Overview of Seismic
Isolation of Nuclear
Power Plants
Jenna Wong
PhD Candidate
UC Berkeley
PEER Annual Meeting
October 26, 2012
Outline
n 
n 
n 
n 
n 
Motivation
Nuclear Power Plant (NPP) Background
Seismic Isolation Background
Analysis
Conclusions and Future Research
Motivation
n 
US Energy Sector
n 
n 
n 
n 
n 
Increased Energy Demand and Environmental
Concerns
Potential for Nuclear Power Renaissance
Policy Perspective: Goal is to ensure the
safety and security of nuclear power plants
(NPPs)
Engineers: Improve design to address
concerns likely to be raised in the licensing
process
Seismic Isolation can be reliable means of
improving seismic safety
Motivation – PEER Project
n 
Long-term project sponsored by:
n 
n 
n 
Goal
n 
n 
Electric Power Research Institute (EPRI)
Korean Electric Power Corporation (KEPCO)
To understand the viability of seismic isolation
in NPPs using Performance Based Design
methodologies
Pilot Studies
n 
n 
Background Information
Preliminary Analysis
Nuclear Power Plant Background
n 
n 
n 
February 2012:
1st Nuclear Reactor
Construction Permit in
35 Years
Changes in licensing
and development of
NPPs since the 1970s
114 reactors in use
today
(NRC, 2011)
Behavior of Building Structure
Base Isolation
Background
with Base Isolation System
Conventional Structure
n 
n 
Base-Isolated Structure
(Symans)
Introduction of laterally flexible layer
between structure and foundation
Structure moves as a rigid body supported
by bearings
Instructional Material Complementing FEMA 451, Design Examples
Seismic Isolation 15 - 7- 7
Base Isolation Background
n 
n 
n 
Period Shift:
T = 2π(M/K)1/2
Balance between
SA and SD (design
of isolation gap)
Higher mode
contributions are
nearly zero for
ideal case
(DIS, 2010)
Elastomeric Bearings
n 
n 
n 
n 
Laminated rubber layers
and steel shims
Damping: 2-20%
Can achieve shear strains
above 200%
Types
n 
n 
n 
(DIS, 2010)
Low Damping Rubber Bearings
(LDRB)
Lead Plug Rubber Bearings
(LPRB)
High Damping Rubber
Bearings (HDRB)
literature, and investigations into their modeling
behavior
serve asconcept.
the basis
forbearing
several
to make and
use of
the pendulum
This
of the studies contained in this report.
a concave spherical surface. The slider
Friction Bearings
3.1
n 
n 
n 
Single-Pendulum Bearings
(polytetrafluoroethylene) composite liner, and the
stainless steel. A picture showing an FP bearing an
The single-concave friction pendulum bearing is the original Friction Pendulum
indicating the above-described components.
System described by Zayas et al. [1987] and represents the first manufactured sliding-bearing
Pendulum-like restoring
force
to make use of the pendulum concept. This bearing consists of an articulated slider resting on
a concave spherical surface. The slider is coated with a woven PTFE
Lining
materials with
(polytetrafluoroethylene) composite liner, and the spherical surface is overlain by polished
friction
coefficients from
stainless steel. A picture showing an FP bearing and a cross-section is shown in Figure 3-1,
1% indicating
to more
than 20%
the above-described components.
Period independent of
Figure 3-1: Photo (left) and section (rig
structure’s weight:
T = 2π(R/g)1/2
Types
19
Spherical concave surface
n 
• 
Single, Double and Triple
Pendulum Friction Bearings
Articulated slider
(Morgan, 2011)
Figure 3-1: Photo (left) and section (right) of a typical FP bearing
Isolation Applications
n 
n 
n 
n 
n 
n 
Structures
Bridges
Off-shore Oil
Platforms
LNG Tanks
Port Cranes
NPPs
(Earthquake Protection)
NPP Isolation - Koeberg
n 
n 
n 
n 
Design by Framatome
Built in 1976 in Koeberg, South Africa
First Seismically Isolated Nuclear Power Plant
Twin 900 MWe PWR Units
Koeberg
n 
Site Conditions
n 
n 
$23G39I+6JJ
PGA = 0.6g
Bearing Details
n 
n 
n 
n 
n 
900 Isolators per
Reactor
<>?+/A'3DK%#'I-2883"+A%#3-#
Neoprene Pads and
Sliders
5% critical damping
2 in. displacement at
point of sliding
µ = 0.18
(Labbe)
Koeberg - Construction
(Spie Batignolles)
n 
n 
n 
n 
Pre-fabricated units
Each unit weighed approximately 4 tons
20-60 units installed per day
Horizontality of unit carefully checked
throughout production and installation process
NPP Isolation - Cruas
n 
n 
n 
Design by Framatome
Built in 1978 in Cruas, France
(4) 900 MWe PWR Units
NPP Isolation - Cruas
n 
Site Conditions
n 
PGA = 0.3g
n 
Bearing Details
n 
n 
n 
900 Isolators per
Reactor
Neoprene Pads
5% Critical
Damping
Cruas
(Labbe)
Other Isolated Nuclear Facilities
M. Hashimoto et al. / Fusion Engineering and Design 41 (1998) 407–414
Fig. 1. Partial isolated building (East–West section).
(Hashimoto)
smic isolation was selected (Fig. 1). In
he tokamak building pit, in which the
machine and its ancillary systems are
will be supported on horizontally commic isolators whichn 
reduce the eartheleration (from 0.4 to 0.2 g). However,
ve motion between the unisolated and
arts of the buildingn will be relatively
m 5 to 15 cm in any horizontal direction
1 to 2 cm vertically, depending on the
stics of the isolators. Pipes, busbars,
n 
study considering the Ministry of International
Trade and Industries (MITI) Notification No. 501
and the seismic design criteria of the Japan Electric Association (JEAG 4601). Stresses at various
points are analyzed by using the computer program SAP-V. The displacement and the acceleration are combined to the operational loading of
each representative service.
For the normal busbars and the HVAC ducts,
special expansion joints are investigated from the
first. For the mechanical rail for remote handling
ITER and Jules Horowitz Reactor
Low Damping Elastomer Bearings
Under Construction in France
(CEA)
NRC Regulations
n 
Types of Isolators
n 
n 
n 
n 
High Damping Rubber Bearings?
n 
n 
n 
Low Damping Rubber Bearing
Lead Rubber Bearing
Friction Pendulum Bearing
Problems with scragging and unpredictable
changes in properties
90%< confidence in the survival of the
isolation system
Limited moat damage or potential for hard
stop
DRAFT FOR COMMENT
1
Table 8-1. Performance and design expectations for seismically isolated nuclear power plants
Isolation system
Ground motion
levels
GMRS+
2
The envelope of
the RG1.208
GMRS and the
minimum
foundation input
3
motion for each
spectral frequency
Isolation unit and
system design and
performance criteria
No long-term change in
mechanical properties.
100% confidence of the
isolation system surviving
without damage when
subjected to the mean
displacement of the
isolator system under the
GMRS+ loading.
EDB4 GMRS
The envelope of
the ground motion
amplitude with a
mean annual
frequency of
exceedance of
-5
1x10 and 167%
of the GMRS+
spectral amplitude
90% confidence of each
isolator and the isolation
system surviving without
loss of gravity-load
capacity at the mean
displacement under EDB
loading.
Approach to demonstrating
acceptable performance of
isolator unit
Production testing must be
performed on each isolator
for the mean system
displacement under the
GMRS+ loading level and
corresponding axial force.
Prototype testing must be
performed on a sufficient
number of isolators at the
5
CHS displacement and the
corresponding axial force to
demonstrate acceptable
performance with 90%
confidence. Limited isolator
unit damage is acceptable
but load-carrying capacity
must be maintained.
Superstructure
design and
performance
Umbilical line
design and
performance
The superstructure
design and
performance must
conform to NUREG0800 under GMRS+
loading.
Umbilical line
design and
performance must
conform to
NUREG-0800
under GMRS+
loading.
There should be less
than a 10%
probability of the
superstructure
contacting the moat
or hard stop under
EDB loading.
Greater than 90%
confidence that
each type of safetyrelated umbilical
line, together with
its connections,
remains functional
for the CHS
displacement.
Performance can
be demonstrated
by testing, analysis
or a combination of
6
both.
Moat or hard stop design
and performance
The moat is sized such that
there is less than 1%
probability of the
superstructure contacting the
moat or hard stop under
GMRS+ loading.
CHS displacement must be
equal to or greater than the
90th percentile isolation
system displacement under
EDB loading.
Moat or hard stop designed to
survive impact forces
associated with 95th percentile
EDB isolation system
7
displacement. Limited
damage to the moat or hard
stop is acceptable but the
moat or hard stop must
perform its intended function.
1. Analysis and design of safety-related components and systems should conform to NUREG-0800, as in a conventional nuclear structure.
2. 10CFR50 Appendix S requires the use of an appropriate free-field spectrum with a peak ground acceleration of no less than 0.10g at the foundation level.
RG1.60 spectral shape anchored at 0.10g is often used for this purpose.
3. The analysis can be performed using a single composite spectrum or separately for the GMRS and the minimum spectrum.
-5
4. The analysis can be performed using a single composite spectrum or separately for the 10 MAFE response spectrum and 167% GMRS.
5. CHS=Clearance to the Hard Stop
6. Seismic Category 2 SSCs whose failure could impact the functionality of umbilical lines should also remain functional for the CHS displacement.
7. Impact velocity calculated at the displacement equal to the CHS assuming cyclic response of the isolation system for motions associated with the 95th
percentile (or greater) EDB displacement.
How might Isolation Benefit
Current NPP designs
n 
20
A simple numerical “stick” model was found
in the open literature to represent an AP
1000 standard plant design (Westinghouse)
Simplified Numerical Stick Model
Study Input Parameters
n 
n 
From EPRI/Bechtel
study of SSI modeling
Structures Considered
•  Auxiliary/Containment
Building (ASB)
(T1 = 0.31 sec)
!
'%(
5)+
%#&
.)0
)6+
9)+,
0*,
96+,
#$%
9)+,
.+/
9)+,
4+
96+,
*54123
0*,
*5478
9)+,
*5*
9)+,
*5*78
4+78
.+)
)6+123
9)+, *54
)6+78
*
)
Structure (CIS)
(T1 = 0.08 sec)
2D idealization used
for pilot study
5)+123
.)0123
/*,
•  Containment’s Internal
n 
0*,
!
'%(
"
%#&
#$%
)*+,
.+)
)
)*+,
21
Figure 2-6
*
-
:;1<=
@18A3B1C
;27>=?
.+)
::::(AD=:1E:%3==F:#1C3ABC7=C3:&=DD=F
.+/
::::@1G:1C:%3==F:#1C3ABC7=C3:&=DD=F
.)0
::::H1I:1E:%3==F:#1C3ABC7=C3:&=DD=F
.)0123 ::::%3==F:#1C3ABC7=C3:&=DD=F:J23?BKK=?
)
::::L12C<A3B1C:1E:'2MBFBA?N:%OB=F<:(2BF<BCK
4+
::::@1G:1C:'2MBFBA?N:%OB=F<:(2BF<BCK
4+78 ::::@1G:1C:'2MBFBA?N:%OB=F<:(2BF<BCK:PADD:#=C3=?
)6+
::::H1I:1E:'2MBFBA?N:(2BF<BCK
)6+123 ::::H1I:1E:'2MBFBA?N:(2BF<BCK:J23?BKK=?
)6+78 ::::H1I:1E:'2MBFBA?N:(2BF<BCK:PADD:#=C3=?
5)+
::::H1I:1E:%OB=F<:(2BF<BCK
5)+123 ::::%OB=F<:(2BF<BCK:J23?BKK=?
*
::::(AD=:1E:#1C3ABC7=C3:$C3=?CAF:%3?2832?=
*5*
::::@1G:1C:#1C3ABC7=C3:$C3=?CAF:%3?2832?=
*5*78 ::::@1G:1C:#1C3ABC7=C3:$C3=?CAF:%3?2832?=:PADD:#=C3=?
*54
::::H1I:1E:#1C3ABC7=C3:$C3=?CAF:%3?2832?=
*54123 ::::H1I:1E:#1C3ABC7=C3:$C3=?CAF:%3?2832?=:J23?BKK=?
*5478 ::::H1I:1E:#1C3ABC7=C3:$C3=?CAF:%3?2832?=:PADD:#=C3=?
Ground Motions Used
Chapter 8
Demonstration Hazard Calculations
n 
n 
n 
Simplified generalized
modal analysis was
done in Matlab based
on response spectrum
Two sites from
Seismic Source
Characterization Study
considered.
Spectra based on
Hazard estimates for
NUREG 1.60 and PGA Manchester, New Hampshire and
Savannah, Georgia sites were
estimates
Figure 8.1-1
Map showing the study area and seven test sites for the CEUS SSC Project
8-28
22
used to generate spectra
Comparison of models
Fixed Base Model
Base Isolated Model
Elastic Superstructure
Base Mat
23
Equivalent linear
isolator properties
Tiso
ξiso
Response Spectrum Analysis
n 
n 
n 
n 
Period of Isolation = 2, 2.5, 3, 3.5 and 4 s
Damping Ratio of Isolator = 2,10,15 and
20%
Method of Analysis: Generalized Modal
Analysis
Program: MATLAB (code courtesy of Dr.
Tracy Becker)
Savannah Site with Shallow Soil
Results in Largest Response
v  Results
shown here
for 4x10-4 probability
of exceedence on soft
clay
v  Shears at levels in
isolated structure are
about 1/7 of those
for fixed base case
v  Other hazard levels
and soils exhibit
similar trends
25
Story Shears for Fixed
and Isolated Cases
Isolated
Effect of Isolator Period and Damping on
Base Shear and Isolator Displacement
Isolator Displacement Demand
Base Shear Demand
Low damping
Higher
Damping
All demands decrease with increased damping of isolators
With increasing isolator period, isolator displacement increases,
but base shear decreases (tradeoff needed)
26
(Savannah: 4x10-4 Hazard on Soft Clay)
Floor Response Spectra
n 
n 
n 
Period of Isolation = 2, 2.5, 3, 3.5 and 4 s
Damping Ratio of Isolator = 2,10,15 and
20%
3 Key Locations
n 
n 
n 
n 
n 
Control Room (ASB)
Fuel Building’s Roof (ASB)
Operating Desk (CIS)
Method of Analysis: Generalized Modal
Analysis
Program: MATLAB (code courtesy of Dr.
Tracy Becker)
Dependence of Floor Spectra On
Different Isolator Periods
n 
n 
Floor spectra
calculated at different
levels for fixed base
and isolated plant
Fixed base has high
spectral values at
high frequencies
Isolated plant has
high spectral values
near frequency of
isolation system
Floor Response Spectra at Z=116.5ft for Manchaster Shallow Soil ASB with Iso ζ=0.02 and Tiso=
0.4
0.2
0
0
10
ζ=0.02 and Tiso=2.5s
0.2
0
0
1
10
10
ζ=0.02 and Tiso=3s
0.4
0.2
0
0
10
ζ=0.02 and Tiso=3.5s
1
10
0.2
0.1
0
0
1
10
10
ζ=0.02 and Tiso=4s
0.2
0.1
0
0
1
10
10
Fixed
1
0.5
0
0
1
10
28
1
10
0.4
SA (g)
n 
10
Frequency (Hz)
Comparison of Ratio of Floor Spectral
Values for Isolated and Fixed Base NPP
In high frequency range,
u  Spectral values decrease with
increasing isolator period
u  Reduction of floor spectrum by
60 to 80% in this range possible
compared to fixed base case.
In low frequency range,
u  Significant amplifications occur
at natural frequency of isolator,
u  Amplification can be by order of
magnitude,
u  For long period isolators, the
fixed base spectral accelerations
may be quite low, and a large
amplification near the isolation
frequency may not be
important. However, this needs
29
to be confirmed.
Sa(Isolated)/Sa(Fixed Base)
Realistic Floor Spectra – Evidence
from Earthquake Records
Fukushima Daiichi Emergency Operations Facility
30
But spectra may be sensitive to
modeling of structure and isolators
P-Δ effects in bearings
31
Simple lumped
mass stick
models will not
capture vertical
response effects
!
Coupled Vertical-Lateral
Mode Shapes in Asymmetric
Structures
!
Floor spectra in high frequency
range is sensitive bearing properties
P-Δ effects in bearings
Vertical ground excitations
will trigger horizontal
vibrations in
superstructure
32
!
Strongly nonlinear systems trigger
high frequency vibrations
Ma + Cv + Keffd = -mag - Qy
Responds at effective
period of isolator
Impact like loading
triggers response at
natural frequencies of
supported structure
Time History Analysis
n 
Ground Motion Time Histories
n 
n 
n 
n 
n 
n 
30 simulated time histories
Hard Rock and Soil
43 miles from the New Madrid source
Magnitude 7.6 Earthquake
Amplification factor = 1.5 to simulate new
seismic characterization
Model
n 
n 
ASB with representative LDRB and LPRB
bearings
SAP2000
Time History Analysis – Results
(LDRB)
n 
Similar response between the amplified
and original ground motions
Time History Analysis – Results
(LPRB)
n 
n 
Linear regressions do not fit data well
outside range of peak values
Difficult to use equivalent linear models
for nonlinear bearings
Time History Analysis – Results
(LPRB)
n 
n 
Linear regressions do not fit data well
outside range of peak values
Difficult to use equivalent linear models
for nonlinear bearings
Future Work
n 
n 
n 
n 
n 
Effects of Vertical Motion
Shape of Hysteretic Loops
Asymmetric Structures (coupled H-V
response)
Soil-Structure Interaction
Experimental Work
Conclusions
n 
n 
n 
n 
Isolation has shown to effectively reduce
shears and drifts at various locations for both
models
Isolation has the ability to maintain
effectiveness for variations in ground motions
Performance Based Design can really provide
an effective means of addressing seismic
issues concerning isolation application
Future research and development offers a
great opportunity for collaboration across
various engineering fields
Thanks!
39