uncertainties in modelling and simulation of …sokocalo.engr.ucdavis.edu/~jeremic/ modeling and...
TRANSCRIPT
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Uncertainties in Modelling and Simulation ofEarthquake Soils Structures Interaction
Boris Jeremic
University of California, DavisLawrence Berkeley National Laboratory, Berkeley
Earthquake Research InstituteUniversity of Tokyo
Tokyo, Japan, January 2016
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Outline
MotivationIntroductionUncertainties
Modeling and Parametric UncertaintyModeling Uncertainty: Seismic Motions for NPPsParametric Uncertainty: Uncertain Material
Real ESSI Simulator SystemReal ESSI Simulator
Summary
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Introduction
Outline
MotivationIntroductionUncertainties
Modeling and Parametric UncertaintyModeling Uncertainty: Seismic Motions for NPPsParametric Uncertainty: Uncertain Material
Real ESSI Simulator SystemReal ESSI Simulator
Summary
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Introduction
Motivation
I Improve seismic design of soil structure systems
I Earthquake Soil Structure Interaction (ESSI) in time andspace, plays a major role in successes and failures
I Accurate following and directing (!) the flow of seismicenergy in ESSI system to optimize for
I Safety andI Economy
I Development of high fidelity numerical modeling andsimulation tools to analyze realistic ESSI behavior:Real ESSI simulator
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Introduction
Predictive CapabilitiesI Verification provides evidence that the model is solved
correctly. Mathematics issue.
I Validation provides evidence that the correct model issolved. Physics issue.
I Prediction under Uncertainty (!): use of computationalmodel to predict the state of SSI system under conditionsfor which the computational model has not been validated.
I Modeling and Parametric Uncertainties
I Predictive capabilities with low Kolmogorov Complexity
I Modeling and simulation goal:predict and inform, rather than (force) fit
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Uncertainties
Outline
MotivationIntroductionUncertainties
Modeling and Parametric UncertaintyModeling Uncertainty: Seismic Motions for NPPsParametric Uncertainty: Uncertain Material
Real ESSI Simulator SystemReal ESSI Simulator
Summary
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Uncertainties
Modeling Uncertainty: Simplified Models
I Simplified modeling: Features (important ?) are neglected(6D ground motions, inelasticity)
I Modeling Uncertainty: unrealistic and unnecessarymodeling simplifications
I Modeling simplifications are justifiable if one or two levelhigher sophistication model shows that features beingsimplified out are not important
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Uncertainties
Parametric Uncertainty: Material and Loads
Example: Elastic Stiffness
5 10 15 20 25 30 35
5000
10000
15000
20000
25000
30000
SPT N Value
You
ng’s
Mod
ulus
, E (
kPa)
E = (101.125*19.3) N 0.63
−10000 0 10000
0.00002
0.00004
0.00006
0.00008
Residual (w.r.t Mean) Young’s Modulus (kPa)N
orm
aliz
ed F
requ
ency
Transformation of SPT N-value: 1-D Young’s modulus, E (cf. Phoon and Kulhawy (1999B))
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Outline
MotivationIntroductionUncertainties
Modeling and Parametric UncertaintyModeling Uncertainty: Seismic Motions for NPPsParametric Uncertainty: Uncertain Material
Real ESSI Simulator SystemReal ESSI Simulator
Summary
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Modeling and Simulation of Nuclear Power Plants
I Nuclear Power Plants (NPPs) design based on a numberof simplified assumptions!
I Linear elastic material behavior
I Seismic Motions: 1D or 3 × 1D, or real 3D (6D)
I Savings in construction cost possible with more accuratemodeling of NPPs
I Improvements in safety of NPPs also possible, even withhigher seismic motions, as inelastic effects "eat up"(dissipate) seismic energy
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Nuclear Power Plants: 6D or 1D Seismic MotionsI Assume that a full 6D (3D) motions at the surface are only
recorded in one horizontal directionI From such recorded motions one can develop a vertically
propagating shear wave in 1DI Apply such vertically propagating shear wave to the same
soil-structure system
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Synthetic Test MotionsI Develop free field models with sources withinI Sources are simple, point (mostly), line and surfaceI Sources will send both P and S wavesI Variation in strike and dipI Simulation programs, Real ESSI Simulator and SW4
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Synthetic Test Motions, 6D vs 1DI Danger of picking one component of motions for 1D or
3×1D (it is done all the time!)I Excellent (forced) fit, but not a prediction and information is
lost (goal is to predict and inform and not (force) fit)
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Stress Test Source SignalsI Gauss
I Ricker (1st, 2nd)
-0.005
-0.004
-0.003
-0.002
-0.001
0
0.001
0.002
0.003
0 2 4 6 8 10 12 14 16 18 20Ric
ker2
nd F
unct
ion
Am
plitu
de
Time [s]
0
2e-05
4e-05
6e-05
8e-05
0.0001
0.00012
0.00014
0.00016
0 2 4 6 8 10
Ric
ker2
nd F
FT
Am
plitu
de
Frequency [Hz]
I Ormsby
-0.04
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0 1 2 3 4 5 6
Orm
sby
Fun
ctio
n A
mpl
itude
Time [s]
0
5e-05
0.0001
0.00015
0.0002
0.00025
0.0003
0.00035
0.0004
0.00045
0.0005
0 5 10 15 20 25 30
Orm
sby
FF
T A
mpl
itude
Frequency [Hz]
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Layered and Dyke/Sill Models
I Uniform soil/rock, to show surface wavesI Horizontally layered geology (a), to show
bending/refraction and more surface wavesI Dyke/Sill intrusion within layered geology (b), to show
effects of local geology on free field motions
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Variable Sources
I Epicenter is 2500m away from the location of interestI Source depth 850m (softer layers) and 2500m (hard rock)I Variation: source locations, strike and dip, magnitude,
frequenciesI Different wave propagation path to the object location
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Layered System , Displacement Traces
I Surface wavesI Surface waves not filtered outI Incoherent motions
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Layered System, Variable Source Depth
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Dyke/Sill Intrusion, Variable Source Depth
I Incoherent motionsI Surface wavesI Lower amplitudes than with
layered only model!
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Dyke/Sill Intrusion, Variable Source Depth
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Dyke/Sill as Seismic Energy SinkI Dyke/Sill (right Fig), made of stiff rock, is an energy sink,
as well as energy reflectorI Variable wave lengths behave differently, depending on
dyke/sill geometry and location
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
Importance of Realistic Seismic Motion Fields
I Developed synthetic (!) free field motions need to excite anumber of (all!) possible responses from a nuclear facility
I Knowledge of detailed geology is needed, geometry andmaterial properties, including inelasticity of shallow layers
I Reduction of modeling uncertainty
I Direct use for Realistic ESSI simulations
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Modeling Uncertainty: Seismic Motions for NPPs
6D vs 1D NPP ESSI Response Comparison
(MP4)
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Outline
MotivationIntroductionUncertainties
Modeling and Parametric UncertaintyModeling Uncertainty: Seismic Motions for NPPsParametric Uncertainty: Uncertain Material
Real ESSI Simulator SystemReal ESSI Simulator
Summary
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Uncertain Material Parameters and Loads
I Decide on modeling complexity
I Determine model/material parameters
I Model/material parameters are uncertain!I MeasurementsI TransformationI Spatial variability
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Uncertainty Propagation through Inelastic System
I Dynamic Finite Elements
Mu + Cu + Kepu = F
I Incremental elastic-plastic constitutive equation
∆σij = EEPijkl ∆εkl =
[Eel
ijkl −Eel
ijmnmmnnpqEelpqkl
nrsEelrstumtu − ξ∗h∗
]∆εkl
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Probabilistic Elasto-Plasticity: von Mises Surface
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Gradient Theory of Probabilistic Elasto-Plasticity:Verification, Elastic-Perfectly Plastic
0.03 0.02 0.01 0.00 0.01 0.02 0.03εx
1.0
0.5
0.0
0.5
1.0
σx [M
pa]
RBFMCS
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Wave Propagation Through Uncertain Soil
0 100 200 300 400 500Shear modulus G [MPa]
0.000
0.002
0.004
0.006
0.008
0.010
0.012
0.014
min COVmax COV
30 20 10 0 10 20 30 40 50 60Shear strength τ [kPa]
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
min COVmax COV
0.0 0.1 0.2 0.3 0.4 0.5Time [s]
0.03
0.02
0.01
0.00
0.01
0.02
0.03
Disp
lace
men
t [m
]
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Uncertain Elastic Response at the Surface (COV = 120%)
0.0 0.1 0.2 0.3 0.4 0.5Time [s]
0.08
0.06
0.04
0.02
0.00
0.02
0.04
0.06
0.08
0.10
Disp
lace
men
t [m
]
MM - SD
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Displacement PDFs at the Surface (COV = 120%)
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Displacement CDFs (Fragilities) at the Surface (COV = 120%)
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Parametric Uncertainty: Uncertain Material
Probability of Exceedance, disp = 0.1m (COV = 120%)
0.0 0.1 0.2 0.3 0.4 0.5Time [s]
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Prob
. of e
xcee
danc
e (d
= 0
.1m
)
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Real ESSI Simulator
Outline
MotivationIntroductionUncertainties
Modeling and Parametric UncertaintyModeling Uncertainty: Seismic Motions for NPPsParametric Uncertainty: Uncertain Material
Real ESSI Simulator SystemReal ESSI Simulator
Summary
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Real ESSI Simulator
Real ESSI Simulator SystemI The Real ESSI-Program is a 3D, nonlinear, time domain,
parallel finite element program specifically developed forHi-Fi modeling and simulation of Earthquake Soil/RockStructure Interaction problems for NPPs (infrastructureobjects) on ESSI-Computers.
I The Real ESSI-Computer is a distributed memory parallelcomputer, a cluster of clusters with multiple performanceprocessors and multiple performance networks.
I The Real ESSI-Notes represent a hypertextdocumentation system (Theory and Formulation, Softwareand Hardware, Verification and Validation, and CaseStudies and Practical Examples) detailing modeling andsimulation of ESSI problems.
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Real ESSI Simulator
Real ESSI Simulator SystemI Developed with funding from US-DOE, US-NRC, US-NSF
and CNSC-CCSN
I The Real ESSI simulator system is designed based onpremise of high fidelity, high performance, modeling andsimulation
I Reduction of modeling uncertainty and propagation ofparametric uncertainty
I Real ESSI simulator, also known as Vrlo Prosto, ,, Muy Fácil, Molto Facile, , Πραγματικά
Εύκολο, , , Très Facile, VistinskiLesno, Wirklich Einfach.
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Real ESSI Simulator
Real ESSI Modelling
input
loading stage
increment
iteration
output
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Real ESSI Simulator
Real ESSI: Components
I Variety of solid, structural and special finite elementsI Variety of material modelsI Variety of static and dynamic solution advancement
algorithmsI Seismic (realistic) input using DRMI High performance parallel computing for elastic-plastic
FEM and multiple performance CPUs and networks usingPlastic Domain Decomposition method
I Stochastic Elastic-Plastic Finite Element Method(SEPFEM)
I Extensive Verification, working on Validation
Jeremic
ESSI under Uncertainty
Motivation Modeling and Parametric Uncertainty Real ESSI Simulator System Summary
Summary
Concluding Remarks
I Modeling and simulation of infrastructure object requireshigh sophistication
I Uncertainties (modeling and parametric) influences results
I Those uncertainties need to be addressed and propagatedto results and used in decision making
I Goal is to predict and inform, and not force fit
Jeremic
ESSI under Uncertainty