exponential magnitude distribution but where specific faults or causative
structures have not been identified. Two sets of background sources were
identified based on alternative methods to estimate maximum magnitude, and
each set of background sources covers the entire CEUS (and surrounding
territory). An updated earthquake catalog was created and used to estimate rates
of activity within the sources, the rate of activity varying spatially to reflect the
historical occurrences of small and moderate earthquakes. Thus, for example,
sub-regions of the CEUS that have experienced relatively many historical
g 2-2 h
earthquakes would have a higher rate of activity than sub-regions that have
experienced relatively few historical earthquakes.
For site-specific licensing applications or site-specific safety decisions, these
seismic sources would be reviewed on a site-specific basis to determine if they
need to be updated. Such evaluations would be appropriate in a licensing
application, where focus could be made on site-specific applications. However,
for a screening-level study of multiple plants for the purpose of setting priorities,
the use of these seismic sources as published is appropriate.
In addition, for applications in a regional study, it is sufficient to include
background sources within 320 km (200 miles) of a site, and specifically to
include only parts of those background sources that lie within 320 km of the site.
This follows the guidance
in [18] regarding
g
ga
g examination of sources within the
sufficient to include the New Madrid, Charlevoix, and the Charleston seismic
zones if they lie within 1,000 km of a site. Beyond 1,000 km, ground motion
equations have not been well-studied, and such distant earthquakes do not
generally cause damage to modern engineered facilities. For other RLME sources
and sub-regions of background sources with higher rates of activity, it is
sufficient to include them in the analysis if they lie within 500 km of a site, based
on test hazard results published in the CEUS Seismic Source Characterization
project.
Seismic Sources for the WUS For Western United States (WUS) plants,
characterizing of seismic sources is much more site-specific. These sites are
Diablo Canyon and San Onofre in California, Palo Verde in Arizona, and
Columbia in Washington. For the California sites, local faults dominate the
seismic hazard; for the Columbia site, local faults, background sources, and
subduction zone earthquakes are a consideration. For the Arizona site,
background sources and distant faults (including the San Andreas Fault) are
important. The development of seismic sources should be made on a site-specific
basis for these four sites by conducting a SSHAC Level 3 study [13, 23].
2.3
Ground Motion Attenuation
Ground Motion Estimates for the CEUS In 2004, EPRI [16] published a set
of ground motion prediction equations (GMPEs) for the CEUS, which included
both aleatory and epistemic uncertainties. In 2006, EPRI [17] published an
updated set of aleatory uncertainties to use with the 2004 equations. These
GMPEs estimate the aleatory and epistemic uncertainty in ground motion for
the mid-continent region of the CEUS and for the Gulf of Mexico region.
Beginning in 2012, EPRI has been evaluating the 2004-2006 GMPEs in light of
new ground motion models published in the technical literature and in light of
recorded ground motion data obtained during earthquakes in the CEUS and
south-eastern Canada. The overall goals of the project are to determine (a) if the
2004-2006 GMPEs should be updated in light of the new models and data, and
(b) if so, how to quantitatively update those GMPEs so they reflect the new
g 2-3 h
information. A decision to update the 2004-2006 GMPEs was confirmed on

August 14, 2012, and the updated models are expected in mid-February 2013.
It is anticipated that, as in EPRI 2004-2006, multiple models with weights will
be determined for the 2013 updated GMPEs and for the aleatory uncertainties.
It is also anticipated that equations will be developed for the two regions (midcontinent and Gulf of Mexico). In cases where the travel path of seismic waves
between a potential earthquake source and a site is predominantly in one region,
equations for that region should be used. In cases where the travel path crosses
from one region to the other, with a substantial fraction of the total travel path of
seismic waves in each region, hazard calculations can be made using either the
more conservative equations, or using a weighted average of hazard results based
on the approximate fraction that seismic waves travel through each region.
Because the EPRI 2012 ground motion update project is proceeding with
updating the EPRI 2004-2006 GMPEs, those updated equations, if approved by
the NRC, should be used to calculate ground motions for seismic hazard
calculations for all CEUS sites for Step
Otherwise the EPRI 2004-2006 GMPEs should
be used.
Currently some CEUS NPPs are developing SPRAs. Consistent with the current
SPRA standard requirement of using the most recent seismic hazard
information, they are using the EPRI 2004-2006 ground motion attenuation
model with the CEUS Seismic Source Characterization model for the seismic
hazard portion of their SPRAs. These CEUS NPPs should, in Step 7a, address
the effect of the new site hazard based on the updated EPRI 2004-2006
GMPEs.
Ground Motion Estimates for the WUS In the WUS, earthquake ground
motions can be estimated using recorded motions, and the seismic hazard is often
p
rg earthquake
hq
dominated by the possible
occurrence of a moderate-to-large
at close
years by the NGA-2 equations. Nuclear plant sites in the WUS should perform a
SSHAC Level 3 study [13, 23] in order to make site-specific decisions on which
equations are appropriate for their sites or to develop site-specific relationships.
2.4
Site Seismic Response
Every site that does not consist of hard rock should conduct an evaluation of the
site amplification that will occur as a result of bedrock ground motions traveling
upward through the soil/rock column to the surface. Critical parameters that
determine which frequencies of ground motion might experience significant
amplification (or de-amplification) are the layering of soil and/or soft rock, the
thicknesses of these layers, the initial shear modulus and damping of these layers,
their densities, and the degree to which the shear modulus and damping change
with increasing ground motion. The methods to calculate possible site
amplification are well-established, but at some sites the characterization of the
g 2-4 h
profile and layering is limited. For these sites, analyses must be conducted, as
described below, that account for uncertainties in soils and layer properties, and
this often results in significant uncertainties in site amplification. This Section
also provides a method for defining the elevation(s) for the SSE to GMRS
comparison for use in the 2.1 seismic screening.
2.4.1
Site Response for Sites with Limited Data
Many sites, particularly those licensed in the early 1970s, do not have detailed,
measured soil and soft-rock parameters to extensive depths. These sites will be
handled using the following guidelines (see Appendix B for a more detailed
discussion).
Shear-wave Velocity (Vs) For soil sites where Vs is estimated from
compression-wave measurements, or was measured only at shallow depths,
template profiles will be used based on experience with other, well-documented
sites. The template profiles will be adjusted and/or truncated to be consistent
with measured or estimated Vs in the upper 30 m of soil, called Vs30, to obtain a
reasonable profile to use for analysis that includes the potential effects on ground
motion of soils at large depths.
For firm rock sites (typically underlain by sedimentary rocks) that have little
measured Vs data, a Vs profile will be adopted that is consistent with shallow
estimates or measurements and that increases with depth using a gradient typical
of sedimentary rocks. A consistent gradient has been documented for
sedimentary rock sites in various locations around the world, and a profile
developed in this way will give reasonable results for the potential effects on
ground motion of sedimentary rock at large depths.
For sites with limited, or indirect data on Vs, multiple profiles or base cases
should be developed to account for the epistemic uncertainty. Typically three
base cases should be developed. To account for the variability in Vs over the scale
of the footprint of a NPP, which is treated as an aleatory uncertainty,
randomization about the base cases should be implemented. Additional
discussion regarding the methodology to incorporate the various types of
uncertainty is provided in Appendix B.
Dynamic Soil and Soft-rock Properties Other soil and soft-rock properties such
as dynamic moduli, hysteretic damping, and kappa (a measure of inherent near
surface site damping) will be adopted using published models. The same will be
done for soil and soft-rock densities, if they have not been measured and
reported.
2.4.2
Horizons and SSE Control Point
This Section provides a method for defining the elevation(s) for the SSE to
GMRS comparison for use in the 2.1 seismic screening. The SSE to GMRS
comparison for 2.1 screening per the 50.54(f) letter are recommended to be
applied using the licensing basis definition of SSE control point. The SSE is part
g 2-5 h
of the plant licensing basis which is typically documented in the FSAR. Three
specific elements are required to fully characterize the SSE:
Peak Ground Acceleration
Response Spectral Shape
Control Point where the SSE is defined
The first two elements of the SSE characterization are normally available in the
part of the Final Safety Analysis Report (FSAR) that describes the site seismicity
(typically Section 2.5). The control point for the SSE is not always specifically
defined in the FSAR and, as such, guidance is required to ensure that a
consistent set of comparisons are made. Most plants have a single SSE, but
several plants have two SSEs identified in their licensing basis (e.g., one at rock
and one at top of a soil layer).
For purposes of the SSE-to-GMRS comparisons as part of the 50.54(f) 2.1
seismic evaluations, the following criteria are recommended to establish a logical
comparison location:
1. If the SSE control point(s) is defined in the FSAR, use as defined.
2. If the SSE control point is not defined in the FSAR then the following
criteria should be used:
a. For sites classified as soil sites with generally uniform, horizontally
layered stratigraphy and where the key structures are soil-founded
(Figure 2-2), the control point is defined as the highest point in the
material where a safety-related structure is founded, regardless of the
shear wave velocity.
b. For sites classified as a rock site or where the key safety-related structures
are rock-founded (Figure 2-3), then the control point is located at the
top of the rock.
c. The SSE control point definition is applied to the main power block area
at a site even where soil/rock horizons could vary for some smaller
structures located away from the main power block (e.g., an intake
structure located away from the main power block area where the
soil/rock horizons are different).
g 2-6 h
Figure 2-2
Soil Site Example
Figure 2-3
Rock Site Example
The basis for the selected control point elevation should be described in the
submittal to the NRC. Deviations from the recommendations described above
should also be documented.
g 2-7 h
2.5
2.5.1
Hazard Calculations and Documentation

PSHA and Hazard Calculations
The PSHA will proceed with (1) the CEUS Seismic Source Characterization
models [14] or a regional WUS seismic source characterization (Section 2.2
above), with (2) GMPEs for the CEUS or the WUS (Section 2.3 above), and
with (3) a site seismic response analysis (quantified as described in Section 2.4
and Appendix B). Several assumptions are appropriate regarding the PSHA
calculations as follows:
For CEUS sites, seismic sources should be included for the range of distances
indicated in Section 2.2. For WUS sites, the Technical Integration team for the
SSHAC Level 3 study with input from the Participatory Peer Review Panel
should determine which seismic sources should be included in the PSHA.
As indicated in Section 2.3, for the CEUS the updated EPRI GMPEs should be
used for purposes of the 50.54(f) 2.1 seismic evaluations, if approved by the
NRC; otherwise, the EPRI 2004-2006 ground motion models [16, 17] should be
used. In addition, estimates of ground motion for source-site configurations with
seismic wave travel paths across both the mid-continent and Gulf of Mexico
regions should be handled as described in Section 2.3. For the WUS, a SSHAC
Level 3 study should be performed to select or develop appropriate GMPEs.
For the purposes of responding to the Seismic Enclosure 1 of the March 12,
2012 Request for Information [1], updates to seismic sources to account for
historical seismicity since 2008 (the last year of the earthquake catalog in the
CEUS Seismic Source Characterization study) are not required. Similarly,
updates to seismic sources to account for more recent earthquakes are not
necessary.
The CAV (Cumulative Absolute Velocity) filter developed by EPRI [19] may be
applied to account for the damageability of ground motions from small
magnitude earthquakes. However, if the CAV filter is applied, the lower-bound
magnitude for the PSHA should be set at M 4.0, and the CAV model should
not be applied for M greater than 5.5 (see Attachment 1 to Seismic Enclosure 1
of Reference [1]). In place of the CAV filter a minimum magnitude of M 5.0
may be used.
Site amplification factors should be calculated as described in Section 2.4. As
discussed in that section, multiple models of site amplification factors (and
associated uncertainties) should be developed, indicating the log-mean and logstandard deviation of control-point motion divided by input rock motion, for
various spectral frequencies. For input to site hazard calculations, these multiple
models should be combined, with weights, to derive the overall log-mean and
log-standard deviations of site amplification for each spectral frequency, as
described in Appendix B. The soil uncertainties should be incorporated into the
seismic hazard calculations using a formulation similar to Eq. (6-5) in [24],
wherein the site amplifications (with uncertainties) are incorporated into the
g 2-8 h
hazard integral to estimate the distribution of site amplitudes given earthquake

magnitude and distance. The implementation should estimate the distribution of
rock amplitude as a function of M and R, and the site amplification (given the
rock amplitude) for the value of M at which site amplifications were calculated.
This is sufficiently accurate since site amplifications are not highly dependent on
M and R.
The control-point elevation seismic hazard curves should be used to calculate a
GMRS for the site, using the method of [21]. The GMRS depends, in this
calculation, on the 10-4 and 10-5 spectral accelerations at each spectral frequency.
The control point should be defined at the same elevation as the design basis
SSE. Given that the site amplification factors are calculated assuming freesurface conditions above the control point, the GMRS will be consistent with
that assumption.
2.5.2
Seismic Hazard Data Deliverables
Soil Profile and Properties A description of the development of the base case
profile as it relates to the local geology should be described. In addition, for each
base case, the soil profile used to calculate site amplification factors should be
described, including layer boundaries, properties (Vs and density), modulus and
damping curves used for each layer, and uncertainties in these properties.
Site Amplification Factors Site amplification factors should be documented as
log-mean amplification factors and log-standard deviations of amplification
factors as a function of input rock acceleration, for the spectral frequencies at
which GMPEs are defined.
2.5.3
Base-Rock
Seismic Hazard Data at Control Points and
Hazard Data at Control Point Seismic hazard curves should be documented for
the control-point elevation corresponding to the mean hazard and common
fractiles. These curves should represent seismic hazard at the spectral frequencies
for which GMPEs are available. The control-point elevation hazard curves
should be represented for annual exceedance frequencies from 10-3 to 10-7.
Hazard curves should be provided in graphical and tabular format along with the
site response amplification function, SSE and GMRS.
The majority of the discussion in Sections 2.4 and 2.5 is based on using site
amplification with Method 3 from NUREG/CR-6728 [24] as described in
Appendix B. For plants using Method 2 for site amplification in accordance
with NUREG/CR-6728, the hard rock seismic hazard curves and the site
amplification factors to the control-point elevation should be reported in addition
to the control-point elevation hazards noted above.
g 2-9 h
Section 3: GMRS Comparisons

and Screening of Plants
3.1
Background on Screening
Following completion of the updated seismic hazard as described in Section 2, a

screening process is needed to determine which plants are required to perform
new seismic risk evaluations. The horizontal GMRS calculation
discussed/defined in Section 2 is being used to characterize the amplitude of the
new seismic hazard at each NPP site, as defined by the NRC [1]. The GMRS
should be compared to the horizontal 5% damped SSE as shown in Diamonds 3a
and 3b of Figure 1-1. If the SSE is exceeded, then licensees may have the option
to perform the screening of the GMRS to the IPEEE HCLPF spectrum (IHS).
The IHS is the response spectrum corresponding to the HCLPF level
documented from the seismic IPEEE program, as shown in Diamonds 3c
through 3e in Figure 1-1. The use of the IHS for screening is contingent upon
satisfying specific adequacy criteria, as described in Section 3.3. This screening
process, along with examples, is described in more detail in the Sections below.
3.2
SSE Screening Task (SSE-to-GMRS Comparison)
The SSE is the plant licensing basis earthquake and is uniquely defined for each
NPP site. The SSE consists of:
A PGA value which anchors the response spectra at high frequencies

(typically 33 Hz for the existing fleet of NPPs),
A response spectrum shape which depicts the amplified response at all

frequencies below the PGA (typically plotted at 5% damping), and
The control point applicable to the SSE (described in Section 2 of this

report). It is essential to ensure that the control point for both the SSE and
for the GMRS is the same.
The first step in the SSE screening process is to compare the SSE to the GMRS
in the 1 to 10 Hz part of the response spectrum (see Diamond 3a in Figure 1-1).
If the SSE exceeds the GMRS in the 1 to 10 Hz region, then a check of the
greater than 10 Hz part of the spectrum is performed as shown in Diamond 3b.
If the SSE exceeds the GMRS in the greater than 10 Hz region, then no further
action is required for NTTF Recommendation 2.1 seismic review (Box 4 in
Figure 1-1). If there are exceedances in the greater than 10 Hz region, then a
g 3-1 h
high-frequency confirmation should be performed (Box 3f in Figure 1-1) as

described in Section 3.4.
An example comparison of an SSE with a GMRS is shown in Figure 3-1. In this
example, only a high frequency confirmation is needed.
Figure 3-1
Example Comparison of GMRS to SSE (5% Damping)
If the initial review of the SSE to GMRS (Diamond 3a in Figure 1-1) does not
demonstrate that the SSE envelops the GMRS in the 1 to 10 Hz region, then,
depending upon the nature of the exceedance, the licensees have the option of:
1) Conducting a screening evaluation for narrow band exceedances as described
in Section 3.2.1, or
2) Conducting a screening evaluation using the IPEEE HCLPF capacity as
described in Section 3.3, or
3) Bypassing the screening evaluations and performing the seismic risk
evaluation using either an SPRA or SMA approach, as appropriate, as
described in Section 6 of this report.
3.2.1
Special Screening Considerations
There are two special screening considerations:
GMRS Comparisons and Screening of Plants at Low Seismic Hazard Sites

g 3-2 h
Narrow Band Exceedances in the 1 to 10 Hz Range
3.2.1.1
GMRS Comparisons and Screening of Plants at
Low Seismic Hazard Sites
A low seismic hazard site is defined herein to be a site where the GMRS peak
5% damped spectral acceleration (SAp) at frequencies between 1 and 10 Hz do
not exceed 0.4g, which is shown in Figure 3-2 as the Low Hazard Threshold
(LHT). Because of the low likelihood of any seismically designed SSC being
damaged by ground motion with an SAp less than this LHT, the following relief
from having to perform a full SMA or SPRA is considered to be warranted for
plants at sites where the GMRS is less than this LHT in the 1 to 10 Hz range.
Figure 3-2
Example Comparison of GMRS to SSE and LHT (5% Damping)
Figure 3-2 shows an example where the SSE spectral accelerations exceed the
GMRS spectral accelerations at frequencies below 10 Hz except for low
frequencies. Because the SSE response spectral accelerations reduce rapidly as
frequencies reduce below 2.5 Hz, the situation shown in Figure 3-2 can occur at
low seismic hazard sites. For most SSCs, such exceedance below 2.5 Hz is nonconsequential because the fundamental frequency of these SSCs exceeds 2.5 Hz.
Low-frequency exceedances (below 2.5 Hz) at low seismic hazard sites (SAp less
than LHT) do not require a plant to perform a full SMA or SPRA. Instead, it is
sufficient to first identify all safety-significant SSCs that are potentially
susceptible to damage from spectral accelerations at frequencies below which the
g 3-3 h
highest frequency fL (fL < 2.5 Hz) acceleration exceeds the SSE spectral
acceleration. Examples of SSCs and failure modes potentially susceptible to
damage from spectral accelerations at low frequencies are:
1) Liquid sloshing in atmospheric pressure storage tanks
2) Very flexible distribution systems with frequencies less than fL
3) Sliding and rocking of unanchored components
4) Fuel assemblies inside the reactor vessel
5) Soil liquefaction
After identifying all safety-significant SSCs that are potentially susceptible to
lower frequency accelerations, new HCLPF capacities using the GMRS shape
can be computed for these potentially low-frequency susceptible SSCs. The
HCLPF to GMRS seismic margin needs to be computed and reported. As long
as the HCLPF is greater than the GMRS for all of these potentially lowfrequency susceptible SSCs, the plant is screened out from having to perform
additional seismic evaluations.
If the IPEEE HCLPF1 capacity evaluations are considered to be sufficient for
screening (as described in Section 3.3.1), the IPEEE HCLPF response spectral
accelerations may be used for this HCLPF/GMRS comparison for screening
potentially low-frequency susceptible SSCs at low seismic hazard sites. The
IPEEE HCLPF response spectral accelerations also reduce rapidly as frequencies
reduce below 2.5 Hz so that the GMRS spectral accelerations might also exceed
the HCLPF spectral accelerations at low frequencies. In this case, new HCLPF
capacities can be computed for these potentially low-frequency susceptible SSCs
using the GMRS response spectrum shape instead of the IPEEE response
spectrum.
3.2.1.2
Range
Narrow Band Exceedances in the 1 to 10 Hz
If the GMRS exceeds the SSE in narrow frequency bands anywhere in the 1 to
10 Hz range, the screening criterion is as follows: In the 1 to 10 Hz range, a
point on the GMRS may fall above the SSE by up to 10% provided the average
ratio of GMRS to SSE in the adjacent 1/3 octave bandwidth (1/6 on either side)
is less than unity. There may be more than one such exceedance point above the
SSE in the 1 to 10 Hz range provided they are at least one octave apart. Figure
3-3 shows an example of this narrow-band criterion. If the GMRS meets the
criteria, no SMA or SPRA is required for the NTTF Recommendation 2.1
seismic review.
If the IPEEE HCLPF1 capacity evaluations are considered to be of sufficient
quality for screening, the IPEEE HCLPF response spectral accelerations may be
used for a HCLPF/GMRS comparison in narrow frequency bands. In this case,
1
IPEEE based screening is not applicable to Spent Fuel Pools because they were not included in
the IPEEE evaluations. See Section 7 for Spent Fuel Pool evaluation criteria.
g 3-4 h
the SSE is replaced by the IPEEE-HCLPF spectrum to determine if a plant can

be screened-out from further seismic review.
0.8
SSE or IPEEE HCLPF

GMRS
0.7
0.6
Acceleration (g)
0.5
0.4
0.3
0.2
0.1
0
0.1
10
Frequency (Hz)
100
Figure 3-3
Screening Example Narrow Exceedances at 2 Hz and 6 Hz
(5% Damping)
3.3
IPEEE Screening Task
The second method to demonstrate plant seismic adequacy based on screening

from further review consists of a comparison of the GMRS to the IPEEE
HCLPF spectrum, which is described in Section 3.3.2 below. The use of the
IPEEE HCLPF spectrum in the screening process is depicted in Boxes 3c, 3d,
and 3e in Figure 1-1. Note that IPEEE screening is not applicable to SFPs
because SFPs were not included in the scope of IPEEE evaluations [11]. See
Section 7 for SFP evaluation criteria.
For plants that conducted an SPRA, focused scope SMA, or full scope SMA
during the IPEEE, the screening is an optional approach that consists of the
comparison of the IPEEE HCLPF spectrum (IHS) to the new GMRS. If the
IPEEE HCLPF is used for screening, the IPEEE will be required to pass an
adequacy review (Diamond 3c in Figure 1-1). If the IPEEE demonstrates
sufficient quality, the next step in this screening process is to compare the IHS to
the GMRS in the 1 to 10 Hz part of the response spectrum (see Diamond 3d in
Figure 1-1). If the IHS exceeds the GMRS in the 1 to 10 Hz region, then a
check of the greater than 10 Hz part of the spectrum is performed, as shown in
Diamond 3e. If the IHS exceeds the GMRS in the greater than 10 Hz region,
then no further action is required for the NTTF 2.1 seismic review (Box 4 in
Figure 1-1). If there are exceedances in the greater than 10 Hz region, then a
g 3-5 h
high-frequency confirmation should be performed (Box 3f in Figure 1-1) as

described in Section 3.4.
3.3.1
IPEEE Adequacy
Background
Seismic risk assessments performed as part of the Individual Plant Examination
of External Events (IPEEE) for Severe Accident Vulnerabilities (Generic Letter
pp
ap y to levels higher
gh than
88-20, Supplement
4)) [2] that demonstrate p
plant capacity
criteria, in which case these plants would not need to perform new seismic risk
analyses. IPEEE submittals using either SPRA or SMA analyses can be
considered for screening, but in either case the analysis must have certain
attributes to be considered for review by the NRC staff.
Use of IPEEE Results for Screening
Certain criteria are necessary if licensees choose to screen a facility based on
IPEEE results. The criteria for screening have been grouped into four categories:
General Considerations
Prerequisites
Adequacy Demonstration
Documentation
Responses to the items in the Prerequisite and Adequacy Demonstration

categories should be provided in the hazard submittal to the NRC.
General Considerations
IPEEE reduced scope margin assessments cannot be used for screening. Focused
scope margin submittals may be used after having been enhanced to bring the
assessment in line with full scope assessments. The enhancements include (1) a
full scope detailed review of relay chatter for components such as electric relays
and switches, and (2) a full evaluation of soil failures, such as liquefaction, slope
stability, and settlement.
The spectrum to be compared to the GMRS for screening purposes should be
based on the plant-level HCLPF actually determined by the IPEEE and
reported to the NRC. If this is less than the review level earthquake (RLE)
spectrum, then the RLE must be shifted appropriately to reflect the actual
achieved HCLPF. In cases where modifications were required to achieve the
HCLPF submitted in the IPEEE, verify that the changes were implemented
(and describe the current status) in the submittal. This information is also
required as part of the Recommendation 2.3 seismic walkdown. Similarly, the
uniform hazard spectrum (UHS) for IPEEE SPRA should be anchored at the
plant-level HCLPF.
g 3-6 h
Prerequisites
Responses to the following items should be provided with the hazard evaluation.
In order to use the IPEEE analysis for screening purposes and to demonstrate
that the IPEEE results can be used for comparison with the GMRS:
1) Confirm that commitments made under the IPEEE have been met. If not,
address and close those commitments.
2) Confirm whether all of the modifications and other changes credited in the
IPEEE analysis are in place.
3) Confirm that any identified deficiencies or weaknesses to NUREG-1407
[11] in the plant specific NRC Safety Evaluation Report (SER) are properly
justified to ensure that the IPEEE conclusions remain valid.
4) Confirm that major plant modifications since the completion of the IPEEE
have not degraded/impacted the conclusions reached in the IPEEE.
If any of the four above items are not confirmed and documented in the hazard
submittal to the NRC, then the IPEEE results may not be adequate for
screening purposes even if responses are provided to the adequacy criteria
provided below.
Adequacy Demonstration
The following items, and the information that should be provided, reflect the
major technical considerations that will determine whether the IPEEE analysis,
documentation, and peer review are considered adequate to support use of the
IPEEE results for screening purposes.
With respect to each of the criteria below, the submittal should describe the key
elements of (1) the methodology used, (2) whether the analysis was conducted in
accordance with the guidance in NUREG-1407 [11] and other applicable
guidance, and (3) a statement, if applicable, as to whether the methodology and
results are adequate for screening purposes. Each of the following should be
addressed in the submittal to the NRC.
1) Structural models and structural response analysis (use of existing or new
models, how soil conditions including variability were accounted for)
2) In-structure demands and ISRS (scaling approach or new analysis)
3) Selection of seismic equipment list or safe shutdown equipment list
4) Screening of components
5) Walkdowns
6) Fragility evaluations (generic, plant-specific analysis, testing, documentation
of results)
7) System modeling (diversity of success paths, development of event and fault
trees, treatment of non-seismic failures, human actions)
8) Containment performance
g 3-7 h
9) Peer review (how peer review conducted, conformance to guidance, peer

review membership, peer review findings and their disposition)
Documentation
Licensees that choose to implement the use of the IPEEE results for screening
purposes should provide a response for each of the criteria in the Prerequisite and
Adequacy Demonstration categories in their hazard submittal to the NRC.
Licensees should also provide an overall conclusion statement asserting that the
IPEEE results are adequate for screening and that the risk insights from the
IPEEE are still valid under current plant configurations. The information used
by each licensee to demonstrate the adequacy of the IPEEE results for screening
purposes should be made available at the site for potential staff audit.
3.3.2
Development of HCLPF Spectrum
The IHS is developed directly from the plant HCLPF capacity established in the
IPEEE program. The IPEEE-reported HCLPF values were typically calculated
by each plant during the 1990s and documented in the IPEEE submittal reports
sent to the NRC by the licensees. These HCLPF values for many of the plants
are also documented in NUREG-1742 Perspectives Gained from the
Individual Plant Examination of External Events (IPEEE) Program, April 2002
[4]. For those plants that performed an SMA, the IHS is anchored to the lowest
calculated HCLPF of any SSC, and the shape of the IHS is consistent with the
RLE used for the SMA (typically the NUREG/CR- 0098 shape). For those
plants that conducted an SPRA as part of the IPEEE program, a plant HCLPF
value was typically calculated (or can be calculated) from the plant core damage
frequency (CDF) and the IHS should be anchored at that value. The shape of
the IHS should correspond to the UHS associated with the seismic hazard
utilized within the SPRA. Typically, the shapes of the UHS are similar between
the 10-4 and the 10-5 return period UHS and, thus, either shape could be used for
the purpose of generating the IHS. These two return periods are considered to be
the appropriate ones for use in the generation of the IHS since the cumulative
distribution of the contribution to the CDF has typically been shown to be
centered in this return period range.
3.3.3
Comparison of IPEEE HCLPF Spectrum to GMRS
An example of the comparison of a GMRS to the IHS is shown in

Figure 3-4. The IHS exceeds the GMRS in the 1 to 10 Hz range, and thus the
lower frequency criteria (Diamond 3d of Figure 1-1) have been met. However,
for this example, the higher frequency criteria (Diamond 3e in Figure 1-1) have
not been met since the GMRS exceeds the IHS in this range. It is noted that (a)
the control point for the IHS will typically be defined in a similar way as for the
SSE, which is described in Section 2.4.1, and (b) the treatment of Narrow Band
Exceedance is the same as discussed in Section 3.2.1 for SSE.
g 3-8 h
Figure 3-4
Example Comparison of GMRS to IHS (5% Damping)
3.4
Treatment of High-Frequency Exceedances
Equipment important to safety within operating NPPs has been seismically

p
eq
qualified for the SSE defined for each plant.
The equipment
has also been
IPEEE program. The SSE
evaluated, in general, for a RLE
and RLE ground motions, however, do not typically include significant
frequency content above 10 Hz. Seismic hazard studies conducted in the late
1990s developed UHS that had spectral
peaks
occurringg in the 20 to 30 Hz
pe
pe
range. EPRI Report
po NPImplementation,
pl
[26], included an appendix titled
-Frequency Ground Motions in
Seismic Margin Assessment for Severe Accident Policy Resolution. This
appendix reviewed the bases for concluding that high-frequency motions were, in
general, non-damaging to components and structures that have strain- or stressbased potential failures modes. It concluded that components, such as relays and
other devices subject to electrical functionality failure modes, have unknown
acceleration sensitivity for frequencies greater than 16 Hz. Thus, the evaluation
of high-frequency vulnerability was limited to components that are subject to
intermittent states.
In the IPEEE program, the consideration of high-frequency
vulnerability of
ig
components was focused on a list of
relays mutually agreed to by the
industry and the NRC, with known earthquake or shock sensitivity. These
specific model relays,
as low ruggedness
relays were identified in
sp
ys, designated
gn
gg
ear Power Plant Relay
Rather than considering highfrequency capacity vs. demand screening, relays on this list were considered
g 3-9 h
program outliers and were evaluated using circuit analysis, operator actions, or
component replacements.
EPRI published the following reports during initial new plant licensing activities
to provide additional information regarding the potential high-frequency
vulnerability of NPP SSCs:
on: The Effects of

High-Frequency
gh eq
y Ground Motion on Structures, Components, and
28].
Screening of Components Sensitive to High-Frequency Vibratory

October 2007 [29].
Report 1015108 [28] summarized a significant amount of empirical and
theoretical evidence, as well as regulatory precedents, that support the conclusion
that high-frequency vibratory motions above about 10 Hz are not damaging to
the large majority of NPP structures, components, and equipment. An exception
to this is the functional performance of vibration sensitive components, such as
relays and other electrical and instrumentation devices whose output signals
could be affected by high-frequency excitation. Report 1015109 [29] provided
guidance for identifying and evaluating potentially high-frequency sensitive
components for plant applications that may be subject to possible high-frequency
motions.
In response to the current NTTF activities, EPRI has established a test program
to develop data to support the high frequency confirmation in Step 3f of Figure
1-1 as well as fragility data for a SPRA (Step 6a) or SMA (Step 6b) of Figure 1-1
for potential high-frequency sensitive components. The test program will use
accelerations or spectral levels that are sufficiently high to address the anticipated
high-frequency in-structure and in-cabinet responses of various plants.
Therefore, it will not be necessary for those plants where GMRS > SSE or IHS
only above 10 Hz to perform dynamic analysis of structures to develop ISRS.
3.4.1
Scope of High-Frequency Sensitive
Components
The following types of failure modes of potentially high-frequency sensitive
components and assemblies have been observed in practice:
Inadvertent change of state
Contact chatter
Change in output signal or set-point
Electrical connection discontinuity or intermittency (e.g., insufficient contact

pressure)
Mechanical connection loosening
Mechanical misalignment/binding (e.g., latches, plungers)

g 3-10 h
Cyclic strain effects (e.g., cracks in solder joints)
Wiring not properly restrained
Inadequately secured mechanical fasteners and thumb screw connections
These failure modes are considered below to determine the appropriate scope of
potentially high-frequency sensitive components requiring additional information
to perform the NTTF 2.1 seismic screening in Figure 1-1, Step 3f.
3.4.1.1
EPRI 1015109 Potentially High Frequency
Sensitive Components
EPRI Report 1015109 [29] reviewed potentially high-frequency sensitive
components and recommended change of state, contact chatter, signal
change/drift, and intermittent electrical connections as the most likely failure
modes. These are the first four failure modes highlighted in the above list.
Failures resulting from improper mounting design, inadequate design
connections and fasteners, mechanical misalignment/binding of parts, and the
rare case of subcomponent mechanical failure, are associated with the same
structural failure modes as those experienced during licensing basis qualification
low frequency testing conducted in accordance with the Institute of Electrical
and Electronics Engineers (IEEE) Standard 344 [25]. Because the equipment
experiences higher stresses and deformations when subjected to low-frequency
excitation, these failure modes are more likely to occur under the low-frequency
qualification testing.
The evaluation of potentially high-frequency sensitive components in new plants
was therefore directed to mechanically actuated bi-stable devices, such as relays,
contactors, switches, potentiometers and similar devices, and those components
whose output signal or settings (set-points) could be changed by high-frequency
vibratory motion. Table 3-1 shows the components identified in EPRI Report
1015109 [29] as being potentially sensitive to high-frequency motion.
3.4.1.2
Equipment
AP1000 Potentially High Frequency Sensitive
During licensing reviews for the AP1000, Westinghouse and the NRC identified
a broader list of potentially high-frequency sensitive components and assemblies
(Table 3-2) to be evaluated in the AP1000 Design Control Document [30].
g 3-11 h
Table 3-1
EPRI 1015109 Potentially High Frequency Sensitive Items
Electro-mechanical relays (e.g.,

control relays, time delay relays,
protective relays)
Circuit breakers (e.g., molded case

and power breakers low and
medium voltage)
Control switches (e.g., benchboard,

panel, operator switches)
Process switches and sensors (e.g.,

pressure, temperature, flow,
limit/position)
Electro-mechanical contactors (e.g.,

MCC starters)
Auxiliary contacts (e.g., for

MCCBs, fused disconnects,
contactors/starters)
Transfer switches (e.g., low and

medium voltage switches with
instrumentation)
Potentiometers (without locking

devices)
Digital/solid state devices (mounting

and connections only)
The primary difference between the list of components in EPRI 1015109 [29]
and the AP1000 list [30] is that the EPRI 1015109 list is focused on potentially
sensitive subcomponents, and the AP1000 list is focused on assemblies that
would include those subcomponents. For example, the potentially sensitive parts
of a Battery Charger or a 250 Vdc Motor Control Center are the relays, switches,
and contactors noted in the EPRI 1015109 component list [29]. Therefore,
evaluating the potential sensitivity of the items in the EPRI 1015109 list would
also address the items in the AP1000 list.
Three key exceptions on the AP1000 list [30] are transformers, batteries, and
valves (motor-operated valves (MOVs), air-operated valves (AOVs), solenoid
valves (SVs). Transformers are primarily passive systems with strain- or stressbased potential failures modes. Some transformers may include subcomponents
on the EPRI 1015109 list [29], but they would be addressed as noted above.
Battery cells have a material aging phenomenon that occurs over time. There is
no indication that cell electrical degradation is influenced by the frequency
content of the cell support motion being either high-frequency or low-frequency.
Batteries do not fail during support motion, but rather fail to produce the rated
amp-hour capacity following the support motion. It is judged that the postearthquake electrical capacity is a function of cell age and the RMS acceleration
level of the input motion rather than the frequency content of the motion.
Batteries that are less than ten years in age would not experience post-earthquake
degradation due to cell shaking.
Valves have been subjected to significant high-frequency test motions due to
Boiling Water Reactor (BWR) hydrodynamic loads and have not demonstrated
high frequency unique sensitivities. EPRI Report 1015108 [29] provides an
example of previous MOV operator combined seismic and BWR hydrodynamic
qualification testing with inputs up to 100 Hz. This example valve operator is
the same as used in other plant designs. These types of tests also show that
g 3-12 h
additional high frequency content does not affect equipment function. In

addition, line mounted valves and operators are subjected to 5-100 Hz sine sweep
vibration testing as part of normal valve qualification to simulate normal plant
induced vibration environments.
Table 3-2
AP1000 Potentially High Frequency Sensitive Items
125V Batteries
Battery Chargers
250Vdc Distribution Panels
120Vdc Distribution Panels
Fuse Panels
Fused Transfer Switches
Battery Disconnect Switches
Termination Boxes
250Vdc Motor Control Centers
250Vdc Switchboard
Regulating Transformers
Inverters
6.9KV Switchgear
Reactor Trip Switchgear
Level Switches (Core Makeup

Tank, Containment Flood)
Neutron Detectors (Source Range,

Intermediate Range, Power range)
Radiation Monitors (Containment

High Range Area, Control Room
Supply Air)
Speed Sensors (Reactor Coolant

Pump)
Transmitters (Flow, Level,

Pressure, Differential Pressure)
Control Room (Workstations,

Switch Station, Display Units)
Protection and Safety Monitoring

Systems (System Cabinets, Transfer
Switches, Neutron Flux
Preamplifiers, High Voltage
Distribution Boxes)
Motor Operated Valves (Motor

Operators, Limit Switches)
Other Valves (Squib [Explosive

Opening] Operators, Limit
Switches)
Air Operated Valves (Solenoid

Valves, Limit Switches)
3.4.1.3
Component Types to be Evaluated
The list of component types to be evaluated in the above noted high frequency
test program was developed based on the reviews described in Sections 3.4.1.1
and 3.4.1.2 and is provided in Table 3-3. A subset of these component types are
used in the Phase 1 testing effort described below. The complete list of
component types in the table will be considered in a follow-on Phase 2 testing
effort.
Test samples will be selected from the component types in Table 3-3 to represent
the components installed in operating nuclear power plants. A review of selected
utility components lists will be used to inform the test sample selection.
g 3-13 h
Table 3-3
High Frequency Confirmation Component Types
Electro-mechanical relays (e.g.,

control relays, time delay relays,
protective relays)
Circuit breakers (e.g., molded case

and power breakers low and
medium voltage)
Control switches (e.g., benchboard,

panel, operator switches)
Process switches and sensors (e.g.,

pressure, temperature, flow,
limit/position)
3.4.2
Electro-mechanical contactors (e.g.,

MCC starters)
Auxiliary contacts (e.g., for

MCCBs, fused disconnects,
contactors/starters)
Transfer switches (e.g., low and

medium voltage switches with
instrumentation)
Potentiometers (without locking

devices)
Phase 1 Testing
The high-frequency test program consists of two phases. The first phase pilot
effort has focused on (1) developing a recommended high-frequency test protocol
to be used in the full test program, and (2) acquiring sufficient data to allow
development of criteria for comparison of fragility levels obtained using highfrequency wide-band and narrow-band motions.
3.4.2.1
Phase 1 Test Samples
The components included in the Phase 1 test program were selected to provide a
representative sample of the types of components listed in Section 3.4.1.3, as well
as a variety of expected seismic capacity levels. The list of components used for
Phase 1 testing is provided in Table 3-4.
Table 3-4
Phase 1 Test Samples
Electro-mechanical relay (600V

industrial control relay)
Electro-mechanical relay (socket

mounted control relay)
Electro-mechanical relay
(pneumatic timing relay)

industrial control relay)
Electro-mechanical contactor (with

auxiliary and overload contacts)
Electro-mechanical relay (lockout

relays, two configurations
considered)

control relay prior HF testing
history)
Electro-mechanical relay (induction

disk protective relay)
Process switch
(pressure switch)
Electro-mechanical relay (auxiliary

relay - hinged armature)
g 3-14 h
3.4.2.2
Phase 1 Testing Protocol
A number of test parameters were investigated in Phase 1, as described below.

Primary Frequency Range of Interest
For the component types listed in Section 3.4.1.3, licensing basis seismic
qualification testing is typically performed over a frequency range up to 33 Hz.
For floor- or wall-mounted components, in-structure response spectra typically
peak in the 4 to 10 Hz range and reach the Zero Period Acceleration (ZPA) in
the 15 to 20 Hz range. For in-cabinet mounted components, IEEE C37.98 [31]
recommends a response spectrum shape with peak spectral accelerations in the 4
to 16 Hz range and a ZPA at 33 Hz.
Some of the new ground motion estimates have peak accelerations in the 25 to
30 Hz range, which may produce significant in-structure or in-cabinet motions
in the 20 to 40 Hz range. Figure 3-5 shows an example ground motion where instructure and in-cabinet high-frequency motions may be significant.
Because licensing basis seismic qualification testing adequately addresses the
lower frequency range, the high-frequency test program will focus on this higher
frequency range. The primary focus of the high-frequency testing program is the
20 to 40 Hz frequency range. Phase 1 testing initially considered a broader
frequency range of 16 to 64 Hz to insure that the focus on the 20-40 Hz range is
sufficient.
Example Site
Spectral Acceleration, g
0.7
5% Damping
0.6
SSE (RG 1.60)
0.5
GMRS
0.4
0.3
0.2
0.1
0
0.1
10
100
Frequency. Hz
Figure 3-5
Example High Frequency Ground Motion Response Spectrum
g 3-15 h
Test Input Motions

Three types of test input motions were investigated in Phase 1: sine sweeps,
random multi-frequency (RMF) motions and filtered RMF motions. In each
case, the input motions were increased in amplitude until either the components
failed the acceptance criteria (typically 2 ms contact chatter per ANSI C37.98
[31]), or had anomalous behavior, or the test machine limits were reached.
Sine Sweep Input Motions This test series used single-axis sine sweep inputs
with constant acceleration levels over the 16 to 64 Hz range. The components
were tested in each primary direction (e.g., front-to-back, side-to-side, vertical)
in the de-energized (non-operate) state with subsequent tests in the energized
(operate) state. The objective of this test series was to develop a plot of chatter
threshold frequency vs. peak input motion acceleration as a means of displaying
the regions of high-frequency sensitivity for each component.
RMF Input Motions This test series used wide-band multi-frequency tri-axial
independent random motions with response spectra covering three separate
amplified frequency ranges as shown in Figure 3-6. The three frequency ranges
were 16 to 32 Hz, 24 to 48 Hz, and 20 to 40 Hz. The general shape of the
amplified spectral region was patterned after the normalized test shape from
IEEE C37.98 [31] with the peak acceleration region being 2.5 times the ZPA,
but with the frequency ranges shifted as shown in Figure 3-6. A set of three
motions was generated for each frequency range. Each axis of motion of each set
was independent but had the same general response spectrum shape and
amplitude. The purpose of these tests was to determine the fragility level of each
device associated with each set of RMF motions for a given frequency range.
Filtered Random Multi-Frequency (FRMF) Input Motions This test series
used wide-band multi-frequency independent random input motions along two
primary axes with a set of narrow-band filtered motions along the third axis as
depicted in Figure 3-7. The narrow-band motions were applied along the third
axis, one at a time, at the indicated 1/6 octave frequencies between 17.8 and 44.9
Hz (21.2 Hz, 23.8 Hz, 26.7 Hz, 30.0 Hz, 33.6 Hz, 37.8 Hz). The RMF
motions applied to the other two axes had strong motion frequency range of 17.8
Hz to 44.9 Hz. The FRMF motions were applied separately in the component
front-to-back direction and the side-to-side direction. Note that the FRMF
testing was intended to simulate either in-structure response or high frequency
local panel in-cabinet response, which is expected to be dominated by front-toback or side-to-side responses; therefore, the filtered motions were limited to
those two directions. Each filtered motion had the appearance of multiple sinebeat motions superimposed on a wide band random backbone motion. The
purpose of these tests was to determine the fragility level of each device
associated a given FRMF motion. The comparison of the fragility response
po
spectra for both the FRMF and RMF motion allows a clipping factor to be
defined that can be used to convert an in-structure or in-cabinet demand
(response spectrum) to an effective wide band motion for comparison to a RMF
fragility test spectrum.
g 3-16 h
Figure 3-6
Random Multi-Frequency Test Input Motions
Spectral Acceleration (Normalized)
6
1 2 3 4 5 6
HF SB1
HF SB2
HF SB3
HF SB4
HF SB5
HF SB6
HF SAwb
0
10
100
Frequency, Hz
Figure 3-7
Filtered Random Multi-Frequency Test Input Motions
3.4.2.3
Phase 1 Test Results
Component High Frequency Sensitivity

No chatter or change of state occurred for any Phase 1 relay in the energized state
for any input motion, thus a relay in the energized state is considered to not be
frequency sensitive
g 3-17 h
Four devices did not have any chatter in the de-energized state for the highest
input levels tested. These devices may be considered as not frequency sensitive.
Two devices were expected to not have any high-frequency sensitivity based on
the high demonstrated low frequency ruggedness (12.5 g spectral), however,
these models had anomalous behavior in the greater than 30 Hz range. The
remaining 5 relay models demonstrated various high-frequency sensitivities for
the de-energized state.
Test Input Motion Results
The general conclusions from each of the three test input motion types are
described below.
Sine Sweep Input Motions The sine sweep tests were primarily exploratory
tests but they did not appropriately simulate earthquake motions at the
component mounting locations and at high amplitudes, they over predicted
component sensitivity.
RMF Input Motions The RMF tests over the three peak frequency ranges (16
to 32 Hz, 20 to 40 Hz, and 24 to 48 Hz) proved to be efficient to perform, and
effective at identifying high frequency component sensitivity. Some of the Phase
1 test components were a little more sensitive to the test motions from 16 to 32
Hz than the other two test ranges; however, this was due to the lower frequency
energy included in those tests. That lower frequency energy is already included
standard relay testing performed using the IEEE relay qualification standard
[31]; therefore, it is more indicative of lower frequency sensitivities than high
frequency sensitivities. Between the remaining two RMF frequency ranges, the
20 to 40 Hz range is more consistent with the expected high frequency ground
motions shown in Figure 3-5. The Phase 1 components were also slightly more
sensitive to that frequency range input than the 24 to 48 Hz RMF motion.
FRMF Input Motions The FRMF tests may be the most accurate at
simulating the kinds of earthquake motions at the component mounting
locations, but they were very time consuming to perform. Comparisons of the
FRMF peak spectral accelerations that produced component chatter with the
RMF peak spectral accelerations confirmed that previous narrow band clipping
factors (e.g. [39], Appendix Q) are also generally applicable to high frequency
motions.
Phase 1 Overall Conclusions
The Phase 1 study indicates that the high-frequency sensitivity of contact devices
is generally device specific. Thus, the best means to identify such frequency
sensitivity is to test the devices. Additional testing in Phase 2 may facilitate more
general conclusions for some categories of potentially sensitive devices (e.g.
miniature relays, potentiometers).
The use of the 20-40 Hz multi-frequency random input motion appears to be the
best compromise for determining frequency sensitivity. The use of other input
g 3-18 h
motions requires considerable effort and do not appear to provide any better
resolution for determining high-frequency sensitivity.
Filtered multi-frequency narrow-band inputs resulted in peak spectral fragility
values that were 2-3 times the spectral fragility values obtained using the wideband multi-frequency inputs. Thus, it appears that the clipping factors used for
low frequency fragility are valid for high-frequency fragility. (This is still under
study)
3.4.3
Phase 2 Testing
Phase 2 testing will be performed to address the component types identified in

Section 3.4.1.3. The complete test results will be compiled as appropriate to
support utility high-frequency confirmation screening in Figure 1-1, Step 3f, as
well as SPRA or SMA evaluations in Figure 1-1, Steps 6a and 6b.
3.4.3.1
Phase 2 Test Protocol
Based on the Phase 1 testing, the 20 to 40 Hz RMF response spectrum shape

will be used to develop the test motions for the Phase 2 test protocol. These
motions will be used to determine the fragility spectra for each component.
3.4.3.2
Expanded Sample
The test sample list for Phase 2 testing will be selected to address the range of
component types identified in Section 3.4.1.3. Components will be selected to
represent a distribution of manufacturers and specific model numbers.
Components will also be selected to address a variety of contact mechanical
motions (e.g., plunger- and clapper-type relays) and physical forms (e.g., socket
and bolted mounting configurations). The number of components in any
component type category may be adjusted depending on the expected degree of
high-frequency sensitivity. In addition, the specific model numbers selected may
be adjusted depending on the component availability. To the extent practical, the
distribution of test samples will be selected to achieve the broadest possible
conclusions.
g 3-19 h
Section 4: Seismic Hazard and

Screening Report
The NRC 50.54(f) information request associated with NTTF Recommendation
2.1 seismic is delineated in [1]. Within 1.5 years of the March 12, 2012 date of
the information request, each CEUS addressee is requested to submit
information related to the seismic hazard and the screening portions of the
program (see Boxes 1, 2, and 3a-3e of Figure 1-1). An example of the type of
information that has been requested, which could form the table of contents for
that report, is listed below.
Introduction
1. Responding to 50.54(f) letter
2. Brief history of seismic licensing basis (summary of SSE and which
codes, standards, and methods were used in the design of Seismic
Category I SSCs)
3. Brief description of method used to develop GMRS and outcome of
screening comparisons
Seismic Hazard Results: GMRS

1. Regional and Local Geology
a. Regional Geology
i. 1-2 paragraphs describing tectonic setting and history
b. Local Geology
i. 1-2 paragraphs described any prominent geologic features,
complexity of geologic features (folding and faulting)
2. Probabilistic Seismic Hazard Analysis
a. Probabilistic Seismic Hazard Analysis
i. Summary of sources used (sub-set of CEUS Seismic Source
Characterization sources or site-specific for WUS sites)
ii. Ground Motion Prediction Equations used or developed
b. Base Rock Seismic Hazard Curves (if hard rock site)
i. Common fractiles and mean for spectral frequencies for which
GMPEs are available
3. Site Response Evaluation (if not a hard rock site)
g 4-1 h
a. Description of Subsurface Materials and Properties

i. Soil/rock types, layering, and properties
b. Development of Base Case Profiles and Nonlinear Material
Properties
i. Resources used and basis for base case profiles
1. Base case shear wave velocity profiles
2. Selected Shear Modulus and Damping curves
c. Randomization of Profiles
i. Randomization method and parameters
ii. Constraints applied on layer thicknesses and velocities
iii. Kappa values
d. Input Spectra
i. Fourier amplitude spectra and response spectra including input
elevation
ii. Any modifications to input spectra (kappa correction)
e. Methodology
i. Brief description of Random Vibration Theory (RVT) or time
series approach
ii. Parameters used in RVT or time histories used
f.
Amplification Functions
i. Amplification functions
ii. Amplification versus Input Amplitude including uncertainty
bands for each of the spectral frequencies
g. Control Point Seismic Hazard Curves

i. Common fractiles and mean for spectral frequencies for which
GMPEs are available
4. Ground Motion Response Spectrum
a. Uniform Hazard Response Spectra
i. 10-4 and 10-5 UHRS
b. GMRS
i. Table of 10-4 and 10-5 UHS, Design Factor values, and
GMRS
Safe Shutdown Earthquake Response Spectra

1. Spectral Shape and Anchor Point PGA for 5% critical damping
a. Brief description from FSAR
2. Control Point Elevation(s)
g 4-2 h
a. Description from FSAR or assumptions used to determine control

point elevation
Special Screening Considerations

1. GMRS and SSE Comparison
a. Discussion of results
b. High-frequency, Narrow Band Exceedance (if applicable)
2. Evaluation of IPEEE Submittal
a. see Section 3.3-1
3. GMRS and IHS Comparison
a. If applicable, discussion of results (narrow-band exceedance if
applicable)
4. Screening for Risk Evaluation (SPRA or SMA see Section 6.2)
a. If applicable, discussion of results
Interim Actions*
1. Any interim actions taken or planned while risk evaluation is being
performed
Conclusions
1. Summary of results
2. Path forward based on Screening Evaluations
actions taken or planned to address the higher seismic hazard relative to the
design basis, as appropriate, prior to the risk evaluation
of information which could be included in this response are:
Modifications or upgrades that the addressee decides to undertake prior to

the seismic risk evaluation described in Section 6 of this report.
Addressee intentions relative to conducting an SPRA or SMA.
Description of the types of exceedances (low-frequency range, highfrequency range, narrow-frequency band, etc.) and the types of SSCs which
may be affected by that exceedance (e.g., high-frequency exceedance could
affect chatter sensitive devices which are going to be addressed by the EPRI
testing program described in Section 3.4 of this report).
g 4-3 h
Section 5: Prioritization
(Schedule)
The resolution for the 2.1 seismic information requests [1] consists of first
generating the new seismic hazard information for all sites, followed by the
screening assessments described in the previous sections. Those plants required to
perform the seismic risk evaluation are then required to be prioritized in terms of
their schedule for submittals (Diamond 5, Figure 1-1). This prioritization occurs
after seismic hazard and screening submittals described in Section 4 of this report
are submitted to the NRC. That report is scheduled to be completed by Fall
2013 for CEUS sites and Spring 2015 for the WUS sites.
The intent of the prioritization is to take into account:
the amount of the seismic hazard exceedance

o
GMRS to SSE Ratios
GMRS to IHS
the available resources (industry-wide and individual utility multi-unit fleets),

and
available plant mitigation strategies
Consideration should be given to plants that are already conducting SPRAs.

Lessons learned from these early SPRA implementations will be valuable
resources to the industry and NRC.
The exact criteria/methods to be used for this prioritization are being discussed
between the NRC and the nuclear utility industry as part of ongoing discussions
on the resolution of the 2.1 seismic program.
g 5-1 h
Section 6: Seismic Risk

Evaluation
6.1
Background on SPRA and SMA
SPRA and SMA studies have been conducted for many of the U.S. NPPs over
the last twenty years. Initially they were conducted to answer safety concerns in
heavily populated areas. The next widespread application was for satisfaction of
the USNRC request for information regarding severe accident vulnerabilities in
Generic Letter 88-20, Supplement 4 [2]. The USNRC is currently encouraging
the use of PRA for making risk-informed decisions and has developed a RiskInformed Regulation Implementation Plan [32] and associated regulatory guides.
The Licensees in turn are using PRAs for Changes to Licensing Basis, Changes
to Technical Specifications, Graded Quality Assurance, Significance
Determination Processes, etc. Seismic PRAs are also required for each new NPP
one year prior to fuel load. SPRAs and SMAs are now also being recommended
as paths to conduct the seismic risk evaluations within Tasks 6 and 7 of Figure 11.
6.1.1
SPRA Methods and Procedures
Current U.S. NPPs were designed to withstand a conservatively selected large

earthquake ground motion (the SSE) with adequate margins introduced at
different stages of design, analysis, qualification, and construction. However, it is
understood that larger earthquake ground motions (although rare) could occur.
The basic objective of the SPRA is to estimate the probability of occurrence of
different levels of earthquake ground shaking that may affect the plant, and to
assess the plant response to such ground motions. Following the historical PRA
practice, the results of this plant seismic assessment are presented in terms of
seismically induced CDF and large-early release frequency (LERF). SPRAs
completed to date, have shown that the seismic contribution to the overall CDF
and LERF at some NPPs could be significant and occasionally can even be
dominant. Therefore, a quantitative assessment of the seismic risk (e.g., SPRA)
can be an important component of the overall risk-informed decision making
process.
The key elements of a SPRA can be identified as:
Seismic Hazard Analysis: Used to assess the seismic hazard in terms of the
frequency of exceedance for selected ground motion parameters during a
specified time interval. The analysis involves the characterization of
g 6-1 h
earthquake sources, evaluation of the regional earthquake history, and an

estimation of the intensity of the earthquake-induced ground motion at the
site (Figure 6-1).
Figure 6-1
Example Seismic Hazard Curve
Seismic Fragility Analysis: Estimates the conditional probability of SSC

failures at a given value of a seismic motion parameter, such as PGA, peak
spectral acceleration, floor spectral acceleration, etc. Seismic fragilities used
in a seismic PRA are realistic and plant-specific based on actual current
conditions of the SSCs in the plant, as confirmed through a detailed
walkdown of the plant. The fragilities of all the systems that participate in
the accident sequences are included (Figure 6-2).
g 6-2 h
Figure 6-2
Example Seismic Fragility Curve
Systems/Accident Sequence Analysis: Modeling of the various

combinations of structural and equipment failures that could initiate and
propagate a seismic core damage sequence.
Risk Quantification: Calculates the frequencies of severe core damage and

radioactive release to the environment by using the plant logic model and
accident sequences for which the SSC fragilities are integrated with the
seismic hazard. The analysis is usually carried out by adding some
earthquake-related basic events to the PRA internal events model, as well as
eliminating some parts of the internal events model that do not apply or that
can be screened-out.
The overall SPRA process is characterized in Figure 6-3.
g 6-3 h
Figure 6-3
Overview of the SPRA Methodology
The detailed methods and criteria to develop the seismic fragility, seismic hazard,
and seismic plant logic models are well beyond the scope of this guide.
Fortunately, there are many technical references which document these methods.
Table 6-1 is intended to provide a good list of references on these topics, while
there are obviously many more in the literature.
g 6-4 h
Table 6-1
Partial List of SPRA Technical References
SPRA Topic
SPRA
Recommended Document Title

Seismic Probabilistic Risk Assessment
EPRI 1002989 (Dec 2003)
Implementation Guide
[33]
Seismic Evaluation of Existing Nuclear
Safety Report Series No. 28
Power Plants
[34]
Probabilistic Safety Assessment for
IAEA Tecdoc-724
Seismic Events
(Oct 1993) [35]
Seismic Fragility Applications Guide

Update
EPRI Report 1019200

(Dec 2009) [36]
Seismic Fragility Application Guide
EPRI 1002988 (Dec 2002)

[37]
Methodology for Developing Seismic

Fragilities
EPRI TR-103959
(June 1994) [38]
A Methodology for Assessment of

Nuclear Plant Seismic Margin
EPRI NP 6041 (Oct 1988)

[39]
PRA Procedures Guide: A Guide to

the Performance of Probabilistic Risk
Assessments for Nuclear Power Plants
NUREG/CR-2300 (1983)
[54]
Recommendations for Probabilistic

Seismic Hazard Analysis: Guidance on
Uncertainty and Use of Experts
NUREG/CR-6372 (1997)
[13]
Practical Implementation Guidelines

for SSHAC Level 3 and 4 Hazard
Studies
NUREG-2117 (2012) [23]
Technical Basis for Revision of

Regulatory Guidance on Design
Ground Motions: Hazard- and RiskConsistent Ground Motion Spectra
Guidelines
NUREG/CR-6728
(Oct 2001) [24]
Seismic Fragility
Seismic Hazard
6.1.2
Reference
NRC SMA Methods and Procedures
The Nuclear Regulatory Commission (NRC) staff developed interim staff

guidance (ISG) [100] on an acceptable method for performing a Seismic Margin
Assessment (SMA) as referred to in the March 12, 2012 NRC letter [1]. This
SMA method includes enhancements to the NRC SMA method originally
described in NUREG/CR-4334 that the NRC deemed necessary to meet the
objectives of the 50.54(f) letter. The NRC ISG approach for SMA is specifically
intended be used to respond to the 50.54(f) letter. The level of effort to perform
a SMA to meet this ISG is nearly equal to that required for a SPRA. The
primary difference is that the SMA reports results in terms of HCLPF values,
rather than risk metrics such as CDF or LERF.
g 6-5 h
A list of the high level features and enhancements to an SMA that are
documented in the draft ISG are listed below. Some of these topics are similar to
staff positions taken during the IPEEE program, and others are additional
enhancements.
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 reducedscope SMA (as described in NUREG-1407).
The systems model should be enhanced over what was contained in either
ig
the original
NRC SMA gguidance (i
(in NUREG/CR-4334 and NURE/CR-1407).
The scope should include certain containment and containment systems, so

as to enable analysis of the plant-level HCLPF for large early release.
The use of the soshould follow certain guidance

gu
p
provided in [[100]. The Convolution 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.
Separately report HCLPF capacities for those sequences with non-seismic

failures and human actions and HCLPF capacities for those sequences
without them.
A formal peer review of the SMA is required. The ISG peer review
requirements are not consistent with the peer review requirements of RG
1.200 or the ASME/ANS PRA Standard.
The details for each of these features and enhancements are described in detail
within the draft ISG [56].
Licensees may propose other methods for satisfying SMA requirements of the
50.54(f) NRC letter. The NRC staff will review such methods and determine
their acceptability on a case-by-case basis.
6.2 Criteria for Selection of Risk Evaluation Method
(SPRA vs. SMA)
As shown in Figure 1-1, plants that do not meet the screening criteria outlined in
Section 3 of this report need to proceed to a seismic risk evaluation. Reference
[1] describes two different approaches for performing the seismic risk evaluation,
an SPRA, or an NRC SMA. The NRC SMA is appropriate for sites where the
re-evaluated seismic hazard is not considerably higher than the design basis
seismic hazard or for sites that have a relatively low seismic hazard level. The
g 6-6 h
SPRA could be used for any of the plants proceeding to the seismic risk
evaluation phase.
The NRC criteria for requiring the use of the SPRA consists of the following:
If the GMRS exceeds the response spectra between 1 and 10 Hz represented

by the higher of the following two spectra, then an SPRA should be
conducted:
1. 1.3 times the SSE
2. Low Hazard Threshold of 0.4g
Figure 6-4 shows an example of a GMRS exceeding the 1.3 SSE and the LHT
spectra in the 1 to 10 Hz range.
Figure 6-4
Example for Selection of SPRA vs. SMA
6.3 Key Elements of Seismic Structural and SSI

Response
6.3.1
Structure Modeling
Many existing structural models (i.e., those used for design basis, USI-A-46 or
IPEEE studies) could be used in structural dynamic analyses that are performed
to support SPRAs or SMAs required as part of the response to the 50.54(f) letter
on 2.1, provided that their adequacy is demonstrated for this purpose. This
requires a review of the existing models to be performed by an experienced
structural engineer(s) (and a peer reviewer) to determine the adequacy of the
g 6-7 h
models for dynamic analysis for application in risk assessments for 2.1 using the
criteria provided below. If necessary, the existing structural models can be
enhanced to meet the structural modeling criteria.
The existing structural models that have been used in dynamic analyses to
develop seismic responses for the design, licensing and qualification of plant
SSCs (e.g., lumped-mass stick models (LMSM)), were reasonably complex for
their original intended purpose at the time they were developed. These models
were used to capture the overall structural frequencies, mode shapes, and seismic
responses. Typically, if a model complexity is increased, the contribution of the
modes identified within the simpler model is decreased as modal mass is shifted
to other modes, often resulting in lower spectral peaks for the significant modes
of the structure. However, more recent experience has shown that, for some
structures, additional complexity of the numerical model may lead to the
identification of important higher modes that may be important for some systems
and components.
Using the existing structural models, in either their current or enhanced state,
will facilitate the completion of the SPRA/SMA effort with the desired accuracy
required as part of the response to the 50.54(f) letter on 2.1.
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
ly ng all
responses, one combined structural model should be used for analyzing
hq
three directions of the earthquake.
See ASCE 4-98 S
for Horizontal and Vertical
[40].
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
sp
eq
y range of interest. See ASCE 4-98 Section 3.1.4.1
response
in the frequency
Part (b) 1 [40].
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 will generally not be considered because it is expected to have
frequencies above 15 Hz, but this should be verified by the structural
engineer). This will ensure that the seismic responses and in-structure
response spectra (ISRS) developed in the 1 to 10 Hz frequency
qu cy range
ge are
reasonably accurate. See ASCE 4Part (b) 2 [40].
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
gi ty should
of rigidity may not be coincident at all levels, and the torsional rigidity
Lumpedbe computed. See ASCE 4-98 S
g 6-8 h
Parts (b) and (c) [40]. Alternatively, a multiple LMSM

mass Stick M
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 assess whether or not one-stick model sufficiently
represents the structure. For example, two-stick models could be more
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
the purposes of 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 inplane diaphragm flexibility may need to be included in the LMSM. The use
of this 1.5 aspect ratio should be reviewed by the structural engineer since
some structures are affected by the in-plane diaphragm flexibility by aspect
ratios lower than the 1.5. As with all these recommendations, alternate
approaches can be used when justified.
The use of existing models must also be justified in the submission to the NRC
using the above criteria.
6.3.2
Seismic Response Scaling
Scaling of ISRS to account for higher ground motions levels is considered a

technically sound approach and has been used in previous SPRAs and SMAs.
Using scaling approaches, where appropriate, will reduce the effort involved in
performing detailed soil-structure interaction (SSI) analyses for the new
hazard/UHS, facilitating the completion of the SPRA or SMA effort for those
plants that are screened-in.
Scaling of responses will be based on
previously developed ISRS,
shapes of the previous UHS/RLE,
shapes of the new UHS/RLE, and
structural natural frequencies, mode shapes, and participation factors.
Example guidance on scaling methods is provided in EPRI documents EPRI

NP-6041-SL Rev. 1 [39] and EPRI 103959 [38].
Scaling can be used in developing ISRS for those cases where the new UHS or
RLE shape is approximately similar to the spectral shape previously used to
generate the ISRS. An example of two response spectra with similar shapes is
shown below in Figure 6-5.
g 6-9 h
Figure 6-5
Example of Ground Response Spectra that are Similar
Scaling of rock or soil sites where the shape of the new hazard spectrum is not
similar to the previous spectrum will require a rigorous justification that
demonstrates the validity of the scaling approach. An example of spectra that are
not similar is shown in Figure 6-5 below. The peak response of these two spectra
is significantly shifted in frequency.
Figure 6-6
Example of Ground Response Spectra that are not Similar
Scaling
es would need to be technically justified on a case
specific basis and would need to appropriately consider any non-linear effects to
the structure or to the soil/rock profile resulting from the new response spectra
shape and amplitude.
g 6-10 h
6.3.3
Fixed-Based Analysis of Structures for
Sites Previously Defined as Rock
For nuclear safety-related structures founded on the commonly used definition of
rock as defined in the design documentation for many operating plants, i.e., shear
wave velocity (Vs) > about 5000 ft/sec, past experience has shown that the
amplified response spectra in the 1 to 10 Hz are generally about the same from a
fixed-based analysis of the structure as from a model that uses SSI analysis.
Therefore, it is reasonable to use fixed-base dynamic analyses for rock-founded
structures even when the rock shear wave velocities are not as high as 9200 ft/sec,
which is the definition of hard rock for new reactors licensed by the NRC. The
original definition of rock (Vs > about 5000 ft/sec) that was used by some plants
in the past can still be used as the criterion for performing a fixed-base dynamic
analysis to develop ISRS that are needed to perform fragility or HCLPF
calculations.
The validity of the above criterion was reviewed using two examples of existing
structures at a nuclear power plant [55]. The first example describes the analysis
of a structure with a fundamental frequency of about 5 Hz, and the second
example used a structure with a fundamental frequency of about 10 Hz in one
horizontal direction. These examples considered fixed-base analysis and SSI
analyses with different Vs values and are discussed in Appendix C. The results
from this study show that there is a slight shift of frequency to the left, and some
changes in spectral peak amplitudes occur when the fixed base is compared to an
SSI analysis with Vs of about 3500 ft/second; however, the comparison of fixedbase analysis is much better with an SSI analysis using Vs of about 5000 ft/sec or
higher.
Therefore, it is appropriate to model a rock-founded structure as fixed base if the
best estimate of Vs is greater than about 5000 ft/sec. For structures founded on
rock with Vs between 3500 ft/sec and 5000 ft/sec, peak-broadening or peakshifting of the ISRS in fragility analyses can potentially alleviate the effect of a
slight frequency/amplitude shift between the SSI and fixed-base analyses. The
determination whether a fixed-base model can be used for rock sites with Vs
values in this range should be made by an experienced structural engineer and
justified in the submittal report to NRC. This assessment should also be peer
reviewed by an experienced structural engineer as part of the peer review process.
6.4 Key Elements of Fragility/Capacity for the
Resolution of NTTF Recommendation 2.1
6.4.1
Hybrid Approach for Fragilities
There are two well-known methods to calculate fragilities of SSCs for use in a
seismic PRA model [36, 37, 38, 39]. These are: (a) the Conservative
Deterministic Failure Margin (CDFM) approach [39] where the HCLPF
capacity is calculated first and the median capacity with an assumed (experiencebased) composite variability (typically in the 0.35 to 0.45 range) is then calculated
from the HCLPF; and (b) the fragility or separation of variables approach [36,
g 6-11 h
37, 38] where the median capacity is calculated, and the randomness and
uncertainty variabilities (logarithmic standard deviations) are then calculated in a
detailed manner for various parameters.
The CDFM approach for developing fragilities is a simpler method that can be
performed consistently by more analysts and is an acceptable approach for
generating fragilities within an SPRA for the majority of components for which a
less detailed assessment is necessary. Because only a handful of components are
risk-significant enough to justify the additional effort required by the separation
of variables method, the CDFM method can provide efficiencies in the overall
effort. Therefore, use of the CDFM approach is useful and beneficial for
calculating fragilities of SSCs for use in seismic PRAs conducted to address the
50.54(f) letter.
In the CDFM fragility approach (also referred to as the Hybrid Method), the 1%
failure probability capacity C1% is computed along with an estimate of the
composite logarithmic standard deviation C and its subdivision into random
variability R and uncertainty U, which, are used to estimate the corresponding
fragility curve. As noted in [51], typically C lies within the range of 0.3 to 0.6. In
fact, if all of the sources of variability discussed in [38] are appropriately
considered, it is not possible to obtain an estimated C less than approximately
0.3.
The Hybrid Method is based on the observation that the annual probability of
unacceptable performance (PF) for any SSC is relatively insensitive to C. This
annual probability (seismic risk) can be computed with adequate precision from
the CDFM Capacity (CCDFM) and an estimate of C. It is shown in [51] that the
computed seismic risk at = 0.3 is approximately 1.5 times that at = 0.4, while
at = 0.6 the computed seismic risk is approximately 60% of that at = 0.4.
Table 6-2 provides recommended values for C, R, U, and the ratio of the
median capacity C50% to the C1% capacity computed by the CDFM Method. The
recommended C values are based on Ref. [51] recommendations and on average
are biased slightly conservative (i.e., slightly low C on average). Because random
variability R is primarily due to ground motion variability, a constant R value of
0.24 is recommended irrespective of the SSC being considered. The
recommended U values are back-computed from the recommended C and R
values. The Beta values for Table 6-2 apply to fragilities tied to ground motion
parameters (e.g., PGA or Peak Spectral Acceleration at 5 Hz). Appendix D
contains a sensitivity study on the computed probability of failure, PF, to the
logarithmic standard deviation used in the hybrid method. The results of the
study in Appendix D demonstrates a lack of sensitivity of the computed seismic
risk exists over the full practical range of seismic hazard curve slopes.
g 6-12 h
Table 6-2
Recommended C, R, U, and C50%/C1% Values to Use in Hybrid
Method for Various Types of SSCs
Composite
C
Random
R
Uncertainty
U
C50%/C1%
Structures & Major

Passive Mechanical
Components Mounted
on Ground or at Low
Elevation Within
Structures
0.35
0.24
0.26
2.26
Active Components
Mounted at High
Elevation in
Structures
0.45
0.24
0.38
2.85
Other SSCs
0.40
0.24
0.32
2.54
Type SSC
Following the generation of the fragilities using the hybrid approach, the fragility
parameters are then used in the systems model to convolve with the hazard. For
those SSCs that are determined to be the dominant risk contributors or are risksignificant in the seismic accident sequences, estimates of median capacity (C50%)
and variabilities (u and r) should be done using the fragility or separation of
variables approach and then used in the integration.
6.4.2
High-Frequency Capacities
Section 3.4 discusses the treatment of high frequency responses of SSCs that may
result from new seismic hazards and/or GMRS shapes developed for CEUS
plants, particularly for the plants on rock sites. In general, the seismic testing
described in Section 3.4 will produce results that can be divided into three
categories. Some components will have high capacities when subjected to the
high frequency motions while others will have moderate capacities, and still
others may have low capacities.
The flow chart in Figure 6-7 provides a general process for performing the High
Frequency Confirmation in Figure 1-1, Step 3f. Components that demonstrate a
high capacity can be screened out from further evaluations. Components that
demonstrate a moderate or low capacity will require further evaluations. The
recommended relay evaluation methods are derived from the evaluation methods
used in the SQUG program [52].
For relays or contactors, some applications can be accepted by performing circuit
analysis to show that contact chatter does not lead to inappropriate plant
operation. Examples include relays that provide control room annunciations or
contactors that turn on or off motors. In the case of the annunciation, following
an earthquake, operators will reset control room annunciations and then evaluate
those annunciations that do not properly reset. In the case of motors, contact
chatter does not typically provide enough power to the motor to drive it to move.
These are examples where relay or contactor chatter is considered acceptable.
g 6-13 h
Other relay chatter affects can be resolved by Operator Actions. Examples of

this strategy include resetting lock-out or seal-in relays that lead to undesired
plant conditions. Care should be exercised to avoid overloading the Operators by
crediting too many Operator Actions.
High Frequency
(HF) Component
Testing
High
Capacity
Component
Acceptable
Credit Circuit
Analysis or
Operator Action
Component
HF Capacity
Low
Capacity
Medium
Capacity
High Frequency
Sensitive
Component List
Replace
Component
Perform Capacity
vs. Demand Eval
HF Demand
Estimate
Figure 6-7
Potentially High-Frequency Sensitive Component Screening
High frequency sensitive components with moderate or low capacities can also be
replaced with comparable components with higher seismic capacities.
In general, the above approach ensures the adequacy of potentially high
frequency sensitive components and specific demand to capacity comparisons or
calculations of their fragilities for inclusion in the SPRA logic models will
typically not be required.
If needed, a detailed capacity to demand evaluation can be done to show that the
seismic capacity of the component exceeds the seismic demand at the component
mounting location. The ability for utilities performing the High Frequency
Confirmation under Figure 1-1 Step 3f to perform high frequency capacity to
demand evaluations may be limited because of the difficulty in calculating a
reliable and accurate high frequency demand at the component mounting
locations. Research activities may be performed as part of the High Frequency
Program in Section 3.4 leading to estimation techniques for high frequency instructure and in-cabinet seismic demands adequate to perform component
capacity to demand evaluations. If needed, plants performing SPRAs can
calculate fragilities of potentially high frequency sensitive components using the
best available estimates of seismic demand.
g 6-14 h
6.4.3
Capacity-based Selection of SSCs for
Performing Fragility Analyses
Capacity-based criteria to determine the SSCs for which fragility analyses should
be conducted have been developed to provide uniform guidance to analysts
performing seismic PRAs (or margin analyses) and to ensure that proper focus is
given to those SSCs that have the potential to be risk-significant. These criteria
were developed as a result of a parametric/sensitivity study [42] that was
undertaken for that purpose. These criteria establish which SSCs will require
explicit calculation of fragility parameters for inclusion in the plant logic models.
SSCs with capacities above the screening level calculated using the criteria are
not expected to have significant impact on the result of the seismic PRA analyses,
the ranking of accident sequences, or the calculated sequence- or plant-level
seismic CDF or LERF values.
It is noted that a standard practice for seismic PRA practitioners has been to use
insights from logic models to determine the need for fragility calculations and to
prioritize SSCs. A preliminary SPRA plant logic model is developed even before
the fragility calculation effort begins. Screening or ranking of SSCs from this
preliminary SPRA logic model can be done by performing parametric sensitivity
analyses with assumed initial fragilities and ranges of fragility values. Those SSCs
that do not contribute significantly to the SCDF of an accident sequence may
not need detailed fragility calculations. These SSCs may be retained in the model
with a screening level capacity value which is described below.
Certain SSCs are inherently seismically rugged and consequently have a very low
probability of failing as a result of a seismic event, as shown in Figure 6-7.
Consistent with long-standing practice in seismic PRAs, seismic failure of such
SSCs need not be included in the PRA logic models. Exclusion of such SSCs
from the logic models does not affect the seismic CDF or the insights derived
from the seismic PRA. Guidance in industry documents [39, 41] is available for
identifying seismically rugged SSCs.
Other SSCs may be less rugged but would still have sufficient capacity such that
their failures would be unlikely to contribute significantly to the SCDF in a
seismic PRA. Screening criteria discussed below are developed for these SSCs.
Detailed fragility calculations are not warranted for SSCs that meet these criteria.
Figure 6-7 illustrates the use of screening level, which is applicable to the SSCs
in the middle box.
g 6-15 h
Figure 6-8
Capacity-based Criteria for Fragility Analyses
Based on the results of the sensitivity study conducted to develop this guidance
[42], the screening HCLPF value of SSCs for a site should be calculated by
convolving the fragility of a single element with the site-specific hazard curve
such that the seismic CDF is at most about 5E-7 per year. This can be done with
trial and error runs using a quantification code or with a spreadsheet with an
assumed composite variability (e.g., c= 0.4). Because each site will have a
different hazard curve, the screening HCLPF value for each seismic PRA needs
to be separately derived. An alternative criterion, equivalent to the above CDFbased HCLPF, is to screen SSCs that have a HCLPF capacity above about 2.5 x
GMRS. The results of the sensitivity study do not indicate that the screening
criteria would be different for soil and rock sites.
The results of the sensitivity analyses performed indicate that the recommended
screening HCLPF capacity derived from a CDF of 5E-7 is conservative for some
hazard curves; a more liberal criterion may be appropriate for some sites and can
be developed from an initial quantification of the logic model. Even though
certain SSCs can be screened-out from having to perform detailed fragility
calculations using the above criteria, their failure should be retained in the
seismic PRA logic model with their capacity equal to the screening level or at a
higher capacity level, if calculated, to allow for a more efficient ranking of
accident sequences. Retention of such failures will ensure that future changes or
sensitivities that could increase their importance are not overlooked and also
addresses the problem of a potential cumulative effect of screened out
components.
g 6-16 h
The results of the SPRA should be reviewed to determine whether or not an

SSC modeled at the screening level could be identified as a significant
contributor to CDF or to LERF sequences. If such an SSC is identified, then
detailed fragility calculations should be performed for that SSC using the
separation of variables method, and the quantification analysis should be rerun
with the new fragility values.
If a component modeled at the screening level is risk significant, then the
screening level has been set too low. One approach to assess whether the
screening level was set at a sufficiently high level is to use a screening level set to
three times the screening level originally used in the logic model for that
component, quantify the seismic CDF/LERF and ensure that the CDF/LERF
estimates are not reduced by more than 20%. It is likely that there will be several
SSCs initially modeled at the screening capacity level, and the above approach, if
performed for one component at a time, may be cumbersome. The procedure can
also be done for multiple SSCs modeled at the screening capacity level
simultaneously (i.e., set the screening level for all such SSCs to three times the
original level and ensure that the CDF/LERF estimates do not reduce by more
than 20%). If either of the estimates change by more than 20%, it may take some
effort to pinpoint which component(s) modeled at the screening level is/are risksignificant. The sensitivity analyses
ys in [42] showed that if the ratio of the
screening level CDF (i.e., 5Emuch
greater than about 3% to 4%, the cumulative impact of SSCs modeled at the
screening level is not expected to be significant (i.e., none of the screening level
SSCs is likely to be risk-significant). This 3% to 4% ratio can be used as a guide
by the PRA analyst to determine if the initial screening criterion for SSCs was
appropriate or the screening capacity level needs to be adjusted (up or down).
To implement the capacity-based screening criteria, engineers can review
previous calculations and reports (e.g., design basis, IPEEE, USI A-46 analyses,
shake-table tests, etc.) to determine and judge if the seismic capacity of a
component or structure for the new seismic hazard is such that no further
calculation of fragility parameters is warranted.
It is expected that the use of the above screening methods will reduce the scope
of the fragility or margin calculations required in the SPRA or SMA, and still
meet the objective of identifying and ranking safety-significant SSCs. It is noted
that, while the use of the above criteria is optional, engineers should not select a
low screening HCLPF level (such as 0.3g) that was used by some plants during
the IPEEEs. The above criteria are expected to result in sufficiently high
screening levels to minimize the surrogate SCDF contribution (modeled at the
screening level). Once the screening level is selected, the list of SSCs can be
ordered so that the ones with the highest SCDF impact are calculated first.
g 6-17 h
6.5 Key Elements of SPRA/SMA Scope and Plant Logic

Modeling
6.5.1
Evaluation of LERF
Enclosure 1 to the 50.54(f) letter recommends that a SPRA performed as the

evaluation for plants that do not screen based on the change in the assessed
seismic hazard address the frequency of large early release (LERF) in addition to
CDF. The LERF risk measure for post-core damage accident sequences is
evaluated in detail in the internal events PRA that is the foundation for the
SPRA. The ASME/ANS Standard [12] appropriately directs use of the internal
events LERF analysis as the starting point for the SPRA LERF assessment:
The approach to anyy external-events PRA typically

yp ly uses as its starting point
the internal--events model
dealing with CDF and the part dealing with LERF are used as starting
points.
Seismic-induced core damage events can contribute to LERF scenarios in three
ways:
1. Seismic-induced failures of systems important to containment performance
and to preventing a large early release may contribute to LERF.
2. Like other core-damage accidents that can lead to a large early release, some
portion of those initiated by a seismic event can contribute to LERF, even if
the scenario does not directly include a seismic challenge to containment.
3. Some accident scenarios that would not constitute large early releases could
effectively be considered to be such due to the potential that the seismic
event could delay or slow evacuation.
The appropriate treatment for these LERF contributors is addressed in the
subsequent subsections.
6.5.1.1
Seismic-induced Failures of Systems
Important to LERF
The starting point for the seismic LERF evaluation is the internal events LERF
model. In the internal events portion of the ASME/ANS PRA Standard, Table
2-2.8-9 identifies the LERF contributors considered depending on the
containment design. Table 6-3 lists the potential contributors from the Standard
and summarizes the recommended approach to addressing the LERF contributor
in the SPRA.
g 6-18 h
Table 6-3
Consideration of LERF Contributors in SPRA
LERF Contributor
Containment isolation failure
Treatment in SPRA
x Include containment (pressure boundary) in SEL
x Include containment isolation failure modes in SEL
x Consider containment isolation function in relay
chatter evaluation
Containment bypass
(a) Interfacing-systems
LOCA
x Consider possibility due to relay chatter
(b) Steam generator tube

rupture (SGTR)
x No significant seismic-induced impact
(c) Induced SGTR
Energetic containment failure

(a) High-pressure melt
ejection
(b) Hydrogen combustion
x Include hydrogen igniters and support systems in SEL

(c) Core-debris impingement

Steam explosion
Shell melt-through
Bypass of pressure suppression
x Consider as part of containment seismic fragility

evaluation
Venting of reactor pressure

vessel or containment
Rupture of isolation condenser

tube(s)
x Include isolation condenser tubes in SEL
Vacuum breaker failure
x Include vacuum breakers in SEL
Hydrodynamic loads under

severe accident conditions
Containment flooding
In-vessel recovery
Containment failure induced

by failure of reactor to scram
Structures, systems and components beyond those relevant to preventing core

damage must be considered for the LERF analysis. This requires that the seismic
equipment list (SEL) be expanded to include the following (based on
applicability to the particular containment design):
i.
The containment structure, including the containment pressure boundary

and pressure suppression function, as applicable.
g 6-19 h
ii.
Equipment that plays a role in containment isolation failure modes from

the internal event LERF model, including containment isolation valves,
containment airlocks relying on inflatable seals to maintain containment
integrity, and required support systems including backup air supplies.
iii.
Hydrogen igniters and required support systems (PWRs with ice

condenser containments and BWRs with Mark III containments).
iv.
The isolation condenser (IC) for BWRs that have them, including IC tube
failure modes.
v.
Containment vacuum breakers, as applicable.
The relay chatter evaluation should be expanded to address the following

functions related to seismic-induced LERF contributors, as applicable:
Containment isolation
Closure of ISLOCA isolation valves
Seismic-induced failure modes for contributors noted in Table 6-3 should be

evaluated and either screened due to high seismic capacity or evaluated by adding
appropriate seismic failure events to the LERF logic model.
Other LERF contributors can be addressed in the same manner as addressed in
the internal events LERF model.
6.5.1.2
Integration of LERF Model with Seismic Core
Damage Model
The seismic-induced core damage scenarios from the Level 1 SPRA should be
extended to address the potential for large early release. It is important to
preserve the potential dependencies between the Level 1 core damage sequences
sp
yp
ys
ys approach
pp
and the containment response
event tree. A typical
systems-analysis
is
to add seismicinternal-events event- and fault-tree models that are adapted from the internalevents-PRA Level 1 and LERF analysis. In this manner, both the random (nonseismic-induced) failures and the seismic-induced failures described in Section
6.5.1.1 can be accounted for.
6.5.1.3
Accidents
Definition of LERF Scenario for Seismic
An accident scenario constitutes a potential

large
p
ge early
ly release if it leads to
(i.e., they are not effectively scrubbed) and the release occurs before there has
been time for pr
protection of the pu
public in the vicinity of the plant (by evacuation
In the event of core damage due to a seismic event, it is possible that
implementation of the emergency plan could be delayed or impeded. For
example, evacuation routes might be disrupted due to seismic damage. Thus, a
g 6-20 h
scenario that would not lead to an early release for an internal initiating event
could be effectively early if the sequence resulted from an earthquake.
This consideration is noted in Section 5-1.3 of the ASME/ANS PRA Standard,
although there are no corresponding high-level or supporting requirements in the
Standard.
Addressing this consideration would require a site-specific examination of the
potential for seismic failures that could have an impact on the emergency plan
(e.g., on systems for notifying the public, on bridges and other structures that
could be in evacuation routes, etc.). Such an examination is beyond the scope of
the evaluation for NTTF Recommendation 2.1, which is focused on the seismic
capacities of plant systems and structures relative to updated estimates of seismic
hazard.
6.6
6.6.1
Comparison to ASME/ANS SPRA Standard and RG1.200

Background
The ASME/ANS PRA Standard [12] presents requirements that are intended to
specify the capability of a PRA in terms of its scope, level of detail and plant
specificity relative to the nature of the intended application for the risk
assessment. In any PRA performed for a risk-informed application, the intent is
that the analyses meet at least the minimum requirements that could be relevant
for the application, at the capability category corresponding to the nature of the
application. The application in this case is to gain an updated understanding of
the risk of seismic events at NPPs in light of new information about seismic
hazard. This includes developing a new or changed understanding of risk outliers
due to seismic events. For any future applications, the SPRA conducted for the
50.54(f) evaluation can be used if it is shown to meet the following Standard
requirements: it is updated to reflect the then plant configurations and
operations; it meets the specific update requirements of the Standard (e.g., SHAH1); and it is shown to be adequate for the intended application.
It is expected that these seismic evaluations will be performed with sufficient
detail and plant specificity to meet the requirements for Capability Category II,
wherever feasible. To meet Capability Category II, the PRA must generally
account for plant-specific configuration and design, and reflect plant-specific
data where doing so could affect the important risk contributors.
For this application, which is aimed at developing an improved understanding of
the impact of new seismic hazard estimates, screening approaches will be used to
limit the scope of detailed analyses for some specific elements of the seismic
PRA. Where more detailed analyses
level of
ly are essential to achieve an adequate
eq
or alternative measures will be taken (such as making plant changes to address
the impacts).
g 6-21 h
The requirements in the Standard are organized according to the three major
elements of a seismic PRA:
Seismic hazard analysis (element SHA in the Standard),
Seismic fragility analysis (element SFR), and
Seismic plant-response modeling (element SPR).
The supporting requirements for each of these three elements are discussed in the
next section in the context of the guidance provided in this document. For any
specific instances in which the simplifications employed to facilitate seismic
evaluations for use in responding to the 50.54(f) letter would not produce
analyses that meet Capability Category II, a justification is provided.
6.6.2
Comparison of 2.1 Seismic Approach to the
SPRA Standard
For this application, the requirements corresponding to Capability Category II of
the ASME/ANS PRA Standard will generally be applied in the performance of
elements of an SPRA. As noted above, exceptions to the Standard requirements
may be taken in limited cases. The intent of the Standard will be met. Each
supporting requirement has been reviewed against the technical approach
recommended in this document to assess the capability category that applies.
The ASME/ANS PRA Standard has recently been revised, and a new version
has been approved by the Joint Committee on Nuclear Risk Management, which
is responsible for its development. The new version has not yet been published.
To be of maximum benefit to users of either version of the PRA Standard, the
comparison has been made to the supporting requirements in both versions.
These are as follows:
ASME/ANS RA-Sa 2009, the currentlyy appr

approved version of the
and
Addenda B to ASME/ANS RA-Sa-2009, the version of the Standard that

has recently undergone
The NRC has endorsed Addendum A of the PRA Standard for use in regulatory
applications via Regulatory Guide 1.200, Revision 2 [43].
In some cases, Regulatory Guide 1.200 provides further clarification or
specification beyond the details in the PRA Standard. In the case of the
requirements for seismic PRA elements SFR and SPR, however, there are no
such clarifications in Regulatory Guide 1.200.
Comments on the guidance in this document relative to the supporting
requirements in the two versions of the PRA Standard are provided in Tables 6-4
through 6-6. Note that, while the content of Addendum B has been approved by
the cognizant Standards committee, it is still in the process of final editing and
g 6-22 h
preparation for publication. The comparisons in Tables 6-4 through 6-6 reflect
the most current text available for Addendum B at the time the tables were
prepared.
The PRA Standard includes both high-level requirements and more detailed
supporting requirements. In Addendum B, the high-level requirements reflect
relatively minor wording changes from Addendum A. Therefore, Tables 6-4
through 6-6 include the high-level requirements as they exist in Addendum B ,
with changes from Addendum A indicated.
This document does not provide explicit guidance for every element of a seismic
PRA, relying instead on reference to other resources. It therefore does not
address the technical details corresponding to every supporting requirement. In
those cases for which this document does not provide guidance, it is expected
that the seismic evaluation will incorporate analyses commensurate with
Capability Category II.
g 6-23 h
SHA-A2
CC I-II-III
CC I-II
g 6-24 h
As the parameter to characterize both hazard

and fragilities, USE the spectral accelerations,
or the average spectral acceleration over a
selected band of frequencies, or peak ground
acceleration.
As the parameter to characterize both hazard

and fragilities, USE the spectral accelerations,
or the average spectral acceleration over a
selected band of frequencies, or peak ground
acceleration.
CC III III
CC III III
Spectral accelerations will be used to

characterize seismic hazard, and are
transmitted for use in fragility evaluations.
PSHAs conducted under the SPID will meet
this requirement directly.
The PSHA is based on an evaluation of

available information on historical seismicity,
paleoseismic data on large-magnitude
recurrence rates, and ground motion models,
and is based on available information on each
is
modeled, including evaluation of
uncertainties, using a defined evaluation
process. Documentation will be included in
hazard report. Licensees in the
CEUS will be using the CEUS-SSC, which is
documented in NUREG-2115. PSHAs
conducted under the SPID will meet this
requirement directly.
In performing the probabilistic seismic hazard

analysis (PSHA), BASE it on, and MAKE it
consist of, the collection and evaluation of
available information and data, evaluation of
the uncertainties in each element of the
PSHA, and a defined process and
documentation to make the PSHA traceable.

analysis (PSHA), BASE it on, and MAKE it
consist of, the collection and evaluation of
available information and data, evaluation of
the uncertainties in each element of the
PSHA, and a defined process and
documentation to make the PSHA traceable.
Relevant Intent of
Guidance in SPID
SHA-A1
Requirements from Addendum B
The frequency of earthquake seismic ground motions at the site shall be based on a site-specific probabilistic seismic hazard analysis (existing or
new) that reflects the composite distribution of the informed technical community. The level of analysis shall be determined based on the
intended application and on site-specific complexity.
Requirements from Addendum A
HLR-SHA-A
Standard
Element
Table 6-4
Comparison of SPID Guidance to ASME/ANS PRA Standard Supporting Requirements: Element SHA
SHA-A5
SHA-A4
SHA--A3
Standard
Element
CC I-II-III
CC I-II-III
g 6-25 h
SPECIFY a lower-bound magnitude (or

probabilistically defined characterization of
magnitudes based on a damage parameter) for
use in the hazard analysis, such that
earthquakes of magnitude less than this value
are not expected to cause significant damage
to the engineered structures or equipment.
CC I-II-III
In developing the probabilistic seismic hazard

analysis results for use in accident sequence
quantification, whether they are characterized
by spectral accelerations, peak ground
accelerations, or both, EXTEND them to
large-enough values (consistent with the
physical data and interpretations) so that the
truncation does not produce unstable final
numerical results, such as core damage
frequency, and the delineation and ranking of
seismic-initiated sequences are not affected.
CC I-II-III
If spectral acceleration or average spectral

acceleration over a band of frequencies is used,
INCLUDE the response frequencies of SSCs
that are significant in the PRA results and
insights.
SPECIFY a lower-bound magnitude (or

probabilistically defined characterization of
magnitudes based on a damage parameter) for
use in the hazard analysis, such that
earthquakes of magnitude less than this value
are not expected to cause significant damage
to the engineered structures or equipment.
CC I-II-III
In developing the probabilistic seismic hazard

analysis results, whether they are characterized
by spectral accelerations, peak ground
accelerations, or both, EXTEND them to
large-enough values (consistent with the
physical data and interpretations) so that the
truncation does not produce unstable final
numerical results, such as core damage
frequency, and the delineation and ranking of
seismic-initiated sequences are not affected.
CC I-II-III
In the selection of frequencies to determine

spectral accelerations or average spectral
acceleration, CAPTURE the frequencies of
those structures, systems, or components, or a
combination thereof that are significant in the
PRA results and insights.
A lower-bound magnitude of 5.0 will be used

in the PSHA and is sufficient to capture
earthquakes that will cause damaging ground
motions. PSHAs conducted under the SPID
will meet this requirement directly.
The PSHA results will be extended to spectral

accelerations corresponding to annual
frequencies of exceedence of 1E-8 or less.
This should be sufficient to produce stable
core damage frequency calculations and to
identify and rank failure sequences. PSHAs
Spectral accelerations will be reported in

hazard calculations that capture spectral
frequencies from 0.1 Hz to 100 Hz. This is
sufficient to capture the frequencies of
structures, systems, and components. PSHAs
Relevant Intent of
Guidance in SPID
g 6-26 h
CC I-II
CC I-II
Inputs to the PSHA will be based on SSHAC

Level 3 studies (including the CEUS Seismic
Source Characterization model for sites in the
CEUS), which use available geological maps
(including tectonic province, crustal thickness,
and volcanic maps) and geophysical maps
(including gravity, heat flow, and crustal
velocity maps). Site response studies will use
geotechnical databases (including topographic
maps, site cross-sections, borehole data, and
seismic velocity measurements).
Seismological data used for the SSHAC Level
3 studies, including the CEUS Seismic Source
Characterization project, includes an updated
historical earthquake catalog that represents a
compilation of available earthquake catalogs
for the region. PSHAs conducted under the
SPID will meet this requirement directly.

analysis USE available or developed
geological, seismological, geophysical, and
geotechnical data that reflect the current state
of the knowledge and that are used by
experts/analysts to develop interpretations and
inputs to the PSHA.

analysis (PSHA), BASE it on available or
developed geological, seismological,
geophysical, and geotechnical databases that
reflect the current state of the knowledge and
that are used by experts/analysts to develop
interpretations and inputs to the PSHA.
Relevant Intent of
Guidance in SPID
SHA-B1
To provide inputs to the probabilistic seismic hazard analysis, a comprehensive up-to-date data shall be compiled that include database, including
geological, seismological, and geophysical data; local site topography; and surficial geologic and geotechnical site properties, shall be compiled. A
catalog of historical, instrumental, and paleoseismicity information shall also be compiled.
HLR-SHA-B
Standard
Element
SHA-B3
SHA--B2
Standard
Element
CC II-III
As a part of the database used, COMPILE a

catalog of historically reported, geologically
identified, and instrumentally recorded
earthquakes. USE reference [5-30]
requirements or equivalent.
CC I-II
ENSURE that the database and information

used are adequate to characterize all credible
seismic sources that may contribute to the
frequency of occurrence of vibratory ground
motion at the site, considering regional
attenuation of ground motions and local site
effects. If the existing probabilistic seismic
hazard analysis (PSHA) studies are to be used
in the seismic PRA, ENSURE that any new
data or interpretations that could affect the
PSHA are adequately incorporated in the
existing databases and analysis.
CC I-II
g 6-27 h
As part of the data used, INCLUDE an

appropriate existing catalog of historically
reported earthquakes, earthquakes reported
through geological investigations, and
instrumentally recorded earthquakes. USE
reference [5-30] requirements or equivalent.
CC I-II
ENSURE that the data and information used

are adequate to characterize all credible
seismic sources that may contribute
significantly to the frequency of occurrence of
vibratory ground motion at the site,
considering regional attenuation of ground
motions and local site effects. If the existing
probabilistic seismic hazard analysis (PSHA)
studies are to be used in the seismic PRA,
ENSURE that any new data or
interpretations that could affect the PSHA are
adequately incorporated in the existing data
and analysis.
Under the SSHAC Level 3 process, a catalog

of historically reported, geologically identified,
and instrumentally recorded earthquakes is
compiled. PSHAs conducted under the SPID
will meet this requirement.
Under the SSHAC Level 3 process, databases

and information are reviewed to ensure that
all credible seismic sources are evaluated.
Databases and information on ground motion
equations are also reviewed to ensure that all
credible ground motion models are evaluated.
Available site data are reviewed as inputs to
local site response calculations to ensure that
all data and interpretations are incorporated
into site response studies. Existing PSHA
studies will not be used. PSHAs conducted
under the SPID will meet this requirement
through the use of published models used as
inputs.
Relevant Intent of
Guidance in SPID
SHA-C2
CC I-II-III
CC I-II-III
g 6-28 h
ENSURE that if an expert elicitation process

is used to characterize the seismic sources then
it needs to be compatible with the level of
analysis discussed in Requirement HLRSHA-A, and USE a structured approach.
CC I-II-III
ENSURE that any expert elicitation process

used to characterize the seismic sources is
compatible with the level of analysis discussed
in Requirement HA-A, and FOLLOW a
structured approach.
CC I-II-III
In the probabilistic seismic hazard analysis,

EVALUATE sources of earthquakes that
have the potential to contribute significantly
to the probabilistic hazard at the site. MAKE
the identification and characterization of
seismic sources taking into account previous
compilations of seismic sources, on regional
and site geological and geophysical data,
historical and instrumental seismicity data,
and geological evidence of prehistoric
earthquakes.
A SSHAC Level 3 approach for seismic

sources is specified in the 50.54(f) letter to
ensure that experts familiar with the regions of
the study are contacted and elicited for
information regarding potential seismic
sources. Under the SSHAC Level 3 process,
a structured approach is followed, including
formal workshops and review of data and
documentation. PSHAs conducted under the
SPID will meet this requirement.
Under the SSHAC Level 3 process, all

potential sources of earthquakes that might
affect a PSHA are examined. The
identification and characterization of those
sources is based on geological and geophysical
data, historical and instrumental earthquakes,
crustal stress, and tectonic interpretations.
Available evidence of prehistoric earthquakes
in the CEUS is evaluated in defining
earthquake sources and their characteristics.
this requirement.

EXAMINE all potential sources of
earthquakes that affect the probabilistic
hazard at the site. BASE the identification
and characterization of seismic sources on
regional and site geological and geophysical
data, historical and instrumental seismicity
data, the regional stress field, and geological
evidence of prehistoric earthquakes.
Relevant Intent of
Guidance in SPID
SHA-C1
To account for the frequency of occurrence of earthquakes seismic ground motions in the site region, the seismic sources with the hazard model
of the probabilistic seismic hazard analysis shall examine all credible sources of potentially damaging earthquakes. For the purpose of
characterizing the occurrence rates for those seismic sources, all earthquakes greater than magnitude 3 shall be considered. Both the aleatory and
epistemic uncertainties shall be addressed in characterizing the seismic sources.
HLR-SHA-C
Standard
Element
SHA-C4
SHA--C3
Standard
Element
CC I-II-III
g 6-29 h
CC I-II-III
If an existing seismic source model is used,

DEMONSTRATE that any new seismic
sources that have been identified or were
uncharacterized when the existing models
were developed are not significant, or
INCLUDE them in the update of the hazard
estimates.
CC II-III
CC II-III
If an existing probabilistic seismic hazard

analysis study is used, SHOW that any
seismic sources that were previously unknown
or uncharacterized are not significant, or
INCLUDE them in a revision of the hazard
estimates.
The seismic sources are characterized by

alternative source representations and source
geometry, maximum earthquake magnitude,
and earthquake recurrence. INCLUDE the
aleatory and epistemic uncertainties explicitly
in these characterizations, where significant.
The seismic sources are characterized by

source location and geometry, maximum
earthquake magnitude, and earthquake
recurrence. INCLUDE the aleatory and
epistemic uncertainties explicitly in these
characterizations.
Existing PSHA studies are not used in

characterizing seismic hazard. This
requirement is not applicable to PSHAs
conducted under the SPID.
Under the SSHAC Level 3 process, the

aleatory and epistemic uncertainties in seismic
sources are characterized for source location
and geometry, magnitude, and activity rate.
For each source, uncertainty in geometry and
location is represented with alternative
interpretations, uncertainties in recurrence
rates are represented with alternative rates,
and uncertainties in maximum magnitude are
represented with distributions of values. All
of these interpretations ensure that the
aleatory and epistemic uncertainties are
included explicitly in the characterizations.
this requirement.
Relevant Intent of
Guidance in SPID
CC I-II-III
g 6-30 h
CC I-II-III
(a) credible mechanisms governing estimates

of vibratory ground motion that can
occur at a site
(b) available historical and instrumental
seismicity data (including strong motion
data)
(c) current attenuation models for the
ground motion estimates
(a) credible mechanisms governing estimates

of vibratory ground motion that can
occur at a site
(b) regional and site-specific geological,
geophysical, and geotechnical data and
historical and instrumental seismicity
data (including strong motion data)
(c) current attenuation models in the ground
motion estimates
(a) For the CEUS, the Updated Ground

Motion Model project results to be used
in the PSHA do not account explicitly
for slip mechanism, because seismic
sources in the CEUS do not distinguish
between different slip mechanisms.
Therefore, explicit modeling of slip
mechanisms is not applicable to CEUS
sites. For the WUS, ground motion
prediction equations do take slip
mechanisms into account and will be
used for the PSHAs.
(b) Under the SSHAC Level 3 process,
empirical data for the region of interest
is used to evaluate published ground
motion models and derive modifications
and weights for those models. The
effect of local site conditions will
explicitly be taken into account by
modeling site amplification using a sitespecific profile. PSHAs conducted
under the SPID will meet this
requirement through the use of
published models used as inputs.
(c) Ground motion models published for
the WUS and CEUS, as appropriate,
will be reviewed and evaluated to develop
appropriate ground motion prediction
models for the PSHAs. PSHAs
In the vibratory ground motion analysis,

INCLUDE
ACCOUNT in the probabilistic seismic

hazard analysis for
Relevant Intent of
Guidance in SPID
SHA-D1
The probabilistic seismic hazard analysis shall examine credible mechanisms influencing estimates of vibratory ground motion that can occur at a
site given the occurrence of an earthquake of a certain type (e.g., strike slip, normal, reverse),magnitude, and at a certain location. Both the
aleatory and epistemic u Uncertainties shall be addressed in characterizing the ground motion propagation.
HLR-SHA-D
Standard
Element
SHA-D3
SHA--D2
Standard
Element
CC II-III
CC II-III
g 6-31 h
INCLUDE both the aleatory and epistemic

uncertainties separately in the ground motion
characterization in accordance with the level
of analysis identified for Requirement HLRSHA-A.
CC I-II-III

used to characterize the ground motion or any
other elements of the ground motion analysis
is compatible with the level of analysis
discussed in Requirement HLR-SHA-A, and
USE a structured approach.
ADDRESS both the aleatory and epistemic

uncertainties in the ground motion
characterization in accordance with the level
of analysis identified for Requirement SHAA.
CC I-II-III

used to characterize the ground motion is
compatible with the level of analysis discussed
in Requirement SHA-A, and FOLLOW a
structured approach.
Both aleatory and epistemic uncertainties in

ground motion characterization have been or
will be addressed to develop ground motion
prediction equations. Multiple equations were
or will be derived, each with multiple aleatory
uncertainties, thereby capturing epistemic
uncertainty in ground motions.
For both the CEUS and WUS, the SSHAC

approach is being used to characterize the
ground motion. For the CEUS, the Updated
Ground Motion Model project followed a
SSHAC Level 2 approach for evaluating
ground motion equations, which was
appropriate for updating a prior ground
motion model. For both CEUS and WUS,
experts familiar with ground motions, for the
respective regions, were or will be contacted
and elicited to ensure that information
regarding ground motion data and equations
is properly characterized. A structured
approach has or will be followed, including a
formal questionnaire and documentation of
responses to the questionnaire for each expert
who has or will participate. PSHAs
requirement.

requirement.
Relevant Intent of
Guidance in SPID
SHA--D4
Standard
Element
CC I-II-III
CC I-II-III
g 6-32 h
If existing ground motion models are used,

DEMONSTRATE that new information not
previously used or unknown would not
significantly affect the PSHA results, or
INCLUDE it in the update of the hazard
estimates.
If an existing probabilistic seismic hazard

analysis study is used, SHOW that any
ground motion models or new information
that were previously unused or unknown are
not significant, or INCLUDE them in a
revision of the hazard estimates.
Existing PSHA studies will not be used to

characterize seismic hazard at any site. This
Relevant Intent of
Guidance in SPID
CC II-III
CC II-III
g 6-33 h
Aleatory and epistemic uncertainties in local

site response will explicitly be addressed in
calculating site response: uncertainties will be
included in shear-wave velocity for each sand,
clay, and soft-rock layer, in the thickness of
each sand and clay layer, in the depth to
bedrock, and in the shear modulus and
damping curves for each sand, clay, and softrock layer. PSHAs conducted under the
INCLUDE both the aleatory and epistemic

uncertainties in the local site response analysis.
ADDRESS both the aleatory and epistemic

uncertainties in the local site response analysis.
SHA-E2
Any surficial geological deposits, and the

geotechnical properties of subsurface geologic
strata, will be taken into account in calculating
the site amplification that might occur from
incoming rock motions at the base of a
sedimentary profile underlying each site.
CC II-III
CC II-III
The effects of site topography, surficial

geologic deposits, and site geotechnical
properties will be taken into account by
reviewing available licensing documents and
In the probabilistic seismic hazard analysis

INCLUDE the effects of site topography,
surficial geologic deposits, and site
geotechnical properties on ground motions at
the site.
ACCOUNT in the probabilistic seismic

hazard analysis for the effects of site
topography, surficial geologic deposits, and
site geotechnical properties on ground
motions at the site.
Relevant Intent of
Guidance in SPID
SHA-E1
The probabilistic seismic hazard analysis SHALL shall account for the effects of local site response.
HLR-SHA-E
Standard
Element
SHA-F2
CC I-II-III
CC I-II-III
g 6-34 h

INCLUDE appropriate sensitivity studies and
intermediate results to identify factors that are
important to the site hazard and that make
the analysis traceable.
CC II-III

INCLUDE appropriate sensitivity studies and
intermediate results to identify factors that are
important to the site hazard and that make
the analysis traceable.
CC II-III
In the final quantification of the seismic

hazard, INCLUDE uncertainties through a
family of hazard curves.
For the CEUS, the CEUS Seismic Source

Characterization project report includes
sensitivity results at 7 test sites that show the
effect on seismic hazard of alternative seismic
sources and their parameters. In addition, for
the CEUS, the Updated Ground Motion
Model project report includes sensitivity
results at the same 7 test sites that show the
effect on seismic hazard of alternative ground
motion models and their uncertainties. The
WUS licensees will also conduct sensitivity
studies as part of their SSHAC Level 3
studies. These sensitivity studies provide
identification and traceability from the seismic
source and ground motion inputs to the
The final quantification of seismic hazard will

include the propagation of both aleatory and
epistemic uncertainties by generating multiple
seismic hazard curves, each representing an
alternative combination of epistemic
uncertainties in input models and parameters,
and each with an epistemic weight. The
propagation of aleatory and epistemic
uncertainties will be displayed through fractile
hazard curves showing the effects of
alternative assumptions on uncertainty in
seismic hazard. PSHAs conducted under the
In the final quantification of the seismic

hazard, INCLUDE and DISPLAY the
propagation of both aleatory and epistemic
uncertainties.
Relevant Intent of
Guidance in SPID
SHA-F1
Uncertainties in each step of the hazard analysis shall be propagated and displayed in the final quantification of hazard estimates for the site.
The results shall include fractile hazard curves, median and mean hazard curves, and uniform hazard response spectra. For certain applications,
the probabilistic seismic hazard analysis shall include seismic source deaggregation and magnitude-distance deaggregation.
HLR-SHA-F
Standard
Element
SHA--F3
Standard
Element
CC II
g 6-35 h
(a) fractile and mean hazard curves for each

ground motion parameter considered in
the probabilistic seismic hazard analysis,
(b) uniform hazard response spectra
(a) fractile and mean hazard curves for each

ground motion parameter considered in
the probabilistic seismic hazard analysis,
(b) fractile and mean uniform hazard
response spectrum
CC II
CALCULATE the following results as a part

of the quantification process, compatible with
needs for the level of analysis determined in
(HLR-SHA-A):
DEVELOP the following results as a part of

the quantification process, compatible with
needs for the level of analysis determined in
(HLR-SHA-A):

(a) fractile and mean hazard curves for peak

ground acceleration and spectral
accelerations, for the spectral frequencies
at which seismic hazard is calculated.
(a) fractile and mean uniform hazard
response spectrum, anchored at the
spectral frequencies at which seismic
hazard is calculated and interpolated in
between.
The following results will be developed and

presented in the PSHA report:
hazard results. Explicit sensitivity results to

alternative site amplification models will not
be included for each site.
Relevant Intent of
Guidance in SPID
CC II
g 6-36 h
CC I-II-III
Existing PSHA studies will not be used to

characterize seismic hazard at any site. This
Use of existing studies allowed.
SHA-H1
CONFIRM that the basic data and

interpretations of any existing studies used
remain valid in light of established current
information, consistent with the requirements
in HLR-SHA-A through G, and
DESCRIBE the bases and methodology used.
When use is made of an existing study for probabilistic seismic hazard analysis purposes, it shall be confirmed that the basic data and
interpretations are still valid in light of current information, the study meets the requirements outlined in A through G above, and the study is
suitable for the intended application.
CC II
The final site spectral shape is based on the

site amplification calculations that take into
account site-specific soil properties. Thus the
final site spectral shape reflects site-specific
considerations. PSHAs conducted under the
HLRSHA-H
CC II
ENSURE that the spectral shape used in the

seismic PRA uses site-specific evaluations
performed for the probabilistic seismic hazard
analysis.
BASE the response spectral shape used in the

seismic PRA on site-specific evaluations
performed for the probabilistic seismic hazard
analysis. REFLECT or BOUND the sitespecific considerations.
Relevant Intent of
Guidance in SPID
SHA-G1
For further use in the seismic PRA, the spectral shape shall be based on a site-specific evaluation taking into account the contributions of
deaggregated magnitude-distance results of the probabilistic seismic hazard analysis. Broad-band, smooth spectral shapes, such as those
presented in NUREG/CR-0098 [5-5] (for lower-seismicity sites such as most of those east of the U.S. Rocky Mountains) are also acceptable if
they are shown to be appropriate for the site. The use of uniform hazard response spectra is also acceptable unless evidence comes to light that
would challenge these uniform hazard spectral shapes.
HLR-SHA-G
Standard
Element
SHA-I2
For those hazards not screened out,

INCLUDE their effect through assessment of
the frequency of hazard occurrence and the
magnitude of hazard consequences.
CC I-II-III
The effects of fault displacements should be

included in SPRAs if not screened out. The
effects of landslides, soil liquefaction, and soil
settlement should be included in the SPRA if
not screened out.
Screening evaluations for fault displacements

should be performed on a site-specific basis2.
Screening evaluations for landslides, soil
liquefaction, and soil settlement should be
performed as necessary in the SPRA.
g 6-37 h
Fault displacements significant enough to be considered would be identified in the plant UFSAR. If significant fault displacements are not
identified in the plant UFSAR, that should be considered an adequate methodology for screening them out for SPRA impacts.
Not included in Addendum A
CC I-II-III
DOCUMENT the basis and methodology

used for any screening out of the seismic
hazards other than the vibratory ground
motion.
(There are no supporting requirements here.)
Relevant Intent of
Guidance in SPID
SHA-I1
A screening analysis shall be performed to assess whether in addition to the vibratory ground motion, other seismic hazards, such as fault
displacement, landslide, soil liquefaction, or soil settlement, need to be included in the seismic PRA for the specific application. If so, the
seismic PRA shall address the effect of these hazards through assessment of the frequency of hazard occurrence or the magnitude of hazard
consequences, or both.
HLR-SHA-I
Standard
Element
SHA-J2
(a) the specific methods used for source

characterization and ground motion
characterization,
(b) the scientific interpretations that are the
basis for the inputs and results, and
(c) if an existing PSHA is used,
documentation to ensure that it is
adequate to meet the spirit of the
requirements herein.
CC I-II-III
(a) the specific methods used for source

characterization and ground motion
characterization,
(b) the scientific interpretations that are the
basis for the inputs and results, and
(c) if an existing PSHA is used,
documentation to ensure that it is
adequate to meet the spirit of the
requirements herein.
CC I-II-III
g 6-38 h
DOCUMENT the process used in the

probabilistic seismic hazard analysis. For
example, this documentation is typically
consistent with reference [5-28] and includes
a description of:

probabilistic seismic hazard analysis. For
example, this documentation is typically
consistent with reference [5-28] and includes
a description of:
CC I-II-III
CC I-II-III
(a) For seismic source and ground motion

characterization, the CEUS-Seismic
Source Characterization and Updated
Ground Motion Model projects
documented the ranges of applicable
models and data, to characterize
epistemic uncertainties. PSHAs
(b) The CEUS-Seismic Source
For the CEUS, the process used for

developing the CEUS seismic sources and
ground motion equations has been
documented with references to supporting
documents, experts contacted (with
questionnaires and responses), and
explanation/justification of alternative models
and weights. Specifically:
For the WUS, seismic hazards will be

developed using the SSHAC Level 3 process,
which is expected to meet the applicable
documentation process.
The PSHA will be been documented with

plots, tables of digital values, and supporting
electronic files to facilitate PRA applications,
upgrades, and peer review. PSHAs conducted
under the SPID will meet this requirement
directly.
DOCUMENT the probabilistic seismic

hazard analysis in a manner that facilitates
PRA applications, upgrades, and peer review.
DOCUMENT the probabilistic seismic

hazard analysis in a manner that facilitates
PRA applications, upgrades, and peer review.
Relevant Intent of
Guidance in SPID
SHA-J1
Documentation of the probabilistic seismic hazard analysis shall be consistent with the applicable supporting requirements.
HLR-SHA-J
Standard
Element
Standard
Element
g 6-39 h
Characterization and Updated Ground

Motion Model projects documented the
scientific interpretations (including
tectonic models, geology, seismology,
and empirical ground motion data) that
were used to select applicable seismic
sources and ground motion equations,
and to reflect the range of interpretations
that is scientifically justified.
(c) Existing PSHA studies will not be used
to characterize seismic hazard at any site.
This requirement is not applicable to
PSHAs conducted under the SPID.
Relevant Intent of
Guidance in SPID
SHA--J3
Standard
Element
CC I-II-III
CC I-II-III
g 6-40 h
DOCUMENT the sources of model

uncertainty and related assumptions associated
with the probabilistic seismic hazard analysis.

with the probabilistic seismic hazard analysis.
(a) Alternative RLME and background

seismic sources reflect different
interpretations on crustal structure and
on the homogeneity of that structure.
PSHAs conducted under the SPID will
meet this requirement through the use of
(b) Alternative parameters for RLME and
background seismic sources reflect
alternative interpretations and
assumptions on homogeneity and
possible earthquake size. PSHAs
(c) Alternative ground motion models
reflect uncertainties on seismic ground
motion in the CEUS. PSHAs
(d) Uncertainties in site profile parameters
reflect uncertainties in shear-wave
velocities, modulus curves, and damping
curves appropriate for sites where direct
data and measurements are limited.
PSHAs conducted under the SPID will
meet this requirement directly.
The CEUS-Seismic Source Characterization

and Updated Ground Motion Model projects
documented the sources of model and
parameter uncertainties as follows:
Relevant Intent of
Guidance in SPID
CC I II-III
g 6-41 h
CC I-II-III
CALCULATE seismic fragilities for SSCs

identified by the systems analysis (see
Requirement SPR-D1).
The SPID does not provide guidance for

identifying SSCs relevant to the seismic
evaluation, but relies on other guidance. The
intent is that the seismic evaluation be
consistent with this requirement.
No substantive change from Addenda A to

Addenda B. A note in the Standard refers to
the screening criteria in HLR-SFR-B. Using
these criteria, some SSCs do not require
explicit fragility calculations, and others may
be assessed using screening values. Note that
Section 6.4.3 of this document provides
explicit guidance for screening SSCs
recognized to be very rugged and for
performing more limited fragility calculations
for certain other SSCs.
DEVELOP seismic fragilities for all those

structures, systems, or components, or a
combination thereof identified by the systems
analysis (see Requirement SPR-D1).
Relevant Intent of
Guidance in SPID
SFR-A1
The seismic-fragility evaluation shall be performed to estimate plant-specific, realistic seismic fragilities of SSCs structures, or systems, or
components, or a combination thereof whose failure may contribute to core damage or large early release, or both.
HLRSFR--A
Standard
Element
Table 6-5
Comparison of SPID Guidance to ASME/ANS PRA Standard Supporting Requirements: Element SFR
SFR-A2
Standard
Element
CC II
BASE the seismic fragilities on plant-specific

data, and ENSURE that they are realistic
(median with uncertainties). Generic data
(e.g., fragility test data, generic seismic
qualification test data, and earthquake
experience data) MAY be used for screening
of certain structures, systems, or components,
or a combination thereof and for calculating
their seismic fragilities by applying the
requirements under (HLR-SFR-F), which
permits use of such generic data under
specified conditions. However,
DEMONSTRATE that any use of such
generic data is conservative.
CC II
g 6-42 h
CALCULATE the seismic fragilities based

on plant-specific data, and ENSURE that
they are realistic (median with uncertainties).
Generic data (e.g., fragility test data, generic
seismic qualification test data, and earthquake
experience data) MAY be used for screening
of certain SSCs and for calculating their
seismic fragilities by applying the
requirements under (HLR-SFR-F), which
permits use of such generic data under
specified conditions. However,
DEMONSTRATE that any use of such
generic data is applicable.
The guidance in the SPID is consistent with

this requirement. It entails applying simplified
or generic values as a screening measure, and
performing detailed calculations for SSCs that
are important to risk (refer to Section 6.4).
Relevant Intent of
Guidance in SPID
SFR-B2
CC I-II-III
ASSESS and DOCUMENT the applicability

of the screening criteria given in EPRI NP6041-SL, Rev. 1 [5-3] and NUREG/CR4334 [5-4] for the specific plant and specific
equipment.
CC I-II
g 6-43 h
CC I-II-III
Deleted.
CC I-II
If screening of high-seismic-capacity
components is performed, DESCRIBE the
basis for screening and supporting documents
and SELECT the screening level high enough
that the contribution to core damage
frequency and large early release frequency
from the screened-out components is not
significant.
The guidance for screening in the SPID does

not explicitly address these documents, but the
approach is consistent with the guidance they
provide.
The deletion of this supporting requirement

in Addendum B is consistent with the general
nature of the PRA Standard.
The guidance in the SPID is consistent with

this requirement. Screening criteria are
addressed in detail in Section 6.4.3.
Reference to NP-6041 and NUREG/CR4334 has been removed in Addendum B but

retained in a note, and its removal from the
supporting requirement does not constitute a
substantive change.
If screening of high-seismic-capacity
components is performed, DESCRIBE fully
the basis for screening and supporting
documents. For example, it is acceptable to
apply guidance given in EPRI NP-6041-SL,
Rev. 1, and NUREG/CR-4334 to screen out
components with high seismic capacity.
However, CHOOSE the screening level high
enough that the contribution to core damage
from the screened-out components is not
significant.
Relevant Intent of
Guidance in SPID
SFR-B1
If screening of high-seismic-capacity components is performed, the basis for the screening shall be fully described (HLR-SFR-B).
HLRSFR-B
Standard
Element
SFR-C2
CC I-II
CC I-II
g 6-44 h
If probabilistic response analysis is performed

to obtain structural loads and floor response
spectra, ENSURE that the number of
simulations done (e.g., Monte Carlo
simulation and Latin Hypercube Sampling) is
large enough to obtain stable median and 85%
nonexceedance responses. INCLUDE the
entire spectrum of input ground motion levels
displayed in the seismic hazard curves.
If probabilistic response analysis is performed

to obtain realistic structural loads and floor
response spectra, ENSURE that the number
of simulations done (e.g., Monte Carlo
simulation and Latin Hypercube Sampling) is
large enough to obtain stable median and 85%
nonexceedance responses. ACCOUNT for
the entire spectrum of input ground motion
levels displayed in the seismic hazard curves.
CC I-II
CC I-II
This consideration is not addressed explicitly

in the SPID.
Addendum B made minor changes.
The guidance provided conforms to each of

the aspects of Supporting Requirement SFRC1. A characterization of site-specific
earthquakes is being provided for each NPP
site as described in Section 2 of this
document. The guidance does not provide
specific recommendations as to use of peak
ground acceleration (PGA) or average spectral
acceleration. It is left to the SPRA analysts to
assure appropriate alignment with the fragility
parameter.
Guidance for seismic-response analyses, and

particularly for determining whether existing
analyses are adequate, is provided in Section
6.3.2.
ESTIMATE the seismic responses that the

components experience at their failure levels
on a realistic basis using earthquake response
spectra in three orthogonal directions,
anchored to a ground motion parameter such
as peak ground acceleration or average spectral
acceleration over a given frequency band or
ENSURE that the spectral shape used bounds
the site-specific conditions.
ESTIMATE the seismic responses that the

components experience at their failure levels
on a realistic basis using site-specific
earthquake response spectra in three
orthogonal directions, anchored to a ground
motion parameter such as peak ground
acceleration over a given frequency band.
ENSURE that the spectral shape used reflects
or bounds the site-specific conditions.
Relevant Intent of
Guidance in SPID
SFR-C1
The seismic-fragility evaluation shall be based on realistic a seismic response that SSCs the structures, or systems, or components, or a
combination thereof experience at their failure levels.
HLRSFR-C
Standard
Element
SFR-C4
SFR-C3
Standard
Element
CC I-II
CC I-II
g 6-45 h
When the existing response analysis models

are judged not to be realistic and state of the
art, or when the existing input ground motion
is significantly different from the site-specific
input motion, PERFORM new analysis to
obtain realistic structural loads and floor
response spectra for use in the seismic PRA.
CC I-II
CC I-II
When the design response analysis models are

judged not to be realistic and state of the art,
or when the design input ground motion is
significantly different from the site-specific
input motion, PERFORM new analysis to
obtain realistic structural loads and floor
response spectra.
If scaling of existing response analysis is used,

JUSTIFY it based on the adequacy of
structural models, foundation characteristics,
and similarity of input ground motion.
If scaling of existing design response analysis

is used, JUSTIFY it based on the adequacy of
structural models, foundation characteristics,
and similarity of input ground motion.
The guidance in Section 6.3 of the SPID is

consistent with these requirements, except (as
noted above), the response analysis need not
be that used in the plant design. Therefore,
this supporting requirement should be
satisfied for Addendum B, and the intent is
satisfied for Addendum A.
This document provides guidance on how to

characterize whether the existing seismic input
motion is significantly different from the new
estimate. It also provides guidance on
determining whether an existing structural
model is realistic and reflective of the state of
the art. This guidance meets the intent
Capability Category II.
condition that new analysis should be
The changes in Addendum B were to replace
Explicit guidance for scaling of response

analysis is provided in Section 6.3.3. This
guidance provides an appropriate technical
approach for Capability Category II. The
guidance also addresses the use of response
analysis other than that performed for design.
Therefore, this supporting requirement should
be satisfied for Addendum B, and the intent is
satisfied for Addendum A.
The only change in Addendum B was to
Relevant Intent of
Guidance in SPID
SFR-C6
SFR-C5
Standard
Element
CC I-II
CC I-II
g 6-46 h
When soil-structure interaction (SSI) analysis

is conducted, ENSURE that it is median
centered using median properties, at soil strain
levels corresponding to the input ground
motions that contribute most to the
seismically induced core damage frequency.
INCLUDE the uncertainties in the SSI
analysis.
CC I-II
CC I-II
When soil-structure interaction (SSI) analysis

is conducted, ENSURE that it is median
centered using median properties, at soil strain
levels corresponding to the input ground
motions that dominate the seismically induced
core damage frequency. ACCOUNT for the
uncertainties in the SSI analysis by varying the
low strain soil shear modulus between the
median value times (1 + Cv) and the median
value divided by (1 + Cv), where Cv is a factor
that accounts for uncertainties in the SSI
analysis and soil properties. If adequate soil
investigation data are available, ESTABLISH
the mean and standard deviation of the low
strain shear modulus for every soil layer. Then
ESTABLISH the value of Cv so that it will
cover the mean plus or minus one standard
deviation for every layer. The minimum value
of Cv is 0-5. When insufficient data are
available to address uncertainties in soil
properties, ENSURE that Cv is taken as no
less than 1.0.
If median-centered response analysis is

performed, ESTIMATE the median response
(i.e., structural loads and floor response
spectra) and variability in the response using
established methods.
If median-centered response analysis is

performed, ESTIMATE the median response
(i.e., structural loads and floor response
spectra) and variability in the response using
established methods.

in the SPID.
Addendum A requirement.
Addendum B changes the action verb from

in the SPID.
No change in Addendum B.
Relevant Intent of
Guidance in SPID
SFR-D2
CC I-II-III
EXAMINE all relevant failure modes of

structures (e.g., sliding, overturning, yielding,
and excessive drift), equipment (e.g.,
anchorage failure, impact with adjacent
equipment or structures, bracing failure, and
functional failure), soil (e.g., liquefaction,
slope instability, and excessive differential
settlement), and EVALUATE fragilities for
critical failure modes.
CC I-II-III
g 6-47 h
CC I-II-III
EVALUATE relevant failure modes

identified in SFR D1 and EVALUATE
fragilities for critical failure modes.
CC I-II-III
IDENTIFY realistic failure modes of

structures ( e.g., sliding, overturning, yielding,
and excessive drift), equipment (e.g.,
anchorage failure, impact with adjacent
equipment or structures, bracing failure, and
functional failure) and soil (i.e., liquefaction,
slope instability, and excessive differential
settlement) that interfere with the operability
of equipment during or after the earthquake
through a review of the plant design
documents and the walkdown.
It should be noted, however, that this

document does provide guidance regarding
liquefaction and soil failure relative to
performing the quality review of IPEEE
quality review for those plants using the IHS
in the screening process.

in the SPID.
Addendum B changed EXAMINE to

EVALUATE and
failure modes, and omitted explicit listing of
the failure modes (since they are already
presented in Supporting Requirement SFRD1.

in the SPID.
Addendum B added examples of failure

modes that needed to be addressed but did
not change the substance of the supporting
requirement.
IDENTIFY realistic failure modes of

structures and equipment that interfere with
the operability of equipment during or after
the earthquake through a review of the plant
design documents and the walkdown.
Relevant Intent of
Guidance in SPID
SFR-D1
The seismic-fragility evaluation shall be performed for critical failure modes of SSCs structures, or systems, or components, or a combination
thereof such as structural failure modes and functional failure modes identified through the review of plant design documents, supplemented as
needed by earthquake experience data, fragility test data, generic qualification test data, and a walk down.
HLRSFR-D
Standard
Element
SFR-E3
SFR-E2
CC I-II-III
CC I-II-III
g 6-48 h
If components are screened out during or

following the walkdown, DOCUMENT the
basis including any anchorage calculations that
justify such a screening.
CC I-II-III
CC I-II-III
If components are screened out during or

following the walkdown, DOCUMENT any
anchorage calculations and PROVIDE the
basis justifying such a screening.
qualifications, walkdown observations, and

conclusions.
DOCUMENT the walkdown procedures,
DOCUMENT the walkdown procedures,

walkdown team composition and its
qualifications, walkdown observations, and
conclusions.
CC I-II-III
CC I-II-III
Screening during the walkdown itself is not

addressed explicitly in the SPID. Note,
however, that Section 6.4.3 of this document
provides explicit guidance for screening SSCs
recognized to be very rugged and for
performing more limited fragility calculations
for certain other SSCs.
Wording changes in Addendum B account for

the need to document the basis for any
screening in addition to that based on
anchorage calculations.

in the SPID.

in the SPID.
CONDUCT a detailed walkdown of the

plant, focusing on equipment anchorage,
lateral seismic support, spatial interactions and
potential systems interactions (both structural
and functional interactions).
CONDUCT a detailed walkdown of the

plant, focusing on equipment anchorage,
lateral seismic support, spatial interactions and
potential systems interactions (both structural
and functional interactions).
Relevant Intent of
Guidance in SPID
SFR-E1
The seismic-fragility evaluation shall incorporate the findings of a detailed walkdown of the plant focusing on the anchorage, lateral seismic
support, and potential systems interactions.
HLRSFR-E
Standard
Element
SFR-E5
SFR-E4
Standard
Element
CC I-II-III
During the walkdown, EXAMINE potential

sources of interaction (e.g., II/I issues,impact
between cabinets, masonry walls, flammable
and combustion sources, flooding, and spray)
and consequences of such interactions on
equipment contained in the systems model.
CC I-II-III
During the walkdown, FOCUS on the

potential for seismically induced fire and
flooding.
g 6-49 h
CC I-II-III
During the walkdown, EVALUATE

potential sources of interaction (e.g., II/I
issues,impact between cabinets, masonry
walls, flammable and combustion sources,
flooding, and spray) and consequences of such
interactions on equipment contained in the
systems model.
CC I-II-III
During the walkdown, EVALUATE the

potential for seismically induced fire and
flooding by focusing on the issues described in
NUREG-1407.

in the SPID.
Addendum B changed the action verb.
The assessment of seismically-induced fire

and flooding is beyond the scope of the
evaluations for which guidance is provided in
the SPID.
Addendum B changed the action verb and

provided additional guidance by reference.
Relevant Intent of
Guidance in SPID
SFR-F2
CC I-II
For all structures, or systems, or components,

or a combination thereof (SSCs) that appear
in the dominant accident cut sets, ENSURE
that they have site-specific fragility parameters
that are derived based on plant-specific
information, such as anchoring and
installation of the component or structure and
plant-specific material test data. Exception:
JUSTIFY the use of generic fragility for any
SSC as being appropriate for the plant.
CC I-II
CC I-II
g 6-50 h
For SSCs that appear in the significant

accident sequences, ENSURE that they have
site-specific fragility parameters that are
derived based on plant-specific information,
such as anchoring and installation of the
component or structure and plant-specific
material test data. Exception: JUSTIFY the
use of generic fragility for any SSC as being
appropriate for the plant.
CC I-II
CALCULATE component seismic-fragility

parameters such as median capacity and
variabilities (logarithmic standard deviations
reflecting randomness and uncertainty) based
on plant-specific data or if necessary based on
earthquake experience data, fragility test data,
and generic qualification test data. Exception:
JUSTIFY the use of generic fragility for any
SSC as being appropriate for the plant.
Section 6.4 of the SPID provides guidance on

the performance of the fragility analysis that is
intended to result in plant-specific estimates
for risk-significant SSCs. This is intended to
satisfy this supporting requirement.
in the PRA Standard.
Addendum B changes the word
The SPID does not provide explicit guidance

on how to calculate fragility parameters, but
SPRA fragility analyses would be expected to
satisfy this requirement.
Section 6.4 of the SPID provides some

guidance on the performance of the fragility
analysis and suggests the use of the CDFM
method to limit the amount of detailed
analysis required.
reference to sources beyond plant-specific data

slightly.
Addendum B changes the action verb from
BASE component seismic-fragility

parameters such as median capacity and
variabilities (logarithmic standard deviations
reflecting randomness and uncertainty) on
plant-specific data supplemented as
appropriate by earthquake experience data,
fragility test data, and generic qualification
test data.
Relevant Intent of
Guidance in SPID
SFR-F1
The calculation of seismic-fragility parameters such as median capacity and variabilities shall be based on plant-specific data supplemented as
needed or if necessary on earthquake experience data, fragility test data, and generic qualification test data. Use of such generic data shall be
justified.
HLRSFR-F
Standard
Element
SFR-F4
SFR-F3
Standard
Element
CC I-II-III
g 6-51 h
CC I-II-III
CALCULATE seismic fragilities for SSCs

that are identified in the systems models as
playing a role in the large early release
frequency part of the seismic PRA. (See
Requirements SPR-A1 and SPR-A3).
CC II-III
CC II-III
DEVELOP seismic fragilities for structures,

or systems, or components, or a combination
thereof that are identified in the systems
models as playing a role in the large early
release frequency part of the seismic PRA.
(See Requirements SPR-A1 and SPR-A3).
CALCULATE seismic fragilities for relays

identified to be essential and that are included
in the systems-analysis model.
DEVELOP seismic fragilities for relays

identified to be essential and that are included
in the systems-analysis model.
This document provides guidance for

identifying the SSCs that could be relevant for
calculating LERF, and for which fragility
analyses might be needed.
No substantive change from Addenda A to

Addenda B.
Section 6.4 of the SPID provides guidance on

the evaluation of relays. This guidance should
be consistent with satisfying this supporting
requirement.
Addendum B changed the action verb.
Relevant Intent of
Guidance in SPID
g 6-52 h
CC I-II-III
CC I-II-III
The SPID identifies specific elements that

need to be documented, but it is expected that
the fragility analysis will be thoroughly
documented in a manner consistent with this
supporting requirement.
DOCUMENT the seismic fragility analysis in

a manner that facilitates PRA applications,
upgrades, and peer review.
DOCUMENT the seismic fragility analysis in

a manner that facilitates PRA applications,
upgrades, and peer review.
Relevant Intent of
Guidance in SPID
SFR-G1
Documentation of the seismic fragility evaluation shall be consistent with the applicable supporting requirements.
HLRSFR-G
Standard
Element
SFR-G3
SFR-G2
Standard
Element
CC I-II-III

with the seismic fragility analysis.
CC I-II-III

seismic fragility analysis. For example, this
typically includes a description of: (a) The
methodologies used to quantify the seismic
fragilities of structures, or systems, or
components, or a combination thereof,
together with key assumptions, (b) The
structure, or system, or component, or a
combination thereof (SSC) fragility values
that includes the method of seismic
qualification, the dominant failure mode(s),
the source of information, and the location of
the component, (c) The fragility parameter
values (i.e., median acceleration capacity, R
and U) and the technical bases for them for
each analyzed SSC, and (d) The different
elements of seismic-fragility analysis, such as
(1) the seismic response analysis, (2) the
screening steps, (3) the walkdown, (4) the
review of design documents, (5) the
identification of critical failure modes for each
SSC, and (6) the calculation of fragility
parameter values for each SSC modeled.
g 6-53 h
CC I-II-III
Deleted.
CC I-II-III

seismic fragility analysis. For example, this
typically includes a description of: (a) The
methodologies used to quantify the seismic
fragilities of SSCs, together with key
assumptions, (b) The structure, or system, or
component, or a combination thereof (SSC)
fragility values that includes the method of
seismic qualification, the dominant failure
mode(s), the source of information, and the
location of the component, (c) The fragility
parameter values (i.e., median acceleration
capacity, R and U) and the technical bases
for them for each analyzed SSC, and (d) The
different elements of seismic-fragility analysis,
such as (1) the seismic response analysis, (2)
the screening steps, (3) the walkdown, (4) the
review of design documents, (5) the
identification of critical failure modes for each
SSC, and (6) the calculation of fragility
parameter values for each SSC modeled..


No substantive change in Addendum B.
Relevant Intent of
Guidance in SPID
SPR-A3
SPR-A2
CC I-II-III
g 6-54 h
CC I-II-III
USE the accident sequences and the systems

logic model from the internal-event at-power
PRA model as the basis for the seismic PRA
model.
CC I-II-III
CC I-II-III
USE the event trees and fault trees from the

internal-event full-power PRA model as the
basis for the seismic event trees.
In the initiating-event selection process,

DEVELOP a hierarchy to ensure that every
earthquake greater than a certain defined size
produces a plant shutdown within the systems
model.
In the initiating-event selection process,

DEVELOP a hierarchy to ensure that every
earthquake greater than a certain defined size
produces a plant shutdown within the systems
model.
CC I II-III
CC I II-III
The SPID does not provide explicit guidance,

but it is expected that any SPRA would build
on the existing models from the internalevents PRA.
The change from Addendum A to Addendum

B makes the reference to the type of models
used more general.

relative to this aspect of the SPRA model, but
it would be expected that the SPRA would

regarding the general nature of the SPRA
model, but it would be expected that the
SPRA would satisfy this requirement.
ENSURE that earthquake-caused initiating

events that give rise to significant accident
sequences and/or significant accident
progression sequences are included in the
seismic-PRA system model using a systematic
process.
ENSURE that earthquake-caused initiating

events that give rise to significant accident
sequences and/or significant accident
progression sequences are included in the
seismic-PRA system model using a systematic
process.
Relevant Intent of
Guidance in SPID
SPR-A1
The seismic-PRA systems model SHALL include seismic-caused initiating events and other failures including seismic-induced SSC failures,
non-seismic-induced unavailabilities, and human errors that give rise to significant accident sequences and/or significant accident progression
sequences.
HLR-SPR-A
Standard
Element
Table 6-6
Comparison of SPID Guidance to ASME/ANS PRA Standard Supporting Requirements: Element SPR
SPR-A4
(Add. B)
(Add. B)
SPR-A5
SPR--A4
(Add. A)
Standard
Element
This supporting requirement is new to

Addendum B and not included in Addendum
A.
CC I-II-III
ENSURE that the PRA systems models

reflect earthquake-caused failures and
nonseismically induced unavailabilities and
human errors that give rise to significant
accident sequences or significant accident
progression sequences.
g 6-55 h
CC I-II-III
Under special circumstances based on the

judgment of the analyst, DEVELOP an ad
hoc systems model tailored especially to the
seismic-PRA configurations or issues being
modeled, instead of starting with the internalevents model and adapting it, as in SPR-A3.
If this approach is used, ENSURE that the
resulting model is consistent with the
internal-events systems model regarding plant
response and the cause-effect relationships of
the failures.
New SPR-A4 in Addendum B.
CC I-II-III
ENSURE that the PRA systems models

reflect earthquake-caused failures and
nonseismically induced unavailabilities and
human errors that give rise to significant
accident sequences or significant accident
progression sequences.
Becomes SPR-A5 in Addendum B.

but it is expected that any SPRA would
employ an appropriate approach, consistent
with this requirement, in applying any ad hoc
models.
This supporting requirement was added for

Addendum B.

but it is expected that any SPRA would build
appropriately incorporate both seismic and
non-seismic failures in the PRA model.
There is no change in the supporting

requirement from Addendum A to
Addendum B, except that the insertion of a
new SPR-A4 in Addendum B changes the
number for this requirement.
Relevant Intent of
Guidance in SPID
(3) success-criteria analysis;

(4) systems analysis;
(5) data analysis;
(6) human-reliability analysis;
(7) use of expert judgment.
CC I II-III
(3) success-criteria analysis;
(4) systems analysis;
(5) data analysis;
(6) human-reliability analysis;
(7) use of expert judgment.
When the Part 2 requirements are used, USE

the Capability Category designations in Part
2, and for consistency USE the same
Capability Category in this analysis.
g 6-56 h
(2) accident-sequence analysis;
(2) accident-sequence analysis;
CC I II-III
(1) initiating-event analysis;
(1) initiating-event analysis;

regarding following relevant elements of the
PRA Standard for internal initiating events,
but it would be expected that the SPRA
would satisfy these requirements.
There was a small wording change to remove
In each of the following aspects of the

seismic-PRA systems-analysis work,
SATISFY the corresponding requirements in
Part 2, except where they are not applicable or
where this Part includes additional
requirements. SPECIFY the basis to support
the claimed nonapplicability of any
exceptions. The aspects governed by this
requirement are
In each of the following aspects of the

seismic-PRA systems-analysis work,
SATISFY the corresponding requirements in
Part 2, except where they are not applicable or
where this Part includes additional
requirements. DEVELOP a defined basis to
support the claimed nonapplicability of any
exceptions. The aspects governed by this
requirement are
Relevant Intent of
Guidance in SPID
SPR-B1
The seismic-PRA systems model shall be adapted to incorporate seismic-analysis aspects that are different from corresponding aspects found in
the at-power full power, internal-events PRA systems model.
HLR-SPR-B
Standard
Element
SPR-B3
( Add. B)
SPR-B4
( Add. A)
SPR-B4a
(Add. B)
SPR-B3
( Add. A)
SPR--B2
Standard
Element
CC I-II
CC I-II
g 6-57 h
PERFORM an analysis of seismic-caused

correlations in a way so that any screening of
SSCs appropriately accounts for those
correlations. USE bounding or generic
correlation values and PROVIDE the basis
for such use.
CC I-II-III
If screening out on the basis of seismic

capacity is performed in the systems model,
PERFORM the screening on a documented
basis.
Deleted in A ddendum B and addressed by

SPR-B5 as follows.

regarding the treatment of seismic
correlations. The SPRA would be expected to
two places. The intent of the requirement did

not change.
A relatively minor change in Addendum B
Acceptable screening criteria are provided in

Section 6.4.3 of the SPID.
This requirement has been deleted in

Addendum B because it is covered by
Supporting Requirement SPR-B5.

regarding the conduct of the HRA for the
seismic PRA, but it would be expected that
the SPRA would satisfy this requirement.
(a) Additional post earthquake workload and

stress that can increase the likelihood of
human errors or inattention
(b) Seismic failures that impact access
(c) Cue availability
CC I II
The requirement related to HRA has been

completely re-written for Addendum B
(including defining a separate requirement for
Capability Category III). The updated
requirement clarifies what is expected of the
seismic HRA.
Relevant Intent of
Guidance in SPID
INCLUDE the following seismic impacts on

Performance Shaping Factors (PSFs) for the
control room and ex-control room postinitiator actions as appropriate to the Human
Reliability Analysis (HRA) methodology
used:
PERFORM an analysis of seismic-caused

dependencies and correlations in a way so that
any screening of SSCs appropriately accounts
for those dependencies and correlations. USE
bounding or generic correlation values and
PROVIDE the basis for such use.
CC I II-III
If any screening is performed, PERFORM it

using defined criteria that are documented in
the PRA.
CC I II-III
In the human reliability analysis (HRA)

aspect, EXAMINE additional postearthquake
stresses that can increase the likelihood of
human errors or inattention, compared to the
likelihood assigned in the internal-events
HRA when the same activities are undertaken
in nonearthquake accident sequences.
Whether or not increases in error probabilities
are used, JUSTIFY the basis for this decision
about what error rates to use.
In the systems-analysis models, for each basic

event that represents a significant seismically
caused failure, INCLUDE the
applicable to a particular SSC and SPECIFY
In the systems-analysis models, for each basic

event that represents a seismically caused
failure, INCLUDE the complementary
SSC.
SPR-B7
( Add. A)
CC I-II-III
CC I-II-III
CC II-III
SPR-B4
( Add. B)
SPR-B5
( Add. B)
INCLUDE the effects of the chatter of relays

and similar devices in the systems model.
EXAMINE the effects of the chatter of relays

and similar devices.
SPR-B6
( Add. A)
CC I-II
g 6-58 h
this activity.
CC I-II-III
If post-earthquake recovery actions are

included in the systems model, INCLUDE
them on a documented basis.
CC I-II-III
If screening out on the basis of seismic

capacity is performed in the systems model,
SPECIFY the screening criterion.
Deleted in Addendum B and replaced by

SPR-B4a as follows.
Not included in Addendum A.
CC I-II-III
ENSURE that any screening of human-error

basic events and non-seismic failure basic
events does not significantly affect the PRA
results.
SPR-B4b
(Add. B)
SPR-B4a
(Add. B)
SPR--B5
( Add. A)
Standard
Element
The SPID does not present explicit guidance

relative to this aspect of the systems models; it
is expected that a SPRA would satisfy this
requirement.
added for Capability Categories I and II in

Addendum B.
The supporting requirement from Addendum

A was retained in Addendum B for Capability
Section 6.4.2 of the SPID provides acceptable

guidance for the treatment of chattering due
to high frequency ground motions.

regarding the treatment of recovery, but it is
expected that any recovery actions would be
appropriate documented in the SPRA.
This is a new requirement in Addendum B.
The SPID provides acceptable guidance

regarding screening based on seismic capacity
in Section 6.4.3.
The requirement related to screening of nonseismic failures has been deleted, and a new
requirement was added in Addendum B
related to screening based on seismic capacity.
Relevant Intent of
Guidance in SPID
SPR-B8
( Add. B)
SPR-B10
(Add. A)
SPR-B7
( Add. B)
SPR-B9
( Add. A)
SPR-B6
( Add. B)
SPR--B8
( Add. A)
Standard
Element
CC I-II-III
each sequence in the seismic-PRA model.
EXAMINE the effect of including an

earthquake-small loss-of
CC II-III
CC I-II
g 6-59 h
ASSUME the existence of an earthquake-ofin

the SPRA accident sequences and system
modeling, unless it is demonstrated that such
a LOCA can be excluded, based on a
walkdown or on another examination of the
possible sources of such a LOCA.
CC II
EVALUATE the likelihood that system

recoveries modeled in the internal-events
PRA may be more complex or even not
possible after a large earthquake, and
ADJUST the recovery models accordingly. It
is acceptable to use generic or conservative
recovery values.
CC I-II-III
CC I-II-III
EXAMINE the likelihood that system

recoveries modeled in the internal-events
PRA may be more complex or even not
possible after a large earthquake, and
ADJUST the recovery models accordingly.
EVALUATE the possibility that a large

earthquake can cause damage that blocks
personnel access to safety equipment or
controls, thereby inhibiting operator actions
that might otherwise be credited.
EXAMINE the possibility that a large

earthquake can cause damage that blocks
personnel access to safety equipment or
controls, thereby inhibiting operator actions
that might otherwise be credited.

on how to evaluate very small loss-of-coolant
accidents, but the SPRA analyses would be
expected to satisfy this requirement.
The requirement was made more definitive in

Addendum B.

relative to this aspect of the treatment of
recovery; it is expected that a SPRA would
The action verb was changed from Addendum

A to Addendum B. Also, the acceptability of
using generic or conservative values was added
for Capability Category II.

operator response; it is expected that a SPRA
would satisfy this requirement.
The only change from Addendum A to

Addendum B was the change in action verb.
Relevant Intent of
Guidance in SPID
SPR-B9
( Add. B)
SPR--B11
(Add. A)
Standard
Element
CC I-II-III
In the seismic PRA walkdown, EXAMINE

the potential for seismically induced fires and
flooding following the guidance given in
NUREG-1407.
g 6-60 h
CC I-II-III
If the seismic PRA walkdown (see SFR-E4)

identifies the potential for seismically induced
fires and flooding, INCLUDE potential
significant contributions to accident sequences
in the systems model.

recovery; it is expected that the walkdowns
would include consideration of seismically
induced fire and flooding. Any detailed
analyses of scenarios associated with such
consequential events would, however, be
performed in a later assessment.
The action verb was changed from Addendum

A to Addendum B.
Relevant Intent of
Guidance in SPID
CC I-II-III
g 6-61 h
CC I-II-III

for developing the SEL, but it would be
expected that the SPRA would satisfy this
requirement.
and the consideration of

some structures was added in Addendum B.
USE the seismic PRA systems model as the

basis for developing the seismic equipment
list, which is the list of all SSCs to be
considered by the subsequent seismic-fragility
evaluation task.
SPR-D1
USE the PRA systems model as the basis for

developing the seismic equipment list to
support the fragility analysis of Section 5-2.2.
INCLUDE structures and passive
components that may not be present in the
internal events model but which require
consideration in the seismic PRA.
SUPPLEMENT the list based on review of
industry SPRA SELs (if available).
The list of SSCs selected for seismic-fragility analysis shall include the all SSCs that participate in accident sequences included in the seismicPRA systems model.

but it would be expected that the SPRA
would satisfy this requirement.
The language has been clarified in Addendum

B.
Relevant Intent of
Guidance in SPID
HLRSPR--D
CC I-II-III
CC I-II-III
To ensure that the systems-analysis model

reflects the as-built, as-operated plant,
JUSTIFY any conservatisms or other
distortions that do not adequately reflect the
as-built, as-operated plant.
To ensure that the systems-analysis model

reflects the as-built, as-operated plant,
JUSTIFY any conservatisms or other
distortions introduced by demonstrating that
the seismic-PRA's validity for applications is
maintained.
SPR-C1
The seismic-PRA systems model shall reflect the as-built and as-operated plant being analyzed.
HLR-SPR-C
Standard
Element
SPR-E4
SPR-E3
SPR-E2
CC I-II
CC I-II
g 6-62 h
In the integration/quantification analysis,

INCLUDE the significant correlations that
affect the results. It is acceptable to use
generic correlation values. If used, PROVIDE
the basis for such use.
CC I-II-III

for treating correlations. The SPRA would be
expected to satisfy this requirement.
The language was clarified in Addendum B.
The SPID provides acceptable guidance for

the screening process, and the validity of the
screening should be confirmed during the
quantification.
The language was clarified in Addendum B.

for performing the quantification, but this
integration is a typical task in SPRA, and the
SPRA would be expected to satisfy this
requirement.
CC I-II-III
USE the quantification process to confirm

and support the screening of SSCs (refer to
SFR-B1).
The language was revised in Addendum B to

refer to guidance in the internal-events
portion of the Standard.
PERFORM the quantification in accordance

with applicable requirements described in
subsection 2.2.7.

ACCOUNT for all significant dependencies
and correlations that affect the results. It is
acceptable to use generic correlation values. If
used, PROVIDE the basis for such use.
CC I-II-III
In the analysis, USE the quantification

process to ensure that any screening of SSCs
does not affect the results, taking into account
the various uncertainties.
CC I-II
In quantifying core damage frequency and

large early release frequency, PERFORM the
quantification on a cut-set-by-cut-set or
accident-sequence-by-accident-sequence basis
(or for defined groups of these), as well as on a
comprehensive/integrated basis. It is
acceptable to use broad groupings.
CC I-II-III
CC I-II-III

for performing the quantification, but this
integration is a typical task in SPRA, and the
requirement.
In the quantification of core damage

frequency and large early release frequency,
PERFORM the integration using the seismic
hazard, fragility, and systems analyses.
In the quantification of core damage

frequency and large early release frequency,
PERFORM the integration using the seismic
hazard, fragility, and systems analyses.
Relevant Intent of
Guidance in SPID
SPR-E1
The analysis to quantify core damage frequency and large early release frequency shall appropriately integrate the seismic hazard, the seismic
fragilities, and the systems-analysis aspects.
HLR-SPR-E
Standard
Element
SPR-E6
(Add. B)
(Add. A)
SPR-E7
SPR-E6
SPR--E5
Standard
Element
New requirement in Addendum B.
CC I-II-III
g 6-63 h
CC I-II-III
In the analysis of LERF, SATISFY the

LERF requirements in the internal-initiators
section (Part 2, Section 2.2.8), where
applicable.
Deleted in Addendum B
CC II
CC I-II
PERFORM appropriate sensitivity studies to

illuminate the sensitivity of the core damage
results to the assumptions used about
dependencies and correlations.

analysis, INCLUDE in the uncertainties in
core damage frequency and large early release
frequency results that arise from each of the
several inputs (the seismic hazard, the seismic
fragilities, and the systems-analysis aspects).
USE the mean hazard, composite fragilities,

and the systems analysis to generate point
estimates for core damage frequency (CDF)
and large early release frequency (LERF).
ESTIMATE the uncertainties in overall CDF
and LERF.
The SPID provides some guidance for

assessment of LERF in Section 6.5.1.
This requirement is entirely new for

Addendum B.

for sensitivity studies for the integrated model.
The SPRA would be expected to satisfy this
requirement.

for assessing uncertainty in the results, but the
requirement.
The requirement was re-written for

Addendum B, and separate requirements are
provided for each of the three capability
categories.
Relevant Intent of
Guidance in SPID

with the seismic plant response model
development.
CC I-II-III

with the seismic plant response model
development.
CC I-II-III
SPR-F3
g 6-64 h
CC I-II-III

seismic plant response analysis and
quantification.
Error in Addendum A corrected in the

supporting requirement in Addendum B.
SPR-F2
CC I-II-III
CC I-II-III
The SPID provides guidance for

documentation of some aspects of the seismic
PRA. The SPRA would be expected to satisfy
this requirement.

this requirement.

this requirement.
DOCUMENT the seismic plant response

analysis and quantification in a manner that
facilitates PRA applications, upgrades, and
peer review.
DOCUMENT the seismic plant response

analysis and quantification in a manner that
facilitates PRA applications, upgrades, and
peer review.
Relevant Intent of
Guidance in SPID
SPR-F1
The seismic-PRA analysis shall be documented in a manner that facilitates applying the PRA and updating it and that enables peer review.
HLR-SPR-F
Standard
Element
6.7
Peer Review
This section describes the peer review requirements for the activities performed
to meet the 50.54(f) letter [1] relative to the seismic 2.1 requests for information.
The peer review need not assess all aspects of the SPRA or SMA against all
technical requirements; however, enough aspects of the PRA shall be reviewed
for the reviewers to achieve consensus on the adequacy of methodologies and
their implementation for each PRA or SMA element. Alternative methods and
approaches to meet the intent of SPRA/SMA technical requirements may be
used if they provide results that are equivalent or superior to the methods usually
used, and it is expected that the peer review team should concentrate on
reviewing such alternate methods and approaches if they are used.
The peer review team shall have combined experience in the areas of systems
engineering, seismic hazard, seismic capability engineering, and other elements
of seismic PRAs or seismic margin methodologies. The reviewer(s) focusing on
the seismic response and fragility work shall have successfully completed the
SQUG Walkdown Screening and Seismic Evaluation Training Course [52] or
equivalent.
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 should also pay attention to potential issues
that could occur at the interface between various activities.
The peer review process includes a review of the following SPRA activities:
Selection of the SSCs included on the SEL
Seismic hazard assessment3
Documentation from the Seismic Walkdowns
Seismic response analyses
Seismic fragility assessments
Seismic risk quantification
Final report
The results of the peer review should be documented in a separate report.

Specific guidance on the key elements of the peer review process is found in
Section 5.3 of the SPRA part of the ASME-ANS PRA Standard [12] entitled
ew for Seismic Events AtThis guidance is felt to be
appropriate for this peer review, with the recommended exception that
independent seismic fragility analyses are not required to be performed by the
3
Seismic hazard assessments submitted to and reviewed by NRC as described in Section 4 would
not require additional reviews under the Peer Review. Additional hazard work performed for the
SPRA (e.g. FIRS estimates) should be reviewed by the Peer Review.
g 6-65 h
peer reviewers. Adequate peer review of the seismic fragilities can be

accomplished (as in past SPRAs and SMAs) based on a review of a sample of the
fragility calculations.
For the NTF 2.1 Seismic resolution, it may be preferable to conduct more
focused peer review activities for individual SPRA elements during
implementation of this program, to the extent pr
practical, rather than waiting until
should include the
all the work is complete. This following elements:
The SPRA findings should be based on a consensus process, and not based
on a single peer review team member
A final review by the entire peer review team must occur after the completion
of the SPRA project
independent throughout the SPRA development activity
The NRC [22] has provided an additional detailed description of the important
elements of the in-process peer review (in their comments to NEI 12-13 on
external event PRA peer review guidelines) which should be incorporated into
any such peer review for the 50.54(f) seismic 2.1 program.
6.8
SPRA Documentation
Documentation criteria for a SPRA are identified throughout the ASME/ANS

Standard [12]. Utilities are expected to retain that documentation consistent with
the Standard.
A report should be submitted to the NRC summarizing the SPRA inputs,
methods, and results. The level of detail needed in the submittal should be
sufficient to enable NRC to understand and determine the validity of all input
data and calculation models used, to assess the sensitivity of the results to all key
aspects of the analysis, to make necessary regulatory decisions as a part of NTTF
Phase 2 activities. It is not necessary to submit all of the SPRA documentation
for such an NRC review. Relevant documentation should be cited in the
submittal, and be available for NRC review in easily retrievable form.
Appendix C to NUREG 1407 [11] documents the detailed documentation and
reporting guidelines associated with the seismic PRA studies conducted as part of
the IPEEE program. The pertinent documentation and reporting for the seismic
IPEEE program are listed below and are intended to serve as guidance to
licensees on elements of the SPRA documentation and/or reporting that may
also be appropriate for this 50.54(f) response:
1. A description of the methodology and key assumptions used in performing
the SPRA.
2. The hazard curve(s) (or table of hazard values) used and the associated
spectral shape used in the analysis. Also, if an upper bound cutoff to ground
motion is assumed, the results of sensitivity studies to determine whether the
g 6-66 h
cutoff affected the overall results and the delineation and ranking of seismic
sequences.
3. A summary of the walkdown findings and a concise description of the
walkdown team and the procedures used. This summary should include any
potential for seismically induced fire or internal flood identified during the
walkdown.
4. All functional/systemic seismic event trees and a description of how nonseismic failures, human actions, dependencies, relay chatter, and soil
liquefaction are accounted for.
5. The estimated CDF and LERF, including a qualitative discussion of
uncertainties and how they might affect the final results, and contributions of
different ground motions to CDF and LERF.
6. A description of dominant functional/systemic sequences leading to core
damage along with their frequencies and percentage contribution to overall
seismic core damage frequencies. The description of the sequences should
include a discussion of specific assumptions and human recovery actions.
7. A table of fragilities used for screening as well as in the quantification. The
estimated fragilities for the plant, dominant sequences, and dominant
components should be reported where possible.
8. A discussion of important non-seismic failures and human actions that are
significant contributors, or have impacts on results, including the rational for
the assumed failure rate given a seismic event.
The 50.54(f) letter [1] requests that specific information be submitted by plants
performing a SPRA including the following.
A list of the significant contributors to SCDF and LERF, including

importance measures (e.g., Risk Achievement Worth, FussellVesely and/or
Birnbaum)
A summary of the methodologies used to estimate the SCDF and LERF,

and results obtained, including the following:
o
o
o
o
o
methodologies used to quantify the seismic fragilities of SSCs, together

with key assumptions
SSC fragility values with reference to the method of seismic
qualification, the dominant failure mode(s), and the source of
information
seismic fragility parameters
important findings from plant walkdowns and any corrective actions
taken
process used in the seismic plant response analysis and quantification,
including the specific adaptations made in the internal events PRA
model to produce the seismic PRA model and their motivation
assumptions about containment performance
g 6-67 h
Description of the process used to ensure that the SPRA is technically

adequate, including the dates and findings of any peer reviews
Identified plant-specific vulnerabilities and actions that are planned or taken.
g 6-68 h
Section 7: Spent Fuel Pool

Integrity Evaluation
The 50.54(f) letter requested that, in conjunction with the response to NTTF
Recommendation 2.1, a seismic evaluation be made of the SFP. More
specifically, plants were asked to consider
all seismically induced failures that
can lead to draining of the SFP Such an evaluation would be needed for any
plants that are not screened from further assessment prior to Step 3a in Figure 11.
This section provides guidance that may be employed in addressing this
consideration for plant-specific evaluations.
7.1
Scope of the Seismic Evaluation for the SFP
The focus of the evaluation process described in this report is on elements of the
SFP that might fail due to a seismic event such that draining of the SFP could
result. This approach is intended to ensure that efforts to gain an understanding
of potential seismic risks needed to respond to the 50.54(f) letter are
appropriately focused on the most risk significant elements..
In developing guidance for the walkdowns associated with NTTF
Recommendation 2.3 [46], the emphasis
was on SFP connections whose failure
ph
-down.
encompasses failures that could lead to uncovering of irradiated fuel stored in the
SFP within 72 hours of the earthquake [46]. This criterion is used for the
evaluations under NTTF Recommendation 2.1 as well; that is, the evaluations
consider possible failures that could lead to uncovering fuel stored in the SFP
within 72 hours.
Failures that could conceivably lead to uncovering of irradiated fuel stored in the
SFP would include the following:
A significant failure of the steel-lined, reinforced concrete structure of the

SFP, causing inventory in the pool to drain out.
Failure of a connection penetrating the SFP structure (drain line, coolingwater line, etc.) below the top of the stored fuel.
Failure of a connection penetrating the SFP structure above the fuel

sufficient to drain significant inventory from the pool such that (in the
g 7-1h
absence of adequate makeup) evaporation and boil-off could cause fuel to be

uncovered within 72 hours.
Extensive sloshing such that sufficient water could be lost from the pool and,
as in the previous item, lead to uncovering of the fuel within 72 hours.
Failure of a cooling-water line or other connection that could siphon water

out of the pool sufficient to lead to uncovering of the fuel within 72 hours.
Tearing of the steel liner due to movement of fuel assemblies as a result of

the earthquake.
Failures that could lead to draining of SFP inventory when the pool and
reactor are configured for refueling operations.
With regard to these possibilities, the evaluation may generally be focused on

connected structures and systems that penetrate the SFP structure, rather than
the basic structure of the SFP itself. Detailed assessments have been made of
SFP structural integrity, including by the NRC on several structures, and these
have found SFP structures to be reasonably rugged.
Previous evaluations in NUREG-1353 [57], NUREG-1738 [47] and
NUREG/CR-5176 [48] characterized the generally robust nature of the design
of SFPs currently in use. NUREG-1738 further identified a checklist that could
be used to demonstrate that a SFP would achieve a high very HCLPF.
ep
Evaluations reported
in NUREG/CR-5176 [48] for two older pl
plants concluded
In addition, previous screening criteria for civil structures in
EPRI NP-6041 [39] (e.g. Table 2-3) provide principles that would be helpful in
evaluating the ruggedness of SFP structures. Either the checklist in NUREG1738 should be used to demonstrate that the structure is sufficiently robust or
another approach can be used if sufficiently justified.
Tearing of the stainless-steel liner due to overall structural failure of the fuel pool
structure would be precluded by the successful completion of the structural
evaluations above. Tearing of the stainless-steel liner due to sliding or other
movement of the fuel assemblies in the pool is considered to be very unlikely
[57].
Depending on the approach taken (as outlined in the sections that follow), an
evaluation of events that could lead to draining the SFP may consider the ability
to make up the inventory lost to ensure that the stored fuel remains adequately
cooled. SFP inventory makeup strategies can be credited provided makeup
resources, including any necessary instrumentation, are seismically rugged and
available; procedures exist to guide response by the operator; and there is
sufficient time for operators to recognize the need for makeup and take action.
Credited operator actions should be reviewed to account for habitability and
accessibility limitations.
Beyond the impact of possible failures on the cooling of the fuel stored in the
SFP, for some plants the loss of inventory from the pool could cause flooding
g 7-2 h
that could affect other systems. The assessment of flooding will be evaluated
separately, as part of the response to a NTTF Tier 3 recommendation.
The remainder of this section outlines a process for identifying and evaluating
features that could lead to draining of the SFP.
7.2
Evaluation Process for the SFP
The process for evaluating the SFP begins with the identification of any
penetrations that should be considered. All penetrations should be identified and
placed into one of the following three categories:
1. Those that are above the level of the fuel in the SFP;
2. Those that are at a level below the top of the fuel in the SFP; and
3. Those that may have the potential to siphon water from the SFP (most
typically, the discharge line from the SFP cooling system).
The sections that follow provide guidance for addressing each of these categories.
Figure 7-1 shows the general process for evaluating SFP penetrations.
Identify SFP
penetrations
Penetration
not able to lead to
uncovering fuel
within 72 hrs?
Perform one of the following

Evaluate seismic
capacity of
penetration and
connected features
No
Evaluate seismic
capacity of makeup
systems
Yes
Screen out and

document evaluation
Yes
Adequate
seismic capacity,
or makeup?
No
Report results of
SFP evaluation
g 7-3h
Figure 7-1
Basic Process for Evaluation of Potential Failures for SFP
Penetrations
7.2.1
Evaluation of Penetrations above Top of Fuel
In most cases, penetrations in the SFP will be located above the top of the
irradiated fuel. Assessment of these penetrations does not need to account for the
potential that a failure would, in and of itself, result in draining the pool level
below the fuel. Failures of these penetrations could, however, still affect SFP
inventory. If the level in the pool could be lowered sufficiently due to a failure
associated with a connection via such a penetration, the volume of water in the
pool serving as a heat sink for the residual decay heat in the fuel assemblies could
be reduced.
In this case, the evaluation should determine whether the potential failure could
lead to uncovering the fuel within 72 hours. It is acceptable to evaluate either the
seismic adequacy of the penetrations or the makeup capabilities to demonstrate
overall SFP adequacy. Plants can choose which of these to address first (that is,
the seismic capacity of the penetration or the availability of makeup adequacy).
The evaluation should include the following.
References identified in Table 6-1 provide acceptable guidance for

computing seismic fragilities that can be used to evaluate the seismic
adequacy of SFP features. Computed SSC HCLPFs would be compared
with GMRS based demands for these evaluations.
For a relatively large potential failure (such as that of the fuel transfer gate),
the analysis should begin with an assumption that the level in the SFP drops
to the bottom of the penetration at essentially the same time as when the
failure occurred. For smaller failures, the time required to lower pool level to
the bottom of the penetration may be significant (refer to Section 7.3.1 for
guidance).
The amount of water lost due to sloshing (refer to Section 7.3.2 for
guidance) should be taken into account.
For a failure associated with a penetration above the top of the fuel, the loss of
inventory through the break will be limited to the level of the penetration.
Therefore, the makeup requirements are only those associated with matching
decay heat. If it is necessary to consider makeup capabilities, the evaluation
should confirm that the makeup systems have adequate seismic capacity to
address the needs for restoring and maintaining SFP inventory.
Maintaining the SFP water level above about two-thirds of the height of the fuel
assemblies in the pool should prevent overheating the fuel [49]. Therefore, the
ability to maintain SFP inventory at a level of about two-thirds of the height of
the fuel assemblies would be considered acceptable.
The makeup required to match decay heat if the SFP does not have fuel
assemblies freshly removed from the reactor may be as low as 20 to 30 gpm. For
g 7-4 h
an SFP that contains freshly offloaded fuel, the decay heat load may be
significantly higher. Plants routinely maintain information needed to calculate
the heat load in the SFP. Guidance for calculating the required makeup rates can
be found in Appendix EE of the report documenting the technical bases for
severe accident management guidance (SAMG) [49].
The evaluation should document the assessment of the penetrations, including
the provisions for makeup to prevent uncovering the stored fuel. The cumulative
effect of inventory losses from penetrations not screened out as seismically
adequate should be considered in the evaluation, in addition to sloshing and boil
off losses. If limitations are identified with respect to the capability of makeup
systems, these results should be reported as part of the SFP seismic evaluation.
7.2.2
Evaluation of Penetrations below Top of Fuel
The SFPs for plants operating in the United States are generally configured so
that they do not have penetrations below the top of the stored fuel. The absence
of penetrations lower in the pool inherently limits the potential to drain
inventory sufficiently to begin uncovering fuel. It is possible; however, that some
SFPs may have penetrations (e.g., drain lines) below the top of the stored fuel
assemblies. There may also be some SFPs for which the bottom of the transfer
gates is below the top of the fuel. A failure associated with such a penetration
could drain the pool level below the top of the fuel.
The evaluation should include the following.
References identified in Table 6-1 provide acceptable guidance for

computing seismic fragilities that can be used to evaluate the seismic
adequacy of SFP features. Computed SSC HCLPFs would be compared
with GMRS based demands for these evaluations.
For a relatively large potential failure (such as that of the fuel transfer gate),
the analysis should begin with an assumption that the level in the SFP drops
to the bottom of the penetration at essentially the same time as when the
failure occurred. For smaller failures, the time required to lower pool level to
the bottom of the penetration may be significant (refer to Section 7.3.1 for
guidance).
The amount of water lost due to sloshing (refer to Section 7.3.2 for
guidance) should be taken into account.
The evaluation should confirm that the makeup systems have adequate seismic
capacity to address the needs for restoring and maintaining SFP inventory. One
consideration is that a significant failure low in the pool has the potential to drain
water from the pool at a rate in excess of readily available makeup provisions. If
the penetration is above about two-thirds of the height of the fuel assemblies in
the pool, however, maintaining the water level at that point should prevent
overheating the fuel [49]. So, for example, if the transfer gate extends down to
2 ft below the top of the fuel, its failure may be acceptable, even though it may
not be possible to restore water level to above the top of the fuel.
g 7-5h
The evaluation should document the assessment of the penetrations, including

the provisions for makeup to prevent uncovering the stored fuel. The cumulative
effect of inventory losses from penetrations not screened out as seismically
adequate should be considered in the evaluation, in addition to sloshing and boil
off losses. If limitations are identified with respect to the capability of makeup
systems, these results should be reported as part of the SFP seismic evaluation.
7.2.3
Evaluation of Potential for Siphoning Inventory
Although designs differ from plant to plant, for some SFPs the discharge line
from the SFP cooling system extends down into the pool. Cool water is
introduced low in the pool, and the suction line takes warm water from closer to
the top of the pool. If the SFP cooling system were to experience a failure, it is
possible that water could be siphoned back through the discharge line and out
the break. To prevent such an occurrence, SFP cooling systems with this
configuration are typically equipped with anti-siphon devices. If the anti-siphon
device were to function improperly, the effect could be similar to a break below
the top of the fuel, as addressed in Section 7.2.2. The process for evaluating
failures in the SFP cooling system that might lead to siphoning inventory from
the pool is outlined in Figure 7-2.
The anti-siphon devices are expected to be relatively rugged; for purposes of this
evaluation, an evaluation should be performed to confirm that, if such a feature is
needed to prevent siphoning water from the pool. If there are questions about the
ruggedness of the feature, the evaluation may follow one of three paths,
depending on what information is most readily available:
The seismic capacity of the anti-siphon feature can be assessed using the
fragility references listed in Table 6-1 and the resulting HCLPF compared to
the GMRS;
The SFP cooling system can be examined to determine if there are effective
isolation features that could be used to terminate the loss of inventory; or
An evaluation of makeup capabilities could be made, as for other breaks

below the level of the fuel.
In the very unlikely event that none of these options is viable, an evaluation can
be made of the seismic capacity of the SFP system (analogous to the assessment
called for in Section 7.2.2 for penetrations below the top of the stored fuel).
g 7-6 h
Identify SFP
cooling system with
discharge low in pool
Anti-siphon
device present and
inherently
rugged?
Perform one of the following

Evaluate seismic
capacity of antisiphon device
No
Evaluate provisions to
isolate potential break
from pool
Evaluate adequacy of
makeup capability to
maintain inventory
Yes
Screen out and

document evaluation
Yes
Adequate
seismic capacity,
isolation, or
makeup?
No
Report results of
SFP evaluation
Figure 7-2
Basic Process for Evaluation of Potential Siphoning of SFP
Inventory
7.3
Guidance for Additional Evaluations
To accomplish the tasks outlined in the preceding section, additional evaluations

may be required. This section provides guidance for the assessment of the timing
of uncovering fuel and for addressing the effects of sloshing.
7.3.1
Drain-down and Evaporative Losses
The evaluation of whether fuel could be uncovered in the event of a failure of an

interconnection at a level above the fuel can be accomplished in a relatively
straightforward manner.
For failures of piping systems connected above the top of the fuel, a flow rate can
be approximated using standard correlations, and assuming a driving head
equivalent to the initial height of water above the top of the connection. This
flow rate can be used to bound the time it would take lower level to that of the
connection.
g 7-7h
For larger connections (such as the gate used for transferring fuel during
refueling), the level can be assumed to drop to the bottom of the connection
nearly instantaneously.
Once level drops to the connection, a calculation can be made to determine the
additional loss of inventory that occurs up to 72 hours in the absence of makeup
flow. This rate of boil off loss can be determined using the correlations provided
in Appendix EE of the report documenting the technical bases for severe
accident management guidance [49].
These evaluations can be used to determine whether the top of the fuel could
begin to be uncovered within 72 hours.
7.3.2
Assessment of the Potential for Sloshing
To support the assessments described in Section 7.2, an estimate is needed of the

amount of water lost from the SFP due to sloshing. An initial, bounding
assessment can be made using the approach described in this section.
The natural frequency (fc1) for the fundamental convective (sloshing) mode of
vertical oscillation of the water surface in a rectangular pool due to shaking input
in either horizontal direction can be expressed as follows:
fc1 = (1/(2)(3.16g / L) tanh(3.16h / L)]
where:
0.5
Equation 7-1
L = pool length in the direction of shaking

h = water depth
g = gravity
Next, the slosh height (hs1) for the fundamental convective mode can be
estimated from:
hs1 = L(SAc1 / g)
Equation 7-2
where: SAc1 = % damped horizontal spectral acceleration at the top of the

pool wall at the frequency fc1 in the direction of motion
In order to account for higher convective modes of sloshing and nonlinear
sloshing effects (more upward splash than downward movement) observed
during stronger shaking, the theoretical slosh height predicted by Equation 7-2
may be increased by 20%. Thus, the total estimated slosh height becomes:
hs = 0.6L(SAc1 / g)
Equation 7-3
For a rectangular pool of length a in the x-direction, and width b in the ydirection, the slosh height due to x-direction shaking, and y-direction shaking
can be computed independently by substituting a and b, respectively, into
Equations 7-1 and 7-3. Next, the total slosh height (hst) can be estimated from:
2
2 0.5
hst = [hsx + hsy ]
Equation 7-4
g 7-8 h
where:
hsx = slosh height due to x shaking

hsy = slosh height due to y shaking
An upper bound estimate of the total volume V of water that might splash out of
the pool can be estimated from:
V = (hst hf)ab
where:
Equation 7-5
hf = freeboard height of the wall above the top of the water
Note that this approach reflects that sloshing in a pool is a very low frequency
phenomenon governed by either the peak ground displacement or the peak
ground velocity of the ground motion. It is independent of the PGA of the
ground motion.
While this approach is expected to produce a reasonable estimate of the slosh
height, it is expected to produce a very conservative estimate of the volume of
water displaced from the pool. It effectively assumes that a solid mass of water
equivalent to the product of the splash height above the side of the pool and the
pool area is lost from the pool.
This relatively simple calculation is adequate for purposes of estimating the loss
of SFP inventory due to sloshing. For most scenarios, it is judged that this
conservative estimate of the inventory lost due to sloshing will not have a
significant effect on the estimate of SFP drain-down. If the inventory lost due to
sloshing has a significant impact on SFP drain-down, a more careful calculation
may be required. Such a calculation would need to account for the time histories
of a range of earthquakes, and is likely to require significant resources, including
an independent review.
g 7-9h
Section 8: References
1. NRC (E Leeds and M Johnson)) Letter to All Power Reactor Licensees
Federal Regulations 50.54(f) Regarding Recommendations 2.1, 2.3 and
9.3 of the Near-Term Task Force Review of Insights from the
Fukushima Dai-Ichi Accident,
arch 12, 2012.
2. NRC Supplement 4 to Generic Letter (GL) 88-20, "Individual Plant
Examination of External Events (IPEEE) for Severe Accident
Vulnerabilities," (ADAMS Accession No. ML031150485).
nts
3. EPRI TR-112932,
Electric Power Research Institute, December 2000.
4. NUREG-1742, "Perspectives Gained from the Individual Plant

Examination of External Events (IPEEE) Program," issued April 2002
(ADAMS Accession Nos. ML021270070 and ML021270674).
5. NRC letter, May 26, 2005, recommendation to create GI-199,
ML051450456.
6. NRC, June 9, 2005, GI-199 established, ML051600272.
7. NRC meeting, February 2008, ML073400477 and ML080350189.
8. NRC Safety/Risk Analysis for GI-299, September 2, 2010,
ML100270582.
9. NRC Information Notice 2010-018, "Generic Issue 199, Implications of
Updated Probabilistic Seismic Hazard Estimates in Central and Eastern
United States on Existing Plants September 2, 2010, ML101970221.
10. NRC, Draft Generic Letter 2011-XX: Seismic Risk Evaluations for
Operating Reactors, July 26, 2011, ML11710783
11. NRC NUREG-1407, Procedural and Submittal Guidance for the
Individual Plant Examination of External Events (IPEEE) for Severe
Accident Vulnerabilities, June 1991.
12. American Society of Mechanical Engineers/American Nuclear Society
(ASME/ANS) RA-Sa-2009.
g 8-1 h
13. Budnitz, et al (1997). Recommendations for Probabilistic Seismic

Hazard Analysis: Guidance on Uncertainty and Use of Experts, Report
NUREG/CR-6372, Lawrence Livermore National Laboratory.
Sponsored by the U.S. Nuclear Regulatory Commission, U.S.
Department of Energy, and the Electric Power Research Institute.
14. NUREG-2115, DOE/NE-0140, EPRI 1021097. Central and Eastern
United States Seismic Source Characterization for Nuclear Facilities, 6
Volumes, January 2012.
15. JLD-ISG-2012-04, Guidance on Performing a Seismic Margin
Assessment in Response to the March 2012 Request for Information
Letter, November 2012, ML12286A029
16. EPRI 1009684, CEUS Ground Motion Project Final Report, Electric
Power Research Institute, December 2004.
17. EPRI 1014381, Program on Technology Innovation: Truncation of the
Lognormal Distribution and Value of the Standard Deviation for
Ground Motion Models in the Central and Eastern United States,
Electric Power Research Institute, August 2006.
Performance-Based Approach to
18. USNRC, Regulatory Guide 1.208,
Define the Site-Specific Earthquake Ground M
2007.
19. EPRI 1014099, Program on Technology Innovation: Use of Cumulative
Absolute Velocity (CAV) in Determining Effects of Small Magnitude
Earthquakes on Seismic Hazard Analyses, August 2006.
20. NEI 12-13
August 2012
21. EPRI 1012044, Program on Technology Innovation: Assessment of a
Performance-Based Approach for Determining Seismic Ground
Motions for New Plant Sites, August 2005.
22.
U.S. Nuclear Regulatory Commission Comments on

NEI 12-13, External Hazards PRA Peer Review Process Guidelines
Dated August
November 16, 2012.
23. NUREG -2117, Practical Implementation Guidelines for SSHAC Level

3 and 4 Hazard Studies: Rev. 1, Office of Nuclear Regulatory Research,
2012, ML12118A445.
24. NUREG/CR-6728, Technical Basis for Revision of Regulatory
Guidance on Design Ground Motions: Hazard- and Risk-Consistent
Ground Motion Spectra Guidelines, Risk Engineering, Inc., October
2001.
25.
Qualification of Class I E Equipment for Nuclear Power Generating
Station
26. EPRI NPImplementation,
g 8-2 h
27.
Seismic Functionality,
28.
ts of
High-Frequency
gh eq
y Ground Motion on Structures, Components, and
29.
of Components Sensitive to HighOctober 2007. (Also available under ML072600202.)
30. Westinghouse, AP1000 Design Control Document, Revision 17,
September 2008.
31. IEEE C37.98-1987, IEEE Standard Seismic Testing of Relays.
32. NRC, Risk-Informed Regulation Implementation Plan, 2000.
33. EPRI 1002989, Seismic Probabilistic Risk Assessment Implementation
Guide, December 2003.
34. IAEA Safety Report Series No. 28, Seismic Evaluation of Existing
Nuclear Power Plants, International Atomic Energy Agency, 2003.
35. IAEA TECDOC-724, Probabilistic Safety Assessment for Seismic
Events. International Atomic Energy Agency 1993.
36. EPRI 1019200, Seismic Fragility Applications Guide Update, December
2009.
37. EPRI 1002988, Seismic Fragility Application Guide, December 2002.
38. EPRI TR-103959, Methodology for Developing Seismic Fragilities, July
1994.
39. EPRI NP-6041-SL,
Electric Power Research Institute, August
1991.
40. ASCE 4-98, Seismic Analysis of Safety-Related Nuclear Structures,
American Society of Civil Engineers.
41. EPRI TR-104871
Electric
Power Research Institute, May 1995.
42. Criterion for Capacity-Based Selection of Structures, Systems and
Components for Fragility Analyses in a Seismic Risk-based Evaluation,
September 2012
43. NRC Regulatory Guide 1.200, An Approach For Determining The
Technical Adequacy Of Probabilistic Risk Assessment Results For RiskInformed Activities, Revision 2, March 2009.
44. Addenda B to ASME/ANS RA-Sa-2009, the version of the Standard
that is currently undergoing balloting.
g 8-3 h
45. Soil Structure Interaction Effects on a Containment Structure,

November 2012
46. EPRI 1025286, Seismic Walkdown Guidance for Resolution of
Fukushima Near Term Task Force Recommendation 2.3: Seismic, June
2012.
47. NRC NUREG-1738, Technical Study of Spent Fuel Pool Accident
Risk at Decommissioning Nuclear Power Plants. February 2001.
48. NRC NUREG/CR-5176, Seismic Failure and Cask Drop Analyses of
the Spent Fuel Pools at Two Representative Nuclear Power Plants.
January 1989.
49. EPRI 1025295, Update of the Technical Bases for Severe Accident
Management Guidance. Electric Power Research Institute, August 2012.
50. NEI-12-06, Diverse and Flexible Coping Strategies (FLEX)
Implementation Guide. Nuclear Energy Institute, Revision B1, May
2012.
51. Kennedy, R.P., Overview of Methods for Seismic PRA and Margin
Analysis Including Recent Innovations, Proceedings of the OECDNEA Workshop on Seismic Risk, Tokyo, August 1999.
52. "Walkdown Screening and Seismic Evaluation Training Course and
Add-On SMA Training Course," Seismic Qualification Utility Group
(1993); available from Electric Power Research Institute.
53.
-S-2008 Standard for Level 1/Large Early

Release Frequency
eq
y Probabilistic Risk Assessment for Nuclear Power
American Nuclear Society, 2009.
54. NRC NUREG/CR-2300, PRA Procedures Guide: A Guide to the

Performance of Probabilistic Risk Assessments for Nuclear Power
Plants, January 1983.
55. Soil Structure Interaction Effects on a Main Steam Valve House
Structure, November 2012
56. Japan Lessons-Learned Project Directorate, JLD-ISG-2012-04
Enhancements to the NRC Method for Seismic Margins Assessments
for Response to the March 2012 Request for Information, Draft Interim
Staff Guidance Revision 0 (Draft Issue for Public Comment),
September 4, 2012 (ML12222A327).
57. NRC NUREG-1353, Regulatory Analysis for the Resolution of Generic
Issue 82, "Beyond Design Basis Accidents in Spent Fuel Pools," April
1989
g 8-4 h
Appendix A: Control Point

Discussion from Standard
Review Plan
NUREG-0800 USNRC Standard Review Plan Rev. 2 1989:
The FSAR 2.5.2 Vibratory Ground Motion described the development
of the site SSE. Typically a peak ground acceleration (PGA) is
determined and a generic spectral shape was defined; e.g., Housner
spectra, Modified Newmark spectra, RG 1.60 spectra.
In FSAR 3.7.1 the implementation of the SSE ground motion for
seismic analysis and design is described. As discussed above the
methodologies for seismic analysis and design varied depending on the
vintage of the Plant.
NUREG-0800 Rev. 2 August 1989 provides the acceptance criteria for the later
set of existing NNP designs:
3.7.1 Seismic Design Parameters states under 1. Design Ground Motion
the following:
"The control motion should be defined to be a free ground
surface...Two cases are identified depending on the soil
characteristics at the site...uniform sites of soil or rock with
smooth variation of properties with depth, the control point
(location at which the control motion is applied) should be
specified on the soil surface at the top of finished grade...for sites
composed of one or more thin soil layers overlaying a competent
material...the control point is specified on an outcrop or a
hypothetical outcrop at a location on the top of the competent
material..."
3.7.2 Seismic System Analysis states under II Acceptance Criteria the
following:
"Specific criteria necessary to meet the relevant requirements of
GDC 2 and Appendix A to Part 100 are as follows:
g A-1 h
4. Soil-Structure Interaction...
C. Generation of Excitation System...
The control point...for profile consisting of component soil or
rock, with relatively uniform variations of properties with depth
the control motion should be located...at top of the finish
grade...For profiles consisting of one or more thin soil layers
overlaying component material, the control motion should be
located at an outcrop (real or hypothetical) at top of the
competent material...
...The spectral amplitude of the acceleration response spectra
(horizontal component of motion) in the free field at the
foundation depth shall be not less than 60 percent of the
corresponding design response spectra at the finish grade in the
free field..."
g A-2 h
Appendix B: Development of
Site-Specific Amplification
Factors
B1.0 Introduction
It has long been recognized that the amplitude and frequency content of ground
motions at a site are strongly influenced by the characteristics of the near-surface
materials. For most sites, however, the properties of the near-surface materials
and the parameters that control the dynamic response are not known with
certainty. The uncertainty in these parameters needs to be accounted for when
developing site specific hazard curves. Ultimately, the goal or objective of the site
response analysis is to produce site-specific hazard curves and response spectra
which reflect the desired exceedance frequencies, that is, preserve the reference
site annual exceedance frequency (AEF) thereby maintaining hazard consistency
for results produced at any elevation in the profile. However, the uncertainty in
characterizing the soil profile and dynamic properties of the near-surface
materials presents a challenge to preserving hazard levels for sites that differ from
some specified reference condition (which is described below).
Previously, the state of practice in calculating a site-specific ground motion has
been to calculate probabilistic reference rock ground motions and then multiply
them by deterministic site-amplification factors [17, 18]. However, as stated
above, there is uncertainty in the layering, spatial distribution and dynamic
properties of near-surface materials. This leads to uncertainty in the estimation of
site amplification functions. To alleviate this problem it is necessary to calculate
the effects of uncertainty on the estimate of the site-amplification functions and
use the resulting site-amplification distribution within a probabilistic
methodology [24, 16, 8].
The first step in developing site-specific seismic hazard curves and response
spectra consists of performing a Probabilistic Seismic Hazard Analysis (PSHA)
that reflects an outcropping reference site condition. The reference site condition
is usually hard or firm rock and is consistent with assumptions made in the
development of the most recent ground motion prediction equations. For central
and eastern North America (CENA) this represents a site with a theoretical
shear wave velocity over the top 1 km of the crust of 2.83 km/sec with a specified
shallow crustal damping parameter [18]. The shear-wave velocity is based on the
empirical Mid-continent compressional-wave velocity model of Pakiser and
g B-1 h
Mooney (1989) [28], taken by EPRI (1993) to represent the CENA, and an
assumed Poisson ratio. For western U.S. (WUS) sites an appropriate reference
condition should be selected that is well-constrained by observational data in the
ground motion prediction equations (GMPEs). Site-specific amplification
functions are then developed relative to the reference site condition.
After completing PSHA calculations for reference rock site conditions, the
development of hazard consistent, site specific horizontal seismic hazard results
may be considered as involving two independent analyses. The first is the
development of frequency and amplitude dependent relative amplification factors
(for 5% damped response spectra, Sa) between the site of interest and the
reference site (SaSITE (f)/SaREFERENCE (f)) that accommodates potential linear or
nonlinear site response. Currently the state-of-practice approach involves
vertically propagating shear-waves and approximations using equivalent-linear
analysis with either a time domain method (e.g. SHAKE) or a more
computationally efficient frequency domain random vibration theory (RVT)
method [32, 10].
Subsequent to the development of the amplification factors, site-specific motions
are computed by scaling the reference site motions with the transfer functions. As
suggested above, probabilistic methods have been developed [24, 8] that
accurately preserve the reference site hazard level and result in full site-specific
hazard curves. These fully probabilistic approaches represent a viable and
preferred mechanism to properly incorporate the site-specific aleatory
(randomness) and epistemic (uncertainty) variabilities of the dynamic properties
and achieve desired hazard levels and performance goals. The following sections
describe the specific steps in the development of the site-specific amplification
functions.
B2.0 Description of Sites Requiring Response Analysis
and Basis for Alternative Models
The level of detail and scope of the geological and geotechnical investigations
conducted during the licensing of the currently operating NPPs was consistent
with the state of the practice at the time of the plants design and licensing.
However, the state of the practice in earthquake engineering has evolved over the
last several decades. As a result, some of the detailed information required to
perform modern site response analyses (consistent with the request in the March
12, 2012 50.54(f) letter [27]) are lacking for some of the older plants. This lack
of information results in increased levels of uncertainty in the site response
analyses. The following sections describe how this uncertainty will be
accommodated in the site response analyses. The amount, quality and
applicability of the available data will determine the analysis procedures.
The information available to develop estimates of site properties and
characteristics will be primarily based on readily available sources (FSAR and
other regional data) for most locations. However, for sites with recent COL and
ESP submittals, the co-located operating plants would be expected to utilize any
g B-2 h
applicable information developed in the ESP and COL site characterization

process to the maximum extent practicable.
Site response analyses will be required for sites in the central and eastern U.S.
(CEUS) (i.e., those sites located east of the Rocky Mountains) when available
information suggests surficial materials will impact design motions at frequencies
below about 50 Hz. The conservative criteria used to determine if site-specific
amplification functions are required is more than 7.5m (25ft) of material with an
average shear-wave velocity less than 2591 m/s (8,500ft/s) over hard rock. Sitespecific response analyses will be required for all sites in the western U.S. (sites
west of the Rocky Mountain front).
Based on the need to determine if a facility requires detailed site response
analyses (the combined stiffness and velocity criteria described above), the first
step in the process is the compilation and evaluation of site geotechnical and
geophysical characteristics. This information should be summarized consistent
with the documentation described in Section 4 of the main report. The available
site-specific information will be highly variable in terms of quantity and
applicability and, to achieve consistency in the assessment of epistemic
uncertainty in dynamic material properties across sites, a general approach has
been developed that is intended to result in an adequate and appropriate level of
uncertainty within the context of limited site information. The recommended
approach is considered appropriate for characterizing uncertainty in amplification
functions utilizing only readily available information. To provide consistency
across the sites the approach developed consists of specific rules that are intended
to quantify epistemic uncertainty in shear-wave velocity, nonlinear dynamic
material properties, and shallow crustal damping for deep soil and firm rock site
conditions. The overall approach is described more fully below.
B2.1 Background on the Treatment of Uncertainties
There are two different types of uncertainty in the development of site-specific
amplification functions (AF(f)). First, at any given site, at the spatial dimensions
of typical nuclear facilities (100-200m (~325-650ft) scale dimensions) there is
expected to be some variability in important site response parameters such as
shear-wave velocities, non-linear dynamic material properties at any depth across
the footprint of the facility, and the overall thickness of soil/soft rock above firm
rock site conditions. It is important to attempt to capture this uncertainty in the
final AF(f) estimates. This is treated as an aleatory (randomness) type of
variability. Current practice represents this variability by developing a candidate
shear-wave velocity profile, depth and overall thickness of soil/soft rock and
associated non-linear dynamic
material properties
(shear
modulus reduction and
yn
prop
(
potential variations in shear-wave velocity and layer thickness are represented by
correlated random perturbations to the base-case values. This is frequently
referred to as a randomization process. A sufficient number of realizations are
used to develop stable statistical estimates (log median and log standard
deviation) of the amplification functions [e.g. 47].
g B-3 h
The second type of uncertainty is epistemic or lack of knowledge uncertainty.

This represents the uncertainty in the development of the base-case models for
site profile, dynamic properties, and seismological parameters. For wellcharacterized sites with abundant high-quality data this uncertainty would be
reduced, possibly eliminating the need to vary some of the site parameters such as
the site profile. This epistemic uncertainty would increase with decreasing
confidence in the available data and information. This uncertainty is evaluated
through the development of alternative base-case models. The approach applied
for the development of alternative base-case models (epistemic uncertainty) is
discussed in more detail in the following sections.
The following information is required to perform the site-specific response
analyses: site shear-wave velocity profiles, non-linear dynamic material
properties, estimates of low-strain site damping (parameterized through the
parameter, kappa), and input or control motions (including relevant seismological
parameters). These various factors are discussed individually in the following
sections.
B3.0 Development of Base-Case Profiles and Assessment
of Epistemic Uncertainty in Profiles and Dynamic
Material Properties
Epistemic uncertainty in depth to hard rock site conditions and dynamic material
properties, which includes shear-wave velocity profiles, site material damping at
low strain (parameterized through kappa), and modulus reduction and hysteretic
damping curves, should be accommodated through the development of
alternative mean cases. The specific methodology utilized to develop the
alternative cases will depend on the amount of information available at a given
site. Conceptually in this context, for poorly characterized sites with few if any
measured dynamic material properties, multiple cases should be developed with
broad ranges of epistemic uncertainty applied in the development of the
parameters of the alternative cases. For sites that have more complete site
characterization information available, smaller epistemic uncertainty factors can
be employed in the development of the alternative cases.
As stated in Section B-2.0, the information available to develop base-case shearwave velocity profiles will be primarily based on readily available sources (FSAR
and other regional data such as petroleum exploration wells). In addition, colocated operating plants are expected to utilize any applicable information
developed in the ESP and COL site characterization process to the maximum
extent practicable. Shear-wave velocity profiles should be based on geophysical
measurements and, if necessary, through inference from all available, applicable
site geotechnical/geological information (e.g., obtained from published shearwave velocities for similar rock types at other locations). Several options are
discussed below to develop base-case velocity profiles depending on the amount
of site information available.
For those cases where limited or no at-site information is available, a minimum
of three profile estimates combined with three kappa estimates and two sets of
g B-4 h
modulus reduction and hysteretic damping curves should be developed. If

significant uncertainty exists in the thickness of soil above firm or hard rock
conditions, this thickness should be treated as an epistemic uncertainty. The
three cases for shear-wave velocity profiles and kappa are referred to hereafter as
base-case, and upper-range and lower-range models. A general set of guidelines
should be employed to develop these cases for dynamic material properties and
associated weights and is described more fully below. The general computational
framework for developing the mean site amplification functions and associated
standard deviations is illustrated in Figure B-1.
B-3.1 Development Process for Base-Case Shear-Wave
Velocity Profiles
In order to predict site response as accurately as possible, and ultimately prevent
error from propagating into other engineering calculations, it is important to
define a detailed shear wave velocity (VS) profile that represents the known or
inferred in-situ velocity structure as realistically as possible. The following
discussion describes the development of the mean or base-case VS profile. The
alternative (upper-range and lower-range) models are derived from the base-case
model utilizing an information-informed epistemic factor. The development of
the upper-range and lower-range models is discussed in Section B-3.2 after the
base-case development.
For sites with sparse or very limited information regarding dynamic material
properties (e.g., a measured shear-wave velocity profile was unavailable), typically
an estimate based on limited surveys (e.g., compressional-wave refraction) is
available over some shallow, limited depth range. For such cases, as well as to
provide a basis for extrapolating profiles specified over shallow depths to hard
rock basement material, a suite of profile velocity templates has been developed,
parameterized with VS30 (time averaged shear-wave velocity over upper 30m of
the profile) ranging from 190m/s to 2,032m/s (620ft/s to 6670ft/s). The suite of
profile templates is shown in Figure B-2 to a depth of 305m (1,000ft). The
templates are from [40] supplemented for the current application with profiles
for VS30 values of 190m/s, 1,364m/s, and 2,032m/s. The latter two profiles were
added to accommodate cases where residual soils (saprolite) are present and
overly hard rock. For both soil and soft rock sites, the profile with the closest
velocities over the appropriate depth range should be adopted from the suite of
profile templates and adjusted by increasing or decreasing the template velocities
or, in some cases, stripping off material to match the velocity estimates provided.
For intermediate cases, such as when only the upper portion of a deep soil profile
is constrained with measured velocities, the VS template profile with velocities
closest to the observed velocity at the appropriate depth should be identified.
This template can then be used to provide a rational basis to extrapolate the
profile to the required depth.
For firm rock sites (typically with shear-wave velocities in the range of about
914m/s (3,000 ft/s) to 2,438m/s (8,000 ft/s), which are often composed of
g B-5 h
Cenozoic or Paleozoic sedimentary rocks such as shales, sandstones, siltstones, or

similar rock types, a constant shear-wave velocity gradient of 0.5m/s/m
(0.5ft/s/ft) should be used as a template and used to estimate the velocities over
the appropriate depth range. This gradient is based on deep measurements in
similar rock types in Japan [20]. The 0.5m/s/m velocity gradient is also consistent
with measurements in sedimentary rocks of similar type at the Varian well in
Parkfield, California [22]. It is recognized that the soil or firm rock gradients in
the original profiles are primarily driven by confining pressure and may not be
strictly correct for each adjusted profile template at each site. However, any
shortcoming in the assumed gradient is not expected to be significant as the
range in multiple base-case profiles accommodates the effects of epistemic
uncertainty in the profile gradient on the resulting amplification functions.
For all sites where limited data exists, or only exists for very shallow depths, it is
necessary to fully evaluate and integrate all existing geological information into
the development of the base-case profile. For sites with soil or soft rock at the
surface and much stiffer materials at relatively shallow depths (less than
approximately 60m (200ft)) the potential for strong resonance in the frequency
range of engineering interest exists. All relevant geological information should be
assessed to ensure this condition is identified.
An example is provided in Figure B-3 to schematically illustrate how a
combination of geological information and geophysical measurements may be
used to develop a base-case profile. The data available at this hypothetical site
consists of shallow shear-wave velocity measurements (a single S-wave refraction
profile) over only about the upper ~30m (100ft) of the profile with a V S30 of
approximately 450m/s. There are also geologic profiles and regional data
available in the FSAR that indicate firm rock is present at a depth about 45m
(150ft) beneath the site. A shear-wave velocity of approximately 1525m/s
(5000ft/s) is inferred for the firm rock based on velocity measurements on
comparable material elsewhere. Regional data indicates the firm, sedimentary
rocks extend to a depth of at least 1 km before crystalline basement rock is
encountered. The information is combined in the following manner to construct
a base case profile. The closest template profile to the 450m/s VS30 estimate is the
400m/s profile. The velocities in the 400m/s template are scaled by a factor of
1.125 (450/400) to adjust to the desired VS30 value. At the 45m (150ft) depth, a
velocity discontinuity is inserted with a velocity of 1525m/s (5000ft/s). Below
this depth the firm rock gradient model of 0.5m/s/m is used to estimate
velocities. This gradient is extended to a sufficient depth such that 2830m/sec is
reached or the depth is greater than the criteria for no influence on response for
frequencies greater than 0.5Hz. The uncertainty in the depth to the soil-firm
rock interface is incorporated in the treatment of epistemic uncertainty as
discussed below.
B-3.2 Capturing Epistemic Uncertainty in Velocity
Profiles
There are basically two approaches for constructing shear wave velocity profiles,
either through inference from geotechnical/geologic information or through the
g B-6 h
use of geophysical measurements. Each approach will inherently have some level
of uncertainty associated with its ability to accurately represent the in-situ
velocity structure. The level of uncertainty will depend on the amount of
information available along with how well the information is correlated with
shear-wave velocity. By adopting the general mean based approach outlined in
Section B-3.1, a level of uncertainty can be assigned to a template velocity
profile, commensurate with the available information, in order to account for the
epistemic uncertainty associated with the in-situ velocity structure.
For sites where geophysical information such as very limited shear-wave velocity
data exists, the estimate for uncertainty in shear-wave velocity is to be taken as:

This value is similar to a Coefficient of Variation (COV) of 0.25 which is
consistent with Toro (1997) [37] for observed spatial variability over a structural
footprint of several hundred meters. The profile epistemic uncertainty factor of
0.35 ( ln) is to be applied throughout the profile and is based on the estimates
of epistemic uncertainty in VS30 developed for stiff profiles [14]. The logarithmic
factor assumes shear-wave velocities are lognormally distributed and was
originally developed to characterize the epistemic uncertainty in measured VS30 at
ground motion recording sites where measurements were taken within 300m
(900ft) from the actual site. The uncertainty accommodates spatial variability
over maximum distances of 300m, and is adopted here as a reasonable and
realistic uncertainty assessment reflecting a combination of: (1) few velocity
measurements over varying depth ranges, and (2) the spatial variability associated
with observed velocities. The application of the uncertainty estimate over the
entire profile that is based largely on VS30 implicitly assumes perfect correlation
that is independent of depth. While velocities are undoubtedly correlated with
depth beyond 30m, which forms the basis for the use of VS30 as an indicator of
relative site amplification over a wide frequency range, clearly the correlation is
neither perfect nor remains high over unlimited depths [11]. More direct support
for the assumption of a of 0.35 comes from the measured range in (VS30)
conditional on proxy inferences. For the four currently employed (VS30) proxies:
surficial geology (42, 41) Geomatrix site category (14), topographic slope (39),
and terrain (43), the overall within class or category uncertainty is about 0.35.
This uncertainty reflects the variability of measured (VS30) about the predicted
value and is relatively constant across proxies (34). The proxy uncertainty of
about 0.35 is a direct measure of the epistemic uncertainty of the predicted value
or estimate and supports the adoption of 0.35 to quantify the velocity uncertainty
for cases with few or absent direct measurements. For the application to site
characterization the of 0.35 has been extended to the entire profile.
An example of the resulting mean ln shear-wave profiles for the 760m/s
template is shown in Figure B-4.
For sites where site-specific velocity measurements are particularly sparse (e.g.,
based on inference from geotechnical/geologic information rather than
g B-7 h
geophysical measurements) a conservative estimate of the uncertainty associated

with the template velocity is to be taken as:

For sites where multiple, detailed shear-wave velocity profiles are available, the
level of uncertainty may be significantly reduced, depending on the number,
depth ranges, and vintage (quality) of the surveys. For sites with an intermediate
level of information available, such as a single shear-wave velocity profile of highquality or shear-wave velocities inferred from measured compressional-wave
velocities, (ln) may be reduced by a factor of two over the better constrained
portions of the profile. A specific factor of two is taken to provide consistency
across the sites. For well characterized sites with multiple measurements the
epistemic uncertainty may be further reduced. In all cases of reduced the range
in base-case profiles reflecting epistemic uncertainty in shear-wave velocity must
represent a realistic expression of the existing information at the site as well as
possible ranges in velocities for the materials considered. For all sites considered,
the shear-wave velocities developed for the upper- and lower-range base-cases
must reflect realistic values for the respective geologic conditions at the site. The
bases for these conclusions should be discussed in the report.
B-3.2.1 Epistemic Uncertainty in Final Hazard
Calculations
It is necessary to represent the epistemic uncertainty in the distribution of
potential shear-wave velocity profiles (mean base-case and a (ln) of 0.35, for
example) in the final site-specific hazard results. Practicality requires this be
accomplished with the minimum number of cases. The recommended approach
for this application is to utilize three cases, the mean base-case and upper range
and lower range base-cases with relative weights applied. An accurate three point
approximation of a normal distribution which preserves the mean utilizes the
50th-percentile (median) and 10th and 90th percentiles, with relative weights of
0.30 for the 10th and 90th percentiles and 0.40 for the median applied [23]. These
values are summarized in Table B-1. The 10th and 90th fractiles correspond to a
profile scale factor of 1.28 . When the 10th or 90th percentiles are
obtained by subtracting or adding 0.45 in natural log units to the shear wave
velocity. For , a value of 0.64 is subtracted or added to the natural log
of the shear wave velocity for the 10th and 90th percentile values. This is
equivalent to an absolute factors of 1.57 or 1.90 applied to the mean base-case
profile for or , respectively. Figure B-5 illustrates
application of these two factors applied to the 760 m/s (1525 ft/s) VS30 template.
Figure B-6 illustrates the same type of curves for the firm rock template derived
using the empirical gradient of [20]. For some individual sites it may be necessary
to deviate from these standard weights if application of the standard factors
results in velocities that are not deemed credible.
Figure B-7 illustrates the development of Upper Range and Lower Range
profiles to accommodate epistemic uncertainties for the hypothetical example
shown in Figure B-3. A ln = 0.35 has been used to develop the 10th and 90thg B-8 h
percentile curves in the upper portion of the profile where sparse Vs

measurements were available. A ln = 0.50 was applied to the lower portion of
the profile where the Vs of the Base Case was inferred from geological
information. The 90th-percentile curve was capped at a value equal to the
2830m/s Vs value assumed for the hard rock basement. This example illustrates
the broad range of velocities encompassed by the Upper Range, Mean, and
Lower Range profiles for sites lacking in good data.
For sites where the depth to firm rock conditions is poorly constrained, that
depth should be treated as a separate epistemic uncertainty as illustrated in Figure
B-1.
B-3.3 Nonlinear Dynamic Material Properties
The potential nonlinear response of near-surface materials to input ground
motions is an important element of the site that needs to be characterized in a
proper site response analysis. To characterize the epistemic uncertainty in
nonlinear dynamic material properties for both soil, and firm rock sites, two sets
of generic modulus reduction and hysteretic damping are recommended. The two
sets of generic curves, one set for soils and a second set for firm rock, are
considered necessary to reflect a reasonable and realistic range in nonlinear
dynamic material properties and associated degree of nonlinearity in site
response, conditional on velocity profile. The use of the recommended sets of
curves provides an acceptable expression of epistemic uncertainty in nonlinear
dynamic properties.
The use of two sets of generic curves recognizes the existence of test results and
site-specific curves at some existing sites as well as nearby COL and ESP
applications. For the purpose of initial screening and prioritization, the range in
nonlinearity implied by the recommended generic curves is intended to
accommodate the range in nonlinearity in site response of site-specific curves
while providing for a consistency in the treatment of epistemic uncertainty in
dynamic material properties (linear as well as nonlinear) across all sites.
For soils, the two sets of proposed curves are the EPRI (1993) [18] and
Peninsular Range [32, 40] results. The two sets of generic curves are appropriate
for cohesionless soils comprised of sands, gravels, silts, and low plasticity clays.
The EPRI (1993) [18] curves, illustrated in Figure B-8, were developed for
application to CEUS sites and display a moderate degree of nonlinearity. The
EPRI (1993) [18] curves are depth (confining pressure) dependent as shown in
Figure B-8.
The Peninsular Range curves reflect more linear cyclic shear strain dependencies
than the EPRI (1993) curves [40] and were developed by modeling recorded
motions as well as empirical soil amplification in the Abrahamson and Silva
WNA (Western North America) GMPE [32, 2]. The Peninsular Range curves
reflect a subset of the EPRI (1993) soil curves with the 51 to 120 ft (15 to 37 m)
EPRI (1993) curves applied to the 0 to 50 ft (0 to 15 m) depth range and the
g B-9 h
EPRI (1993) 501 to 1,000 ft (153 to 305 m) curve applied to the 51 to 500 ft (15
to 152 m) depth range, below which linear behavior is assumed.
The two sets of soil curves are considered to reflect a realistic range in nonlinear
dynamic material properties for cohesionless soils. The use of these two sets of
cohesionless soil curves implicitly assumes the soils considered do not have
response dominated by soft and highly plastic clays or coarse gravels or cobbles.
The presence of relatively thin layers of hard plastic clays are considered to be
accommodated with the more linear Peninsular Range curves while the presence
of gravely layers are accommodated with the more nonlinear EPRI (1993) soil
curves, all on a generic basis. The potential impact on the amplification functions
of the use of these two sets of nonlinear dynamic property curves was evaluated
and is shown in Figures B-9 and B-10. The results indicate that above 1 Hz the
difference can be significant and the resulting epistemic uncertainty needs to be
included in the development of the final amplification functions.
The two sets of soil curves are given equal weights (Table B-1 and Figure B-1)
and are considered to represent a reasonable accommodation of epistemic
uncertainty in nonlinear dynamic material properties for the generic types of soils
found at most in CEUS sites which include:
1. Glaciated regions which consist of both very shallow Holocene soils
overlying tills as well as deep soils such as the Illinois and Michigan basins,
all with underlying either firm rock (e.g., Devonian Shales) and then hard
basement rock or simply hard basement rock outside the region of Devonian
Shales,
2. Mississippi embayment soils including loess,
3. Atlantic and Gulf coastal plain soils which may include stiff hard clays such
as the Cooper Marl,
4. Residual soils (saprolite) overlying hard metamorphic rock along the
Piedmont and Blue Ridge physiographic regions.
For firm rock site conditions, taken generally as Cenozoic or Paleozoic
sedimentary rocks, such as shale, sandstones, or siltstones, two alternative
expressions
of nonlinear dynamic material behavior are
pr
-11) and linear response. The EPRI rock curves were developed
during the EPRI (1993) [18] project by assuming firm rock, with nominal shearwave velocities in the range of about 914m/s to 2134m/s (3,000ft/s to about
7,000ft/s, about 5,000ft/s on average), behaves in a manner similar to gravels [18]
being significantly more nonlinear with higher damping than more fine grained
sandy soils. The rock curves were not included in the EPRI report as the final
suite of amplification factors was based on soil profiles intended to capture the
behavior of soils ranging from gravels to low plasticity sandy clays at CEUS
nuclear power plants. With the stiffness typically associated with consolidated
sedimentary rocks, cyclic shear strains remain relatively low compared to soils.
g B-10 h
gn
ty in the soft-to-firm rock materials is largely confined to
Significant
nonlinearity
As an alternative to the EPRI rock curves, linear response should be assumed.
Implicit in this model is purely elastic response accompanied with damping that
remains constant with cyclic shear strain at input loading levels up to and beyond
1.5g (reference site). Similar to the two sets of curves for soils, equal weights were
given to the two sets of nonlinear properties for soft/firm rock sites as
summarized in Table B-1.
B-3.4 Densities
Because relative (soil surface/reference site) densities play a minor role in sitespecific amplification, a simple model based on the shear-wave velocity of the
mean base-case profile is proposed for those sites where a profile density is not
available. This model relating estimated shear-wave velocity and density is
summarized in Table B-2.
Due to the square root dependence of amplification on the relative density, a 20%
change in soil density results in only a 10% change in amplification and only for
frequencies at and above the column resonant frequency. As a result only an
approximate estimate of profile density is considered necessary with the densities
of the mean base-case profile held constant for the upper and lower range basecase profiles. This approach provides a means of accommodating epistemic
uncertainty in both density as well as shear-wave velocity (Section B-3.1) in the
suite of analyses over velocity uncertainty.
B-4.0 Representation of Aleatory Variability in Site
Response
To accommodate the aleatory variability in dynamic material properties that is
expected to occur across each site (at the scale of the footprint of a typical nuclear
facility), shear-wave velocity profiles as well as G/GMAX and hysteretic damping
curves should be randomized. The aleatory variability about each base-case set of
dynamic material properties should be developed by randomizing (a sufficient
number of realizations to produce stable estimates) shear-wave velocities, layer
thickness, depth to reference rock, and modulus reduction and hysteretic
damping curves. For all the sites considered, where soil and firm rock extended to
depths exceeding 150m (500ft), linear response can be assumed in the deep
portions of the profile [32, 33, 35, 36].
B-4.1 Randomization of Shear-Wave Velocities
The velocity randomization procedure makes use of random field models [37] to
generate VS profiles. These models assume that the shear-wave velocity at any
depth is lognormally distributed and correlated between adjacent layers. The
layer thickness model also replicates the overall observed decrease in velocity
fluctuations as depth increases. This realistic trend is accommodated through
increasing layer thicknesses with increasing depth. The statistical parameters
g B-11 h
required for generation of the velocity profiles are the standard deviation of the
natural log of the shear-wave velocity (VlnVs) and the interlayer correlation (UIL).
For the footprint correlation model, the empirical lnVs is about 0.25 and
decreases with depth to about 0.15 below about 15m (50ft) [32]. To prevent
unrealistic velocity realizations, a bound of 2lnVs should be imposed throughout
the profile. In addition, randomly generated velocity should be limited to
2.83km/s (9200ft/s). All generated velocity profiles should be compared to
available site-specific data as a check to ensure that unrealistic velocity profiles are
removed (and replaced) from the set of velocity profiles used to develop site
response amplification functions. This process should be documented as part of
the site response analysis.
B-4.2 Aleatory Variability of Dynamic Material
Properties
The aleatory variability about each base-case set of dynamic material properties
(EPRI depth dependent vs. Peninsular for example) will be developed by
randomizing modulus reduction and hysteretic damping curves for each of the
thirty realizations. A log normal distribution may be assumed with a ln of 0.15
and 0.30 at a cyclic shear strain of 3 x 10-2% for modulus reduction and hysteretic
damping respectively [32]. Upper and lower bounds of 2 should be applied.
The truncation is necessary to prevent modulus reduction or damping models
that are not physically realizable. The distribution is based on an analysis of
variance of measured G/GMAX and hysteretic damping curves and is considered
appropriate for applications to generic (material type specific) nonlinear
properties [32]. The random curves are generated by sampling the transformed
normal distribution with a ln of 0.15 and 0.30 as appropriate, computing the
change in normalized modulus reduction or percent damping at 3 x 10-2% cyclic
shear strain, and applying this factor at all strains. The random perturbation
factor is then reduced or tapered near the ends of the strain range to preserve the
general shape of the base-case curves [18, 32]. Damping should be limited to a
maximum value of 15% in this application.
B-5.0 Development of Input Motions
The ground
motion used as input
analyses
gr
pu to site response
sp
ly is commonlyy referred
scaled to a response spectrum or, in the case of Random Vibration Theory, a
power spectral density (PSD). Because of the very large number of cases that will
need to evaluated to capture the range of epistemic uncertainty and aleatory
variability in this application (See Figure B-1 and Table B-3) the following
discussion will assume that the much more efficient random vibration theory
(RVT) approach to performing site response analyses will be utilized as opposed
to a time series (TS) based technique. The following sections describe the model
used in the development of the control motions and the parameters of that model
that require an assessment of uncertainty.
g B-12 h
B-5.1 Simple Seismological Model to Develop Control

Motions
The methodology suggested for developing the input or control motions relies on
a widely used, simple seismological model to represent earthquake source,
propagation path and site characteristics ([10] and references therein). The
ground motions recorded at a given site from an earthquake can be represented in
the frequency domain as:
Y(M0, R, f) = E(M0
Where Y(M0,R,f) is the recorded ground motion Fourier amplitude spectrum,
E(M0,f) is the Brune point-source seismic spectrum, P(R,f) represents the
propagation path effects, and G(f) represents the modification due to site effects.
In this equation M0 is the seismic moment of the earthquake, R is the distance
from the source to the site, and f is frequency. The seismic moment and the
earthquake magnitude are related through the definition of the moment
magnitude, M [21]:
M=2/3Log10M0 -10.7.
The P(R, f) term accounts for path effects, geometrical spreading and frequency
dependent deep crustal damping and can be expressed as:
P(R,f) = S(R) exp((-S f R)/(Q(f)VS)).

Where S(R) is the geometrical spreading function, Q(f) is the seismic quality
factor, and VS is the shear-wave velocity in the upper crust.
The G(f) term accounts for upper crustal amplification and frequencyindependent shallow crustal damping:
Where A(f) is the amplification function relative to source depth velocity

conditions and D(f) represents the frequency-dependent damping term (D(f) =
exp(-SN0f)). The A(f) term may be calculated using simplified square-root
impedence methods (A(f) = (Zsource/Zavg)0.5), where Z is the product of density
and velocity (UVS ), for example) or using more detailed full resonant techniques
[44].
Kappa (N0) is an upper crustal site ground motion attenuation parameter that
accounts for the overall damping in the basement rock immediately beneath a
site. The properties and behavior of the upper few hundreds of meters of the
crust has been shown to produce as much as 50% or more of the total attenuation
(Q(f)) of the high-frequency portion of the ground motion spectrum [1, 4]. The
value of kappa influences the shape of the ground motion spectrum observed at a
given site. High values of kappa result in enhanced attenuation of the highfrequency portion of the spectrum.
g B-13 h
The factors in the simple seismological model that affect the spectral shape of the
input motions are kappa, magnitude, attenuation model and source model. These
factors are discussed below. An example of the potential effect of these
parameters on the spectral shape of the input ground motions (Fourier amplitude
spectra and 5%-damped response spectra) is shown in Figures B-12 and B-13
(input parameters are summarized in Table B-7).
B-5.1.1 Magnitude
Conditional on reference site peak acceleration, amplification factors depend, to
some extent, upon control motion spectral shape due to the potential nonlinear
response of the near-surface materials. For the same reference site peak
acceleration, amplification factors developed with control motions reflecting M
5.5 will differ somewhat with those developed using a larger or smaller
magnitude, for example.
Figures B-14 and B-15 show amplification factors developed for the 400m/s VS30
template profile (Figure B-2) using the single-corner source model for
magnitudes M 5.5, 6.5, and 7.5. For this sensitivity analysis the more nonlinear
EPRI G/Gmax and hysteretic damping curves (Figure B-8) were used. The
dependence on control motion spectral shape is observed to decrease with degree
of nonlinearity becoming independent for linear analyses. As Figures B-14 and
B-15 illustrate, the largest amplification reflects the lowest magnitude (M
M 5.5).
Over the frequency range of about 5 to 10 Hz, and the ground motion amplitude
range of most engineering interest (between 0.1 g and 0.75 g), the difference in
the derived amplification
functions between the magnitudes
is minor. The largest
pl
gn
0.75g). The largest difference in amplification is between M 5.5 and M 6.5 with
little difference (< 10%) between M 6.5 and M 7.5. Given the most current
source characterization model in CENA [26] and the distribution of existing
NPP sites, the dominant contribution for the annual exceedance frequencies
(AEF) of 10-4 and below are from magnitudes in the range of about M 6 to M
7+. Given these factors, and the large number of analyses required (Table B-3) a
M 6.5) is proposed for development of the control motions.
single magnitude (M
This is felt to adequately characterize the amplification, with tacit acceptance of
slight conservatism for magnitude contributions above about M 7.
B-5.1.2 Attenuation (Q(f)) Model
As illustrated in Figures B-12 and B-13, major differences in the assumed crustal
attenuation model (note Q(f) depends strongly on geometrical spreading) will
influence the spectral shape of the control motions. However, within a given
tectonic region, the CENA or the WUS for example, changes in the crustal
attenuation model do not contribute significantly to changes in the derived
amplification functions. Appropriate, widely referenced crustal attenuation
models are proposed for the CEUS and WUS sites (Table B-4).
g B-14 h
B-5.1.3 Kappa
In the context of this discussion, the kappa referred to here is the profile damping
contributed by both intrinsic hysteretic damping as well as scattering due to wave
propagation in heterogeneous material. Both the hysteretic intrinsic damping and
the scattering damping within the near-surface profile and apart from the crust
are assumed to be frequency independent, at least over the frequency range of
interest for Fourier amplitude spectra (0.33 to about 25.0 Hz). As a result, the
kappa estimates reflect values that would be expected to be measured based on
pi
ys of wavefields propagating
p paga ng throughout
gh
pr
empirical
analyses
the profiles
at low
s within the
profile [12]. Changes in kappa can exert a strong impact on derived control
motion spectra (Figures B-11 and B-13) and as a result are an important part of
the input model for development of control motions. Hence, similar to the
treatment of uncertainties in shear-wave velocity profiles, multiple base-cases
(mean and upper and lower ranges) may be developed for kappa.
B-5.1.3.1 Development of Base Case Kappa Models
Mean base-case kappa values were developed differently for soil and firm rock
sites.
CEUS Rock Sites: For rock sites with at least 3,000ft (1000m) of firm
sedimentary rock (VS30 > 500 m/s) overlying hard rock, the kappa-VS100 (average
shear-wave velocity over the upper 100ft of the profile) relationship of
log(N) = 2.2189 1.0930*log(VS100),
Where VS100 is in ft/s, is proposed to assign a mean base-case estimate for kappa
[36, 38]. The requirement of a 3000ft (1,000m) thickness of firm materials
reflects the assumption that the majority of damping contributing to kappa
occurs over the upper km of the crust with a minor contribution from deeper
materials (e.g., 0.006s for hard rock basement material). As an example, for a
firm sedimentary rock with a shear-wave velocity of 5,000ft/s (1525m/s), this
relationship produces a kappa estimate of about 0.02s. The assumption that is
typically used implies a kappa of 0.014s is contributed by the sedimentary rock
column and 0.006s from the underlying reference rock (Table B-4), and reflects
an average Qs of about 40 over the 3,000ft depth interval. The Qs value of 40 for
sedimentary rocks is consistent with the average value of 37 observed (measured)
over the depth range of 0m to 298m in Tertiary claystones, siltstones, sandstones,
and conglomerates at a deep borehole in Parkfield, California [22].
For soft/firm rock cases with low estimated velocity values, an upper bound
kappa value of 0.04s should be imposed. The maximum kappa value of about
0.04s reflects a conservative average for soft rock conditions [31, 32].
For cases where the thickness of firm rock is less than about 3,000ft (1000m) and
the relationship cited above is not applicable, the kappa contributed by the firm
g B-15 h
rock profile can be computed assuming a Qs of 40 plus the contribution of the

reference rock profile of 0.006s (Table B-4). For the three base-case firm rock
template profiles shown in Figure B-6, the total kappa values assuming a Qs of
40 are 0.019s, 0.025s, and 0.015s for the mean, lower range, and upper range
base-cases respectively.
Soil Sites:
For soil sites (either in the WUS or CEUS) with depths exceeding 1000m
(3,000ft) to hard rock, a mean base-case kappa of 0.04s should be assumed based
upon observed average values for deep soil sites and low loading levels. The mean
base-case kappa of 0.04s adopted for deep firm soils is lower than the value of
approximately 0.06s based on recordings at alluvium sites located in Southern
California [4, 32]. For soil sites, due to nonlinear effects, low strain kappa may be
overestimated depending upon loading level and the nonlinear dynamic material
properties. To avoid potential bias in the deep, firm soil, low strain kappa, the
value of 0.04s is based on inversions of the Abrahamson and Silva [3] soil site
GMPE [32]. In that inversion, a range of rock site loading levels was used with
the soil value of 0.04s based upon a rock site peak acceleration of 5% g or less,
clearly a low strain estimate. The deep soil mean base-case kappa of 0.04s is
adopted
for both the upper
with the assumption
that the
op
pp and lower range
ge profiles
p
pt
1000m (3,000ft)) firm soils in the CEUS is somewhat less than the 0.054s
inferred by Campbell [12] based on Cramer et al. [15] analyses for effective Qs
within the 960m deep sedimentary column in the Mississippi embayment near
Memphis, Tennessee. The deep firm soil kappa of 0.04s is in fair agreement with
0.052s found by Chapman, et al. [13] for the 775m thick sedimentary column
near Summerville, South Carolina.
CEUS, a nominal kappa value of 0.04s based on an average of many empirical
estimates predominately in the WNA tectonic regime is proposed. Sparse
analyses for deep soil sites in the CEUS suggest 0.04s reflects some conservatism.
However it should be noted the small strain total kappa is rapidly exceeded (i.e.,
becomes less important) as loading level increases due to nonlinear response. The
initial low strain kappa serves primarily as a means of adjusting (lowering) kappa
to accommodate the scattering component due to the profile randomization.
Hence, no significant bias in the final amplification functions at loading levels of
engineering interest is anticipated.
For cases of shallower soils, less than 1000m (3,000ft) to hard rock basement
material, the empirical relation of Campbell [12] should be used for the
contribution to kappa (in milliseconds) from the thickness of the sediment
column (H):
N(ms) = 0.0605*H (thickness in meters).
The assumed basement kappa
appa value of 0.006s ((Table B-4)
4) is used in lieu of
ution to total kappa.
g B-16 h
0.0665s (0.0605 contribution from soil and 0.006 contribution from basement
rock) , considerably larger than the mean base-case value of 0.04s, suggesting a
gr of conservatism at low loadingg levels for CENA firm soils. For continuity,
degree
should be implemented for sites with less than 1000m (3,000ft) of firm soils.
B-5.1.3.2 Representation of Epistemic Uncertainty in
Kappa
The parameter kappa is difficult to measure directly. Since no measurements of
the type required exist at the sites of interest, a large uncertainty is applied in the
site response analyses. Epistemic uncertainty in kappa is taken as 50% (Ln =
0.40, Table B-1; [18]) about the mean base-case estimate for this assessment.
The uncertainty is based on the variability in kappa determined for rock sites
which recorded the 1989 M 6.9 Loma Prieta earthquake [18], and adopted here
as a reasonable expression of epistemic uncertainty at a given site. As with the
shear-wave velocity profiles (Section B-3.2.1), the +/-0.51 natural log units (1.28
) variation is considered to reflect 10% and 90% fractiles with weights of 0.30
and a weight of 0.40 for the mean base-case estimate. The models for epistemic
uncertainty are summarized in Table B-1.
B-5.1.4 Source Model
Alternative conceptual models to represent the earthquake source spectral shape
exist in the literature. A single corner frequency model of the earthquake source
spectrum has been widely used in the simple seismological model described above
[10 and 18]. However, based on the limited ground motion data in CENA as
well as inferences from intensity observations, an alternative empirical two-corner
source model for CENA earthquakes has been developed [5]. The two-corner
source model addresses the potential for CENA source processes to reflect a
significant spectral sag at large magnitude (M
M
nd intermediate frequency
[6], compared to source processes of tectonically active regions. Such a trend was
suggested by the 1988 M 5.9 Saguenay, Canada and 1985 M 6.8 Nahanni,
Canada earthquakes. The two-corner source model for CENA [6] incorporates
the spectral sag between two empirical corner frequencies which are dependent
on magnitude. The two-corner model merges to the single-corner model for M
less than about M 5. Interestingly, the two-corner model has been implemented
for tectonically active regions and shown to be more representative of WNA
source processes than the single-corner model [7], albeit with a much less
pronounced spectral sag than the CENA model.
The debate regarding the applicability of these two source models continues. The
lack of relevant observations for M >6 in CENA precludes identifying either
model as a unique, preferred model. As a result, in the interest of representing
the epistemic uncertainty in this element of the control motions, both single- and
two-corner [6] source models were used with M 6.5 to develop control motions.
The two models were considered to reflect a reasonable range in spectral
g B-17 h
composition for large magnitude CENA sources. As a result, equal weights were
selected as shown in Table B-1 to develop amplification factors using each source
model.
Additionally, for moderately stiff soils, typical for NPP siting, the difference in
amplification between single- and double-corner source models becomes
significant only at the higher loading levels as Figures B-16 and B-17 illustrate.
Figures B-16 and 17 compare the amplification computed for both the singleand double-corner source models using the EPRI modulus reduction and
hysteretic damping curves (Figure B-8), the most nonlinear set of curves for soils.
These results suggest the difference in amplification between single- vs. doublecorner source models are significant enough to consider the implied epistemic
uncertainty in central and eastern North America (CENA) source processes at
M >6).
large magnitude (M
B-5.1.4.1 Development of Input Motions
It is necessary to define the site response over a broad range of input amplitudes
to develop amplification functions. For sites in the CEUS, the Mid-continent
crustal model [18] (Table B-5) with a shear-wave velocity of 2830 m/s, a defined
shallow crustal damping parameter (kappa; [4]) of 0.006 s, and a frequency
dependent deep crustal damping Q model of 670 f0.33 [18] is used to compute
reference motions (5% damped pseudo absolute acceleration spectra). The
selected Q(f), geometrical spreading, kappa, and reference site shear-wave
velocities are consistent with the EPRI GMPEs (Ground Motion Prediction
mply placed
p
op of this
Equations) [19]. The site-specific profiles are simply
on top
defined crustal model which has a reference shear9,300 ft/s) and a reference kappa value of 0.006 s. Distances are then determined
to generate a suite of reference site motions with expected peak acceleration
values which cover the range of spectral accelerations (at frequencies of 0.5, 1.0,
2.5, 5.0, 10.0, 25.0, 100.0 Hz) anticipated at the sites analyzed. To cover the
range in loading levels, eleven expected (median) peak acceleration values at
reference rock are needed to span from 0.01g to 1.50g. This is the minimum
number of intervals felt adequate to capture the amplitude dependence of the
amplification functions. Table B-4 lists the suite of distances for the singlecorner source model and Table B-6 lists the corresponding distances for the
double-corner model.
Amplification factors (5% damping response spectra) are then developed by
placing the site profile on the Mid-continent crustal model at each distance with
the input motion being equal to the reference rock motion convolved with a
diminution function which implements the site specific kappa (e.g. kappa from
the equations in Section B-5.1.3 and a 0.006s contribution from basement rock) ,
generating soil motions, and taking the ratios of site-specific response spectra
(5% damped) to hard rock reference site response spectra. For the higher levels of
rock motions, above about 1 to 1.5g for the softer profiles, the high frequency
amplification factors may be significantly less than 1, which may be exaggerated.
To adjust the factors for these cases an empirical lower bound amplification of
0.5 is to be implemented [18, 3].
g B-18 h
The general framework for the site response calculations are summarized in
Figure B-1 and Tables B-1 and B-3.
B-6.0 Development of Probabilistic Hazard Curves
The procedure to develop probabilistic site-specific soil hazard curves was
described by McGuire et al. (2001) and by Bazzurro & Cornell (2004) [24, 8].
That procedure (referred to as Approach 3) computes a site-specific soil hazard
curve for the spectral acceleration at a selected spectral frequency (or period)
given the site-specific hazard curve for the bedrock spectral acceleration at the
same oscillator period and site-specific estimates of soil response. The soil
response is quantified through the period/frequency-dependent amplification
factor, AF(f). The function AF(f) is given by:
AF(f) = SaSOIL(f)/SaROCK(f),
where f is frequency, and SaSOIL(f) and SaROCK(f), are the 5% damped spectral
accelerations at the soil surface and bedrock, respectively. Since the near-surface
materials frequently exhibit nonlinear behavior, the variation of AF(f) with input
intensity needs to be captured. Most commonly the input intensity is quantified
by SaROCK at the frequency of interest.
In the fully probabilistic approach, the annual probability of exceedance of soil
ground motion level z (GZ(z)) at spectral frequency f is computed as:
Where is the probability that AF is greater than the quantity z/x
given a bedrock amplitude of x, and fX(x) is the probability density function of
SaROCK.
In discretized form, the above equation can be expressed as:

Where px(xj) is the annual probability of occurrence for SaROCK equal to xj. This
probability is obtained by differentiating the appropriate rock hazard curve.
Then, can be computed by assuming AF is lognormally

distributed and a function of x, since
Where PlnAF|x is the mean value of ln AF given SaROCK = x, and VlnAF|x is the
standard deviation of ln AF given SaROCK= x. The term for is simply the
standard Gaussian cumulative distribution function. The parameters PlnAF|x and
VlnAF|x are obtained from the distribution for AF derived from the site response
analyses described above, and are a function of bedrock amplitude x.
g B-19 h
The site amplification functions are to be developed as described in Sections B-1

through B-5. As discussed in those sections, multiple models of site amplification
functions are derived. To compute site-specific hazard results using the equations
above, these multiple models are to be combined, with associated weights (See
Figure B-1 and Table B-1), to derive overall log-mean and log-standard
deviation values for each spectral frequency. For each spectral frequency and
input rock amplitude, the total log-mean, PT ( PlnAF|x in the equation above), and
log-standard deviation, VT (VlnAF|x in the equation above), are calculated as:

Where i indicates individual site amplification models, wi is the weight on each

model, and Pi and Vi are the log-mean and log-standard deviation of each site
amplification model, i.
B-7.0 Hazard-Consistent, Strain-Compatible Material
Properties (HCSCP)
In the approach to develop site-specific motions (hazard curves or
probabilistically-based design response spectra), a probabilistic method should be
used which correctly preserves the annual exceedance probabilities of the generic
hard rock PSHA, while properly incorporating variabilities (aleatory and
epistemic) in site-specific dynamic material properties. For structural analyses,
strain-compatible material properties are desired which are consistent with the
probabilistically-based design motions [46]. To achieve hazard consistency,
strain-compatible dynamic material properties, must reflect both the hazard level
(ground motion and exceedence frequency) as well as the aleatory and epistemic
components in site-specific dynamic material properties that have been
incorporated in developing the soil motions for design or risk analyses.
Simply using control motions based on a generic rock site hazard to drive the
site-specific soil column (Approaches 1 or 2 [24, 45]) will, in general, not result
in strain-compatible properties consistent with the site-specific soil hazard
developed using a fully probabilistic approach (Approaches 3 or 4). Additionally,
this approach is not viable for cases where the reference site hazard was
developed for generic soil or soft rock site conditions. For such cases, the rock
control motions which are appropriate for the base of the soil conditions are not
generally available.
As a result, if a fully probabilistic approach has been employed to develop the
site-specific hazard (such as Approach 3) then an approach should be employed
to develop the strain-compatible dynamic properties that preserves that hazard
consistency. The approach may assume that strain-compatible properties are
approximately lognormally-distributed, consistent with observed strong ground
motion parameters [2], and make use of the distributions of strain-compatible
properties catalogued during development of the suites of amplification factors.
Median strain compatible properties and estimates of sigma in those properties
g B-20 h
can be determined as a function of rock spectral acceleration for each soil layer in
the same manner that a median amplification and associated sigma are
determined. Instead of cataloging amplification values associated with Figure B1/Table B-1, final strain values are catalogued from which strain-compatible
dynamic material properties are then derived.
Specifically, the approach should examine the site-specific horizontal or vertical
soil hazard curves at the annual exceedance frequencies (AEF) of interest,
determine the ground motion (interpolating logarithmically as necessary), and
locate the corresponding amplification factors and associated strain-compatible
properties at the ground motion levels determined from the hazard curve. To
accommodate epistemic variability in site-specific properties, the same weights
used in developing the site-specific hazard curves (Figure B-1 and Table B-1) are
applied to the corresponding strain compatible properties. The weighted median
(mean log) set of strain compatible properties (for each layer) and log standard
deviations, reflecting upper- and lower-range properties, are given by the last two
equations in Section B-6.0. The associated variance includes both the aleatory
component for each epistemic case as well as the variability of mean properties
for each base case.
To examine consistency in strain-compatible properties across structural
frequency, the entire process is performed at PGA (typically 100 Hz), and again
at low frequency, typically 1 Hz. If the differences in properties at high- and lowfrequency are less than 10%, the high-frequency properties may be used since this
frequency range typically has the greatest impact on soil nonlinearity. If the
difference exceeds 10% the HCSCP developed at PGA and those developed at 1
Hz may be combined with equal weights.
The above approach results in a suite of HCSCP consistent with Approaches 3
or 4 and the intent of [46]. These results can then be used to develop bestestimate and upper and lower range values for strain-compatible dynamic
material properties in each soil layer that can be directly utilized in soil-structure
interaction calculations [46]
g B-21 h
B-8.0 References
1.
-surface attenuation and site-effects from

Amer., 87, p. 731-744.
2.
. Research
Lett., 68(1), 9-23.
3.
68(1), 94-127.
the
4.
Seism. Soc. Am., 74(5), 1969-1993.
5. Atkinson, G. M., (1993). Earthquake source spectra in eastern North
America, Bull. Seismol. Soc. Amer., 83, p. 1778-1798.
6. Atkinson, G. M., and D.
30.
7. Atkinson, G. M., and W. J. Silva, (2000). Stochastic modeling of California

ground motions: Bull. Seismol. Soc. Amer., 90, p. 255-274.
8. Bazzurro, P.
94, p. 2090-2109.
9.
-frequency
eq
y ground
gr
motions based on seismological mo
Soc. Am., 73(6), 1865-1894.
10.
-676.
elations
11.
Bull. Seism. Soc. Am., in-press.
-wave Q and o for
12.
Bull. Seism. Soc. Am, 99(4), 2365-2392.
13. Chapman, M. C., P. Talwani, and R. C. Cannon (2003). Ground-motion

attenuation in the Atlantic Coastal Plain near Charleston, South Carolina,
Bull. Seism. Soc. Am., 93, 998 1011.
14.
strong-
-44.
15. Cramer, C. H., J. S. Gomberg, E. S. Schweig, B. A. Waldron, and K.

Tucker (2004). The Memphis, Shelby Country, Tennessee, seismic hazard
maps, U.S. Geol. Surv. Open-File Rept. 04-1294.
g B-22 h
16. Cramer, C. H., (2004). Site-specific seismic-hazard analysis that is

completely probabilistic, Bull. Seismol. Soc. Amer., 93, p. 1841-1846.
abilistic Seismic Hazard Evaluation at
17. EPRI NP-6395Nuclear Plant Sites in the Central and Eastern United States: Resolution of
g design
gn
18.
vol. 1-5, EPRI TR-102293.
19.
Palo Alto, Calif: Electric Power Research Institute, Final Report.
20.
-Dependent
Attenuation Relations of Seismic Motion Parameters at Depth Using
-1804.
21. Hanks, T. C., and H. Kanamori, (1979). A moment magnitude scale, Jour.
Geophys. Res., 84, p. 2348-2350.
-wave
22.
BSSA, 85(6), 1805-1820.
23.
-point
po appr
approximations for
-695.
24.
for Revision of Regulatory Guidance on Design
ig Ground Motions: Hazard
and Risk-Consistent Ground Motion Spect
6728, U.S. Nuclear Regulatory Commission, Washington, D.C.
25. NRC NUREG/CR-5250, (1989). Seismic Hazard Characterization of 69
Nuclear Plant Sites East of the Rocky Mountains, U.S. Nuclear Regulatory
Commission, Washington, D.C.
26. NRC NUREGCommission, Washington, D.C., (published jointly with EPRI and DOE).
(2 2) (E.
(E Leeds and M. Johnson)) Letter to All Power Reactor
27. NRC (2012).
of Federal Regulations 50.54(f) Regarding Recommendations 2.1, 2.3, and
9.3 of the Near-Term Task Force Review of Insights from the Fukishima
Daih 12, 2012.
28.
GSA, Boulder, CO.
29. Pehlivan, M., E. M. Rathge, and R. B. Gilbert, (2012). Incorporating SiteSpecific Site Response into PSHA, The Second International Conference on
Performance-Based Design in Earthquake Geotechnical Engineering,
Taormina, Italy.
g B-23 h
30.
d Dynamic Analysis and Design
Considerations for High Level Nuclear Waste Repositories, ASCE 132-161.
31.
Electric Power Research Institute, Final Report RP 2556-48.
32. Silva, W.J., N. Abrahamson, G. Toro and C. Costantino. (1996).
(
).
Submitted to Brookhaven National Laboratory, Associated Universities, Inc.
Upton, New York 11973, Contract No. 770573.
33.
response
sp
for the Loma Prieta, Northridge,
ge and Imperial
pe
Valley California
onal Symposium on The
effects of Surface Geology on Seismic Motion Seismic
Motion/Yokohama/Japan/1-3
December 1998, Irikura, Kudo, Okada &
pa
Sasatani (eds.), 1137 1143.
34.
of a VS30 map
p for central and eastern North Mmerica (CENA):
(C
median
35.
strong
ng motion amplification
pl
factors for the San Francisco Bayy and Los
Angel
SA2120-59652.
ag N. Gregor,
eg G. Martin, C. Kircher, N. Abrahamson
36. Silva, W.J., R. Darragh,
-fault factors for building
SGS Grant award #98-HQ-GR-1010.
37.
Generic and Siteas an appendix in Silva et al (1996).
38. Van Houtte, C., S. Drouet, and F. Cotton
Bull. Seism. Soc. Am., v.101, p. 2926-2941.

39.
site conditions and amplifica
-1395.
40.
Spectra, 24(1), 243-255.
map
p of ggeologically
gi y
41.
defined site96(4A), 1483-1501.
-wave velocityy characteristics of

-556.
42.
g B-24 h
43. Y
Letters, 81(2), 294.
-root impedence
mp
44.
method for computing si
Amer., in press.
45.
-Based Approach
Ap
to
Define the Site-
46. USNRC, DC/COL-ISGHazard-Consistent Seismic Input for Site Response and Soil-Structure
47. Duke Energy (2010), (B. Dolan), Letter to NRC, William States Lee III
Nuclear Station Units 1 and 2, Transmittal of Unit 1 Foundation Input
Response
Spectra
Analysis,
sp
pe
((FIRS),
), Horizontal and Vertical Component
mp
-Specific Hazard
Consistent Uniform Hazard Response Spectra at the Lee Nuclear Station
g B-25 h
Tables
Table B-1.
Site Independent Relative Weights and Epistemic Uncertainty
Parameter
VP
Relative
Weight
Mean Base-Case Profile
0.40
Lower-Range
0.30
Upper-Range
0.30
Mean Base-Case Kappa
0.40
Lower-Range
0.30
Upper-Range
0.30
0.35
0.40
G/Gmax and Hysteretic Damping Curves
0.15*,
0.30**
Soil
EPRI Cohesionless Soil
0.5
Peninsular Range
0.5
Firm Rock
EPRI Rock
0.5
Linear
0.5
* Modulus variability at cyclic shear strain 3 x 10-2%

** Shear-wave damping variability at cyclic shear strain 3 x 10-2%
Table B-2.
Model to Estimate Density from Shear-Wave Velocity
Shear-Wave Velocity (m/s)
Density (g/cm3)
<500
1.84
500 to 700
1.92
700 to 1,500
2.10
1,500 to 2,500
2.20
>2,500
2.52
g B-26 h
Table B-3.
Maximum Number of Models to Characterize Epistemic
Uncertainty
Parameter
Maximum
Soil
Firm Rock
Soil/Firm Rock
Profile
Curves
Kappa
Magnitude
1,2-Corner
Total Models
72
12
18
36
Table B-4
Suite of Hard Rock Peak Accelerations, Source Epicentral
Distances, and Depths (M 6.5; 1-corner source model)
Expected Peak
Acceleration (%g)
Distance (km)
Depth (km)
230.00
8.0
74.00
8.0
10
45.00
8.0
20
26.65
8.0
30
18.61
8.0
40
13.83
8.0
50
10.45
8.0
75
4.59
8.0
100
0.0
7.0
125
0.0
5.6
150
0.0
4.7
Additional parameters used in the point-source model are:

(1-corner) = 110 bars
= 2.71 cgs
= 3.52 km/s
RC = 60 km, crossover hypocentral distance to R-0.5 geometrical
attenuation
T = 1/fc + 0.05 R, RVT duration, R = hypocentral distance (km)
Qo = 670
= 0.33
kappa(s) = 0.006
g B-27 h
Table B-5
Generic CEUS Hard Rock Crustal Model
Thickness (km)
Vs (km/sec)
U (cgs)
2.83
2.52
11
3.52
2.71
28
3.75
2.78
--
4.62
3.35
Table B-6
Suite of Hard Rock Peak Accelerations, Source Epicentral
Distances, and Depths (M 6.5; 2-corner source model)
Expected Peak Acceleration
(%g)
1
5
10
20
30
40
50
75
100
125
150
Distance (km)
Depth (km)
230.00
81.00
48.00
28.67
20.50
15.60
12.10
6.30
0.0
0.0
0.0
8.0
8.0
8.0
8.0
8.0
8.0
8.0
8.0
7.9
6.4
5.4
Table B-7
Geometrical spreading and attenuation models for the CEUS
and WUS
Region
CEUS
WUS
Geometric Spreading
1/R
(1/60)(60/R)
0.5
1/R
(1/35)(35/R)
0.5
for
R 60km
for
R > 60km
for
R 35km
for
R > 35km
g B-28 h
Anelastic
Attenuation
Q(f) = 670f
0.33
Q(f) = 270f
0.60
Figures
Figure B-1
Logic tree illustrating the process for capturing
uncertainty in the development of site-specific
amplification functions.
This illustration is for a site with limited at-site geophysical and geotechnical
data available. UR and LR indicate Upper-Range and Lower-Range about the
mean Base-Case model.
g B-29 h
500

1000
1500
2000
2500
3000
0
100
50
200
300
100
150
500
Depth (ft)
Depth (m)
400
600
200
700
800
250
900
300
0
2000
4000
6000
8000
Shear-Wave Velocity (ft/s)

Vs 190 m/s
Vs 400 m/s
Vs 760 m/s
Vs 270 m/s
Vs 560 m/s
Vs 900 m/s
Figure B-2
Template Shear Wave Velocity Profiles for Soils, Soft Rock,
and Firm Rock. Rock Profiles Include Shallow Weathered
Zone. Indicated velocities are for VS30.
g B-30 h

0
500
1000
1500
2000
2500
3000
0
-20
-40
-60
Depth (m)
-80
-100
400m/s Template
Vs Measured
Adjusted Template
-120
Firm Rock Gradient
-140
-160
-180
Base Case Model

Vs Estimated
Estimated Firm Rock

Depth
Basement Vs (2830)
-200
Figure B-3
Illustration of how available information is used to
develop a mean base-case profile.
The information available is represented by the measured near-surface soil VS30

(solid black line), estimated depth to firm rock (solid brown line) and estimated
firm rock VS (solid orange line). Proposed mean base-case VS profile is indicated
by dashed red line.
g B-31 h
Shear Wave Velocity (ft/s)

0
2000
4000
6000
8000
10000
200
Depth (ft)
400
600
800
1000
Template Velocity Profile
Template Velocity
ln
1200
Figure B-4
Illustration of the range of velocity
This shows a case implied by the method used to account for epistemic
uncertainty in site specific shear wave velocity profiling where sparse or limited
information is available. Displayed is the 760 m/s WNA reference rock template
velocity (solid curve) with dashed curves representing ln = 0.35.
g B-32 h
Shear Wave Velocity (ft/s)
2000
4000
6000
8000
10000
12000
200
Depth (ft)
400
600
800
1000
Base Case Template Velocity Profile

Template Velocity Factor: +/-0.45 ln units
Template Velocity Factor: +/-0.64 ln-units
1200
Figure B-5
Method used to account for epistemic uncertainty in site
specific shear wave velocity profiling where limited
information is available.
Displayed is the 760m/s reference template velocity (solid black curve) with
dashed curves representing 10th and 90th-percentile values (0.45 natural log
units which corresponds to a ln = 0.35). Dotted red curves are for 0.64
natural log units which corresponds to a ln = 0.5.
g B-33 h
Figure B-6
Illustration of Upper Range and Lower Range Base-Case
profiles (10th and 90th percentiles) developed to represent
the epistemic uncertainty in the Mean Base-Case for firm
rock conditions.
A mean surface velocity of 5000ft/s (1525m/s) was assumed for the Base Case
and the empirical gradient of Fukushima et al. (1995) [19] was applied. A ln
= 0.35 was used.
g B-34 h

0
500
1000
1500
2000
2500
3000
0
-20
-40
-60
Depth (m)
-80
-100
Vs Measured
-120
-140
-160
-180
Base Case Model

Vs Estimated
Estimated Firm Rock

Depth
Upper Range (90th
percentile)
Lower Range (10th
percentile)
Basement Vs (2830m/s)
-200
Figure B-7
Illustration of the development of Upper Range and Lower
Range profiles to accommodate epistemic uncertainties for
the hypothetical example shown in Figure B-3.
A ln = 0.35 has been used to develop the 10th and 90th-percentile curves in
the upper portion of the profile where sparse Vs measurements were available. A
ln = 0.50 was applied to the lower portion of the profile where the Vs of the
Base Case was inferred from geological information. The 90th-percentile curve
was capped at a value equal to the 2830m/s Vs value assumed for the hard rock
basement.
g B-35 h
G/Gmax
0.8
0.6
0 - 20 ft
20 - 50 ft
0.4
50 - 120 ft
120 - 250 ft
0.2
250 - 500 ft
500 - 1000 ft
0
0.0001
0.001
0.01
0.1
Shear Strain (%)

30
0 - 20 ft
Damping (%)
25
20
15
20 - 50 ft
50 - 120 ft
120 - 250 ft
250 - 500 ft
500 - 1000 ft
10
5
0
0.0001
0.001
0.01
0.1
Shear Strain (%)

Figure B-8
Generic G/GMAX and hysteretic damping curves for
cohesionless soil [18]. Note that damping will be limited
to 15% for this application.
g B-36 h
Figure B-9
Comparison of median amplification functions (5%-damped
PSa) derived using the EPRI (1993) [18] (see Figure B-8)
and Peninsular Range [40]) G/GMAX and hysteretic damping
curves.
The results are for the 400 m/sec VS30 template profile and a single-corner
source model for reference rock loading levels of 0.01 to 1.50g.
g B-37 h
Figure B-10
Comparison of median amplification functions (5%-damped
PSa) derived using the EPRI (1993) [18] (see Figure B-8)
and Peninsular Range [40] G/GMAX and hysteretic damping
curves.
The results are for the 400 m/sec VS30 template profile and a single-corner
source model for reference rock loading levels of 0.01 to 1.50g.
g B-38 h
G/Gmax
0.8
0.6
0.4
0.2
0
0.0001
0 - 20 ft
21 - 50 ft
51 - 120 ft
121 - 250 ft
251 - 500 ft
501 - 1000 ft
1001 - 2000 ft
2001 - 5000 ft
0.001
0.01
0.1
0.1
Shear Strain (%)

40
Damping (%)
30
20
0 - 20 ft
21 - 50 ft
51 - 120 ft
121 - 250 ft
251 - 500 ft
501 - 1000 ft
1001 - 2000 ft
2001 - 5000 ft
10
0
0.0001
0.001
0.01
Shear Strain (%)

Figure B-11
Generic G/GMAX and hysteretic damping curves developed for
firm rock in the EPRI (1993) study [12] (from Dr. Robert
Pyke).
Note that damping is limited to 15% in this application.
g B-39 h
Figure B-12
Illustration of effect of various factors in the simple
seismological model on Fourier spectral shape.
g B-40 h
Figure B-13
Illustration of effect of various factors in the simple
seismological model on response spectral shape.
g B-41 h
Figure B-14
Comparison of amplification functions (5%-damped PSa)
computed for magnitudes of M 5.5, 6.5, and 7.5, using the
single-corner source model and the 400 m/sec VS30 stiffsoil template profile (Figure B-2) with the EPRI (1993)
[18] G/GMAX and hysteretic damping curves (Figure B-8).
The input reference rock loading levels varied from 0.01 to 1.50 g (Table B-4).
g B-42 h
Figure B-15
Comparison of amplification functions (5%-damped PSa)
computed for magnitudes of M 5.5, 6.5, and 7.5.
This Figure uses the single-corner source model and the 400 m/sec VS30 stiffsoil template profile (Figure B-2) with the EPRI (1993) [18] G/GMAX and
hysteretic damping curves (Figure B-8). The input reference rock loading levels
varied from 0.01 to 1.50 g (Table B-4).
g B-43 h
Figure B-16
Comparison of amplification functions (5% damped PSa)
computed using the Single- and Double-Corner source models
(Tables B-4 and B-6) for the 400 m/sec VS30 stiff-soil
template profile (Figure B-2) with the EPRI (1993) [18]
G/GMAX and hysteretic damping curves (Figure B-8).
The input reference rock loading levels varied from 0.01 to 1.50 g (Tables B-4
and B-6).
g B-44 h
Figure B-17
Comparison of amplification functions (5% damped PSa)
computed using the Single- and Double-Corner source models
(Tables B-4 and B-6) for the 400 m/sec VS30 stiff-soil
template profile (Figure B-2) with the EPRI (1993) [18]
G/GMAX and hysteretic damping curves (Figure B-8).
The input reference rock loading levels varied from 0.01 to 1.50 g (Tables B-4
and B-6).
g B-45 h
Appendix C: Sensitivity Studies

to Develop Criteria for
Analyzing Rock-Founded
Structures as Fixed-Base
Models
Two examples of models of existing structures at a nuclear power plant were
analyzed in a study to compare the ISRS developed from a fixed-base dynamic
analysis with the ISRS obtained from SSI analyses with various shear wave
velocities. The first example analyzed a containment structure with a
fundamental frequency of about 5 Hz, and the second example was for a Main
Steam Valve House (MSVH) structure which has a fundamental frequency of
about 10 Hz in one horizontal direction.
C1.0 Containment Structure
The containment structure, which has a horizontal fundamental mode of about
5 Hz, was analyzed [45] using a fixed-base model and, subsequently, with SSI
analyses using three sets of shear wave velocities (Vs) for this site: lower bound
(about 3,400 ft/sec), best estimate (about 5,200 ft/sec) and an upper bound
(about 7,900 ft/sec). Figures C-1, C-2 and C-3 show the results of the analyses
at the operating deck of the structure (about 75 ft above the basemat) in the eastwest, north-south, and vertical directions, respectively. All three figures compare
the results from SSI analyses using the lower and best estimate shear wave
velocity values with the results of the fixed-base analysis. The results from the
SSI analysis using the upper bound Vs were very close to the fixed-base results
and therefore are not shown in this Appendix but are included in [45].
From these figures, it can be seen that for the east-west and north-south
directions, the lower bound Vs case resulted in a slight shift of frequency of the
spectral peak as compared to the fixed-base model (about 1 Hz lower with SSI)
because of the rotational effects. For the best estimate case (Vs of 5200 ft/sec),
the frequency shift was smaller in the north-south direction, and there was no
shift in the east-west direction. In the vertical direction, the two SSI cases gave
g C-1 h
identical ISRS, and there was also no frequency shift compared to the fixed-base
case. In all three directions, the spectral peaks determined from the fixed-base
analysis were higher than those from the SSI analyses, and the shapes of the
ISRS in the entire frequency range remained about the same. Therefore, based
on the analysis of this structure, a fixed-base analysis can be used for sites with
rock Vs > 3,500 ft/sec only if one can accept a slight potential frequency shift of
the spectral peak. If not, a fixed-base analysis is only recommended for rock Vs >
5,000 ft/sec. It is noted that peak-broadening or peak-shifting of the ISRS in
fragility analyses can alleviate the effect of a slight frequency shift between the
SSI and fixed-base analyses.
Figure C-1
Reactor Containment Internal Structure, Operating Floor
East-West Direction, 5% Damping
g C-2 h
1.8
Lower Bound (LB), Vs = 3400 fps
1.6
Best Estimate (BE), Vs = 5200 fps

1.4
Fixed Base Analysis
Acceleration (g)
1.2
0.8
0.6
0.4
0.2
0
0.1
10
100
Frequency (Hz)
Figure C-2
North-South Direction, 5% Damping
1.2
Lower Bound (LB), Vs = 3400 fps
Best Estimate (BE), Vs = 5200 fps
Fixed Base Analysis
Acceleration (g)
0.8
0.6
0.4
0.2
0
0.1
10
100
Frequency (Hz)
Figure C-3
Vertical Direction, 5% Damping
C2.0 Main Steam Valve House Structure

The MSVH structure, which has a fundamental mode of about 10 Hz in one
horizontal direction, was analyzed [55] using a fixed-base model and,
subsequently, with SSI analyses using two sets of shear wave velocities (Vs):
g C-3 h
3,500 ft/sec and 5,200 ft/sec (best estimate Vs for this site). Three sets of input
motions were used in the analysis; however, ISRS for only two sets of
representative motions are presented in this Appendix (results of the third set of
input motions are about the same as the first set and are documented in [55]).
The two set of results presented here are: (a) ISRS using time histories derived
from the design basis earthquake (DBE) ground spectrum shape of this plant,
which is similar to that of RG 1.60, and (b) ISRS using time histories derived
from GMRS from a recent PSHA for a hard rock site that contains high
frequency content.
Figures C-4, C-5 and C-6 show the ISRS for case (a) and Figures C-7, C-8 and
C-9 show the ISRS for case (b), in the east-west, north-south, and vertical
directions, respectively. These ISRS are compared at a node (Elevation 306 ft)
that is about 60 ft above the basemat of the structure. These figures compare the
results from the fixed-base case and the SSI cases with (Vs) = 3,500 ft/sec case
and (Vs) = 5,200 ft/sec.
From these figures, it can be seen that the comparison of fixed-base ISRS with
the (Vs) = 5,200 ft/sec case is reasonably good for all directions, whereas the
lower bound Vs =3500 ft/sec case results in a slight shift of frequency and the
spectral peaks are also slightly different in comparison to the fixed-base model.
The shapes of the ISRS in the entire frequency range remain about the same
from all these analyses. Therefore, based on the analysis of this structure, a fixedbase analysis is appropriate for rock sites with Vs > about 5,000 ft/sec. A fixedbase analysis can be used for sites with rock Vs > 3,500 ft/sec only if one can
accept potential small frequency and amplitude shifts of the spectral peaks.
Again, peak-broadening or peak-shifting of the ISRS in fragility analyses can
alleviate the effect of a slight frequency shift between the SSI and fixed-base
analyses.
g C-4 h
1.8
1.6
DBE_N14_5%_X_Hard Rock
DBE_N14_5%_X_Vs = 3500 fps
1.4
DBE_N14_5%_X_Vs = 5200 fps
Acceleration (g)
1.2
0.8
0.6
0.4
0.2
0
0.1
10
100
Frequency (Hz)
Figure C-4
Main Steam Valve House, Elevation 306 ft
East-West Direction, 5% Damping
2.5
DBE_N14_5%_Y_Hard Rock
2
DBE_N14_5%_Y_Vs = 3500 fps

DBE_N14_5%_Y_Vs = 5200 fps
Acceleration (g)
1.5
0.5
0
0.1
10
Frequency (Hz)
Figure C-5
North-South Direction, 5% damping
g C-5 h
100
0.6
DBE_N14_5%_Z_Hard Rock
0.5
DBE_N14_5%_Z_Vs = 3500 fps
DBE_N14_5%_Z_Vs = 5200 fps
Acceleration (g)
0.4
0.3
0.2
0.1
0
0.1
10
100
Frequency (Hz)
Figure C-6
Vertical Direction, 5% damping
4.5
COLA_N14_5%_X_Hard Rock
COLA_N14_5%_X_Vs = 3500 fps

3.5
COLA_N14_5%_X_Vs = 5200 fps
Acceleration (g)
2.5
1.5
0.5
0
0.1
10
Frequency (Hz)
Figure C-7
East-West Direction, 5% damping
g C-6 h
100
3.5
COLA_N14_5%_Y_Hard Rock
3
COLA_N14_5%_Y_Vs = 3500 fps

COLA_N14_5%_Y_Vs = 5200 fps
Acceleration (g)
2.5
1.5
0.5
0
0.1
10
100
Frequency (Hz)
Figure C-8
North-South Direction, 5% damping
6
COLA_N14_5%_Z_Hard Rock
5
COLA_N14_5%_Z_Vs = 3500 fps

COLA_N14_5%_Z_Vs = 5200 fps
Acceleration (g)
0
0.1
10
Frequency (Hz)
Figure C-9
Vertical Direction, 5% damping
g C-7 h
100
Appendix D: Sensitivity of
Computed Annual Probability
of Failure PF to Assumed
Logarithmic Standard
Deviation Used in Hybrid
Method with Capacities
Defined by 1% Failure
Probability Capacity C1%
D1.0 Introduction
In the Hybrid Method, the 1% failure probability capacity C1% is computed by
the CDFM Method. Then an estimate of the composite logarithmic standard
deviation C and its subdivision into random variability R and uncertainty U are
used to estimate the corresponding fragility curve. As noted in Ref. 1, typically C
lies within the range of 0.3 to 0.6. In fact, if all of the sources of variability
discussed in Ref. 2 are appropriately considered, it is not possible to obtain an
estimated C less than approximately 0.3.
The Hybrid Method is based on the observation that the annual probability of
unacceptable performance PF for any Seismic Source Characterization is relatively
insensitive to C. Thus, annual probability (seismic risk) can be computed with
adequate precision from the CDFM Capacity, CCDFM, and an estimate of C. It is
concluded in Ref. 1 that the computed seismic risk at = 0.3 is approximately 1.5
times that at = 0.4, while at = 0.6 the computed seismic risk is approximately
60% of that at = 0.4.
In Section 3, it is demonstrated that the Ref. 1 conclusion concerning the lack of
sensitivity of the computed seismic risk remains valid over the full practical range
of hazard curve slopes. This demonstration uses the Simplified Seismic Risk
Equation defined in Section 2.
g D-1 h
D2.0 Simplified Seismic Risk Equation

Typical seismic hazard curves are close to linear when plotted on a log-log scale
(for example see Figure 1). Thus over any (at least) ten-fold difference in
exceedance frequencies such hazard curves may be approximated by a power law:
K I a KH
H (a)
Equation 1
where H(a) is the annual frequency of exceedance of ground motion level a, K I is

an appropriate constant, and KH is a slope parameter defined by:
1
logA R
KH
Equation 2
in which AR is the ratio of ground motions corresponding to a ten-fold reduction

in exceedance frequency defined by:
SA0.1H
SAH
AR
Equation 3
where SAH is the spectral acceleration at the mean exceedance frequency H and
SA0.1H is the spectral acceleration at 0.1H.
So long as the fragility curve PF(a) is lognormally distributed and the hazard curve
is defined by Equation 1, a rigorous closed-form solution exists for the seismic
risk. This closed-form solution has been derived in a number of references
including Appendix C of Ref. 1 and Appendix A of Ref. 3:
H F50%K H eD
PF
Equation 4
in which
F50%
C50%
CH
Equation 5
KHE2
Equation 6
and
1
2
where H is any reference exceedance frequency, CH is the UHRS ground motion

level that corresponds to this reference exceedance frequency H from the seismic
hazard curve, C50% is the median fragility capacity, and is the logarithmic
standard deviation of the fragility.
g D-2 h
Equation 4 is referred to herein as the Simplified Seismic Risk Equation. The

only approximations in its derivation are that the hazard curve is approximated by
Equation 1 over the exceedance frequency range of interest and the fragility curve
is lognormally distributed.
By defining CH at the 1% failure probability capacity C1%, then H is replaced in
Equation 4 by H1% corresponding to the annual frequency of exceeding C1%.
With these definitions for CH and H, Equations 4 through 6 can be rewritten as:
F50% = (C50%/C1%) = e2.326
Equation 7
(PF/H1% ) = eD1
Equation 8
1 = 0.5 (KH)2
2.326 (KH)
Equation 9
From Equations 8 and 9, it can be observed that the ratio (PF/H1%) is only a
function of the product (KH).
D3.0 Sensitivity of Failure Probability PF to
Table D-1 presents the ratio (PH/H1%) computed over the full practical range of
ground motion ratios AR from 1.5 to 4.5 and values ranging from 0.3 to 0.6
using the Simplified Seismic Risk Equation.
As shown in Table 4 of Reference 1, the failure probability PF values computed
by the Simplified Seismic Risk Equation tend to be about 10% to 20%
conservatively biased when compared to PF computed by numerical convolution
of the hazard curve and fragility curve. This bias is due to the slight downward
curvature of the hazard curve when plotted on a log-log plot as is shown in
Figure D-1. Even with this slight conservative bias, the Simplified Seismic Risk
Equation can be used to compare PF values computed for different values at any
specified AR ratio
AR values ranging from 2.0 to 4.5 are typical for CEUS sites over the 10-4 to 10-6
annual frequency of exceedance range. Within this AR range, the (PF/H1%) ratios
shown in Table D-1 for = 0.3 range from 1.32 to 1.44 of the (PF/H1%) ratios
shown for = 0.4 with an average ratio of about 1.4. Many numerical
convolutions of hazard and fragility curves have confirmed the conclusion that PF
computed using = 0.3 will be less than about 1.5 times the PF computed using
= 0.4 for the same C1% capacity.
Over this same AR range from 2.0 to 4.5, the (PF/H1%) ratios shown in Table D-1
for = 0.6 range from 58% to 64% of the (PF/H1%) ratios computed for = 0.4
with an average of about 60%. Again, many numerical convolutions of hazard
and fragility curves have confirmed this value of about 60%.
g D-3 h
For high seismic western sites, the AR values will typically range between 1.5 and
2.25 over the 10-4 to 10-6 annual frequency range. For AR values less than about
1.8, the lower tail of the lognormal fragility curve below the 1% failure
probability capacity C1% begins to significantly affect computed PF. Many
experienced seismic capacity engineers question the validity of extending the
lower tail of the lognormal fragility curve substantially below the C1% capacity.
However, if one conservatively includes this lower tail on the lognormal fragility
curve, the resulting (PF/H1%) ratios computed by numerical convolution of hazard
and fragility are similar to the results shown in Table D-1. As AR is reduced
below 1.75, the ratio of (PF/H1%) for = 0.3 to (PH/H1%) for = 0.4 will begin to
rapidly reduce below 1.4. Conversely, as AR is reduced below 2.0, the ratio of
(PH/H1%) for = 0.6 to (PH/H1%) for = 0.4 will begin to rapidly increase. In fact
at AR = 1.5, the (PH/H1%) ratio for = 0.6 will be 1.8 times that for = 0.4. This
unexpected result is directly attributable to the tail of the lognormal fragility
curve below the C1% capacity. For sites with AR less than 1.75:
1)
one should assess whether it is appropriate to extend the lognormal

fragility curve below the C1% capacity, and
2)
if the fragility curve is extended below C1%, one needs to carefully

estimate for those situation where might exceed 0.5 (i.e., active
components mounted at high elevation in structures).
D4.0 References
1. Kennedy, R.P. Overview of Methods for Seismic PRA and Margin Analysis
Including Recent Innovations, Proceedings of the OECD-NEA Workshop
on Seismic Risk, Tokyo, August 1999.
2. Methodology for Developing Seismic Fragilities, EPRI TR-103959, Electric
Power Research Institute, June 1994.
3. Kennedy, R.P., Performance-Goal Based (Risk Informed) Approach for
Establishing the SSE Site Specific Response Spectrum for Nuclear Power
Plants. Nuclear Engineering and Design 241, pp. 648-656, 2011 (also
Transactions SMiRT 19 Toronto, August, 2007).
4. Risk Engineering, Inc., REI (2001), Technical Basis for Revision of
Regulatory Guidance on Design Ground Motions: Hazard- and Riskconsistent Ground Motion Spectra Guidelines, US Nuclear Regulatory
Commission Report NUREG/CR-6728, October, 2001.
g D-4 h
Table D-1
Sensitivity of (PF/H1%) to Estimated over
Range of AR Values
AR
1.5
1.75
2.0
2.5
3.0
3.5
4.0
4.5
KH
5.68
4.11
3.32
2.51
2.10
1.84
1.66
1.53
(PF/H1%)
=0.3
0.081
0.121
0.162
0.230
0.282
0.323
0.355
0.382
=0.4
0.067
0.084
0.110
0.160
0.202
0.237
0.266
0.290
g D-5 h
=0.5
0.076
0.069
0.083
0.118
0.151
0.180
0.205
0.226
=0.6
0.120
0.068
0.070
0.093
0.118
0.141
0.162
0.180
Figure D-2
SA (10 Hz) and SA (1 Hz) hazard curves for the eleven sites
normalized by the acceleration value corresponding to mean
10-4 annual probability.
(From Figures. 7.7 and 7.8 of REI, 2001)
g D-6 h

SPID Seismic Evaluation Guidance

Uploaded by

Copyright:

Available Formats

SPID Seismic Evaluation Guidance

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SPID Seismic Evaluation Guidance

Uploaded by

Copyright:

Available Formats

Seismic Evaluation

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

This report describes research sponsored by EPRI. The following

Simpson Gumpertz & Heger Inc.

Pacific Engineering and Analysis

Southern Nuclear Operating Company, Inc.

Following the accident at the Fukushima Daiichi nuclear power

Results and Findings

Prioritization (Schedule) ....... 5-1

References ...................... 8-1

Appendix A: Control Point Discussion from

B-5.1.4 Source Model ..................... B-17

Figure 1-1 Recommended Approach to Respond

Example Narrow Exceedances at 2 Hz

and 6 Hz (5% Damping) .......................... 3-5

Figure 6-7 Potentially High-Frequency Sensitive Component

Figure B-8 Generic G/GMAX and hysteretic damping curves

Figure B-17 Comparison of amplification functions (5% damped

Table 3-1 EPRI 1015109 Potentially High Frequency Sensitive

Items ...................................... 3-12

annual exceedance frequency

American Nuclear Society

American Society of Mechanical Engineers

boiling water reactor

Cumulative Absolute Velocity

core damage frequency

Conservative Deterministic Failure Margin

Central and Eastern North America

Central and Eastern United States

Combined Operating License Application

early site permit

Final Safety Analysis Report

Filtered Random Multi-Frequency

Generic Issues Program

ground motion prediction equations

ground motion response spectra

High Confidence of Low Probability of

Institute of Electrical and Electronics

IPEEE HCLPF Spectrum

Individual Pant Examination of Individual

Interim Safety Guide

in-structure response spectra

large-safety release frequency

Low Hazard Threshold

Lumped-mass stick models

motor control center

Molded case circuit breakers

Main Steam Valve House

Next Generation Attenuation

nuclear power plant

Nuclear Regulatory Commission

Near Term Task Force

Peak Ground Acceleration

power spectral density

probabilistic seismic hazard analysis

review level earthquake

Repeated Large Magnitude Earthquake

Random Variation Theory

Severe Accident Management Guidance

seismic core damage frequency