2013 EERI Annual Meeting Session: Unusual Earthquakes and their Implications for Risk

1. in US Nuclear Facilities
Annie Kammerer
EERI Annual Meeting
February 2013

2.  This title implies there are “usual” and “unusual”
earthquakes. We don’t know enough to make that
distinction and the idea is counterproductive to goal
of reducing risk by instilling false confidence.
 In reality, there is a range of behaviors that can
occur on faults. We need to fully characterize the
distribution of behavior using a methodical
approach (and an open mind) if we hope to assess
ground motion and its inherent uncertainties for
any site (and to reduce risk).
 We shouldn’t confuse real risk and perceived risk

3.  Risk-informed regulatory framework
 Structured hazard and risk assessment methods
that provide both best estimates and uncertainties
(impact of rare events on risk can be quantified)
 Rare ground motion levels used in performance-
based design approaches
 10-4 to 10-5 annual probability of exceedance
 Designs are checked against risk objectives
 Defense-in-depth concept

4. • PSHA is the approach required by the NRC
• Regulatory Guide 1.208 described high level
requirements and points to the SSHAC
guidelines for implementation of PSHA
studies
• PSHA is input to performance-based design
framework described in ASCE 43-05

5. • NUREG/CR-6372, “Recommendations for
Probabilistic Seismic Hazard Analysis: Guidance on
Uncertainty and Use of Experts”
• Developed in the 1980s as a result of differing NRC
and EPRI Seismic Hazard Assessment Studies - the
method used to engage experts differed more than
the technical input
• SSHAC provides a framework for incorporating
experts into scientific assessments through
structured processes and interactions

6. Original report
provides framework.
New report provides
additional details. Both
describe how to
undertake studies that
develop hazard
NUREG/CR-6372 NUREG 2117 assessment models
(1989) (2012)

7.  Objective is to develop a model that
represents the center, body and range of
technically defensible interpretations of the
available data
 Center-best estimate
 Body-shape of the distribution
 Range-extreme values of the distribution
 Achieved through a process with well
defined evaluation and integration phases

8.  Compilation of comprehensive databases
 made available to all participants
 Defined roles and responsibilities for participants
 Technical Integration (TI) Team: Evaluate data, methods and
models and develop distribution capturing center, body and
range of technically-defensible interpretations
 Participatory Peer Review Panel (PPRP): Continuous
process and technical review
 Resource Experts (neutral experts on a dataset or topic)
 Proponent Experts (support an interpretation or model)

9.  Structured sequence of steps, including 3
formal workshops
 WS1: Data needs and critical issues
▪ Probe the datasets available, identify and other
data, and identify and discuss the critical issues
 WS2: Proponent viewpoints and alternatives
▪ Proponents experts go through a process of
discussion, challenge and defense
 WS3: Investigation of the preliminary model

10. Technical Staff TI
& Contractors Team PPRP
Hazard sensitivity
Evaluation of Models to Form Composite Distribution
Preliminary database calculations
Database Compilation
Resource WORKSHOP 1: Hazard Sensitive
Issues and Data Needs
Experts
Process and Technical Review
Additional data collection & analysis
Resource Experts WORKSHOP 2: Review of Database and
Discussion of Alternative Models
Proponent Experts
Final database Preliminary SSC and
GMC models
WORKSHOP 3: Presentation of Models and
Hazard Sensitivity Feedback
Final SSC and GMC models, then final hazard calculations,
Documentation of all technical bases

11. Uncertainty
Aleatory Epistemic
Modeling or knowledge
Natural variability
uncertainty
Reducible with more
Not reducible
information
Addressed through integration Addressed through use of a
over parameter distributions logic tree

12. Uncertainty
Aleatory Epistemic
Integration over distribution of logic tree of technically
expected parameter values defensible interpretations

13. Uncertainty
Aleatory Annual Prob of Exceedance
Epistemic
Aleatory Epistemic
variability gives uncertainty leads
the curve its to uncertainty
shape. 85% bands
Median
15%
Acceleration (g)

14. Seismic Source Ground Motion
Characterization Model Characterization Model
Provides the characterization for all Provides a distribution of predicted
seismic sources that may impact a site of ground motions for a particular
interest. The SSC model is in the form of magnitude distance scenario earthquake.
a logic tree composed of the full suite of
The GMC model is in the form of a logic
alternative technically defensible
interpretations of the earth science data. tree composed of Ground Motion
The logic tree accounts for epistemic Prediction Equation (GMPEs) to account
uncertainty. Aleatory variability is for epistemic uncertainty. Each GMPE
incorporated for specific parameters as incorporates aleatory variability.
appropriate.
PUBLISHED
January 2012
PSHA

17. Seismic Source Ground Motion
Characterization Model Characterization Model
Provides the characterization for all Provides a distribution of predicted
seismic sources that may impact a site of ground motions for a particular
interest. The SSC model is in the form of magnitude distance scenario earthquake.
a logic tree composed of the full suite of The GMC model is in the form of a logic
alternative technically defensible
interpretations of the earth science data. tree composed of Ground Motion
The logic tree accounts for epistemic Prediction Equation (GMPEs). The logic
uncertainty. Aleatory variability is tree accounts for epistemic uncertainty.
incorporated for specific parameters as Each GMPE incorporates aleatory
appropriate. variability through “sigma”.
IN PROGRESS
January 2014
PSHA

18. Seismic Load Capacity
• Determined by PSHA • Fragility curves quantify
• Defined in terms of capacity of individual
hazard curves and structures, systems and
response spectra components
• Uncertainty is explicitly
quantified using modern
approaches
Risk

19. Seismic Load Capacity
Conditional Probability of Failure
Frequency of Exceedance
i
Pi
Seismic Motion Parameter Seismic Motion Parameter
Risk

20. Systems Analysis
Seismic Load Capacity
Event trees, Fault
Conditional Probability of Failure
Frequency of Exceedance
trees, Containment Analysis i
Pi
Seismic Motion Parameter
SPRA Seismic Motion Parameter

21. • Plant capacity and risk assessed in SPRA
• Systems model used to develop plant-level fragility
curves for core damage
• SPRA provides information on the SSCs that
contribute most to risk
• Plant capacity improved through systems design
and redundancy
SPRA

22. Defense-in-depth (IAEA INSAG-10 (1996))
1. The first level is prevention of abnormal operation and system failures
(good design, construction, and maintenance/operations)
2. If the first level fails, abnormal operation is controlled or failures are
detected by the second level of protection (appropriate response to
problems to bring plant back to normal operations)
3. The third level ensures that safety functions are further performed by
activating specific safety systems and other safety features (accident
prevention through additional emergency/safety systems)
4. The fourth level limits accident progression through accident management,
so as to prevent or mitigate severe accident conditions with external
releases of radioactive materials (accident mitigation and containment)
5. The last objective (fifth level of protection) is the mitigation of the
radiological consequences of significant external releases through the off-
site emergency response (emergency response)

23. Deterministic Approach
to Hazard
Probability and risk were
not considered

24. 24

25. Fukushima Lessons
Learned Report with
Significant
Recommendations
for changing NRC
regulations and
practices
NRC ML111861807

26. 26

27.  50.54(f) RFI Letter issued
March 12, 2012
 Enclosure 1: Seismic hazard
and risk reevaluation
 Enclosure 3 Seismic Walkdowns
27

28. Walkdowns to assure plants are meeting licensing
11/2012 basis and to look for potential seismic issues.
2.3 Walkdowns (+outages) Reports due November 2012. Some equipment
delayed until outage.
Ongoing
9/2013 Hazard evaluation due in 18 months for NPPs
Hazard (CEUS) within the CEUS SSC model area. 3 years for
3/2015 western US NPPs performing SSHAC level 3
evaluation (WUS) studies. Plant-specific site response.
Risk results due 3-4 years after hazard. SMAs only
Risk 3 years after allowed for small exceedance levels. SPRAs
2.1 hazard allowed for all exceedances, but required for large
evaluation exceedances.
Near term
Regulatory After receiving the information from the SPRA
Depends on
and SMA analyses, the NRC will determine
findings
Actions appropriate regulatory actions.
10 year Rulemaking Rulemaking to require a reevaluation every 10
Long term 2.2 timeline years.
update
28

29. Japanese Diet (congress)
report from the 1st
independent commission
ever formed in Japan.
Outstanding and complete
explanation, easy to read,
honest, technically superb
and fascinating!