INTEGRATED EARTHQUAKE SIMULATIONINTEGRATED EARTHQUAKE SIMULATIONM. HORI

Earthquake Research Institute, University of Tokyo

CONTENTSFull computation of earthquake hazards and disastersExamples of integrated earthquake simulation

Department of Mechanical Engineering, Tsinghua

UniversityBeijing, RPC, July 7, 2006

Recent Achievement on Earthquake ResearchRecent Achievement on Earthquake Research

Computation of Earthquake Wave PropagationProf. Furumra (ERI, U. Tokyo): use of Earth Simulator

Dynamic structure analysis methods used for designnon-linear due to materials and large-deformation

Development of GIS Storing Structure Data

fault mechanism

propagation

amplificationearthquake

computer model of city

structure response simulation

earthquake simulation

action against earthquake

infrastructurespublic spaces

lifelineshouses/buildings

crisis managementevacuation

recoveryretrofitting

action simulation

time

OVERVIEW OF IESOVERVIEW OF IES

IES

earthquake

structure response

steel structure

concrete structure

geotechnology

buildings/architecture

fault mechanism

propagation

amplification

data:structure

results:damage

data:underground

results:SGM

GIS

action

peopleagentscommunitysociety

earthquake scenario

SGM distribution

model construction

disasterprediction

SYSTEM OF IESSYSTEM OF IESa set of (application) layers of GIS

MACRO-MICRO ANALYSIS METHODMACRO-MICRO ANALYSIS METHOD

micro-analysis

macro-analysis

bounding medium theory

multi-scale analysispessimistic optimistic

?stochastic modeling of underground structure

geological scale

engineering scale

EXAMPLE OF MACRO-MICRO ANALYSIS: YOKOHAMA CITY SIMULATION EXAMPLE OF MACRO-MICRO ANALYSIS: YOKOHAMA CITY SIMULATION

Comparison of synthesized strong ground motion with data observed at 13 seismograph sites

August 11, 1999

139.5E

139.5E

140.0E

140.0E

140.5E

140.5E

35.0N 35.0N

35.5N 35.5N

0 7.5 15

km

epicenter

as06

is06

hd06

Yokohama City

Lat. Long. Depth Strike Dip Rake Mag.35.4N 139.8E 53km 62 85 73 4.0Mw

MODEL FOR MICRO-ANALYSISMODEL FOR MICRO-ANALYSIS

30[km]

40[km]

70[km]

EN

mate.2mate.1

mate.4mate.3m ate.1 m ate.2 m ate.3 m ate.4

p wave veloc.[m /sec] 1040 1730 2950 5200s wave veloc.[m /sec] 600 1000 1700 3000

density[kg/m 3] 1800 2000 2300 2500

139.5E

139.5E

140.0E

140.0E

140.5E

140.5E

35.0N 35.0N

35.5N 35.5N

0

-500

-1000

[m]0

-2500

-5000

[m]0

-1250

-2500

[m]between 1st and 2nd layers between 2nd and 3rd layers between 3rd and 4th layers

MODEL FOR MICOR-ANALYSISMODEL FOR MICOR-ANALYSIS

model 1 (optimistic) model 2 (pessimistic)

model of site as06s

400[m]

as06N

400[m]

[m]

1020

50

30

40

401 402 403 405 414404

700

210

140

100

2100

1800

1600

1500

Ac2

Ac1

T

As2

ρ[kg/m] 3 β[m/sec]layer

properties of layer

400[m]

100

400

200

500

300

b[m/s]600

400[m]

50[m]

MACRO-ANALYSIS RESUTLSMACRO-ANALYSIS RESUTLS

a) case 1

b) case 2

lower bounding medium

upper bounding medium

measuredvelocity [kine]

0.08

0.04

0.00

20 4 6 8time [sec]

velocity [kine]

0.02

0.01

0.00

20 4 6 8time [sec]

agreement with observed data

simple fault mechanism usedaccuracy guaranteed up to 2[Hz]

MICRO-ANALYSIS RESULTS (1)MICRO-ANALYSIS RESULTS (1)

east/west up/down

frequency[Hz]

1

10

100

0.25 1.25 2.25 3.25 4.25

ampl

itude

[cm

]

north/southmeas. opt. pes.

meas. opt. pes.PGV [kine] 0.234 0.316 0.391

SI 0.295 0.348 0.415

meas. opt. pes.PGV [kine] 0.170 0.264 0.292

SI 0.191 0.318 0.339

meas. opt. pes.PGV [kine] 0.049 0.093 0.133

SI 0.053 0.073 0.101

fair agreement with observed data

MICRO-ANALYSIS RESULTS (2)MICRO-ANALYSIS RESULTS (2)

1

0.1

0.01

a) case 1 b) case 2

max. velocity[kin]1

0.1

0.01

site number

lower bounding medium

upper bounding mediummeasured

2 sites have consistent errors, probably due to poor modeling fair agreement with observed data

SIMULATION USING GIS DATASIMULATION USING GIS DATA

Structure Data Stored in GISbasic properties (location, owner)configurationstructure type

Construction of Modelspresumption of standard material and structure properties

standard structure model

Automatic Constructionautomatic construction of GIS data to standard structure modelhighest redundancy needed

EARTHQUAKE DISASTER SIMULATION DEVELOPMENT OF VIRTUAL TOWN/CITY EARTHQUAKE DISASTER SIMULATION DEVELOPMENT OF VIRTUAL TOWN/CITY

size [m] data earthquake structure action

virtual town 300x300 GIS MMAM SDOF for building

none

virtual city 1 1,000x2,000 GIS with CAD MMAM VFEM for buildings

none

virtual city 2 400x440 GIS with CAD MMAM VFEM for infrastructures

none

virtual city 3 300x300 GIS with AI MMAM FEM based on ERD

Multi-Agent Simulation

VFEM Voxel Finite Element MethodMMAM Macro-micro Analysis MethodERD Earthquake Resistant Design

VIRTUAL TOWNVIRTUAL TOWN

Roppongi Area: 300x300[m]

GIS for bore holes(surface layers)

GIS for buildings

DATA AND COMPUTER MODELDATA AND COMPUTER MODEL

number of layer soil type density(g/cm3) Sv(m/s) Pv (m/s)1 surface soil 1.625 120.0 204.02 loam 1.550 135.0 229.53 sand 1.800 400.0 680.04 clay 1.750 200.0 340.05 fine sand 1.900 425.0 722.5

bottom rock 1.850 600.0 1020.0

60[m]

300[m]

300[m]

target area (300x300[m])

boring site

C

B

D

EF

N

A

C

B

D

EF

N

A

C

B

D

EF

N

A

147 buildings

Materials Fundamental Period

Wooden buildings varies from 0.2 sec. to 0.7 sec.

R.C. T=0.02H

S.R.C. T=0.03H

SIMULATION OF VIRUTAL TOWNSIMULATION OF VIRUTAL TOWN

high spatial resolution large variability in response

borehole data

CAD data

VIRTUAL CITY 1VIRTUAL CITY 11km

2km dataCAD data for all buildingsborehole data for underground

N

high frequency dominant low frequency dominant

RESPONSE OF BUILDINGS IN VIRTUAL CITY 1RESPONSE OF BUILDINGS IN VIRTUAL CITY 1

HIGH RESOLUTION OF IESHIGH RESOLUTION OF IES

part of virtual city room

VIRTUAL CITY 2VIRTUAL CITY 2

dataCAD data for all buildingsborehole data for underground

RESPONSE OF INFRASTRUCTURES AND BUILDINGS IN VIRTUAL CITY 2 RESPONSE OF INFRASTRUCTURES AND BUILDINGS IN VIRTUAL CITY 2

RESPONSE OF VIRTUAL CITY 2RESPONSE OF VIRTUAL CITY 2

CLOSE-UP VIEW OF STRUCTURE: DYNAMIC RESPONSE OF BRIDGE 1 CLOSE-UP VIEW OF STRUCTURE: DYNAMIC RESPONSE OF BRIDGE 1

X direction

Y direction

Z direction

VIRTUAL CITY 3VIRTUAL CITY 3

soil layerconcrete pier

steel pier

ground mold gas pipe line 300[m]

300[m

]

PLUG-IN OF NUMERICAL METHODSPLUG-IN OF NUMERICAL METHODS

Methodology of IntegrationApplication of DataBase management techniques

Federation-type DataBase with aid of interpreting programs (agents)

PLUG-IN: request programmer to modify source codes such that I/O can be made through common protocols

heavy task for programmer due to variety of structure data

interpreting program: supports data exchange between kernel and each simulation program

SP-Csimulation program

SP-A

SP-D

SP-B

interpreter program A

IES

MEDIATOR AS INTERPETING AGENTMEDIATOR AS INTERPETING AGENT

major role of mediator1. data exchange2. execution of program

kernel

simulation program

mediator

data

visualization

UGSM

micro-analysis

analysis

visualize

take data

select point

analyze response

put

make VRML

take

take location

response analysisVRML

library

GIS

IES structure with mediator

structure

FEM

location, displacement data

node

nodes, type, element

propertymaterial

location, propertyGauss point

A VRML

VMRL_A

VRML

POV

AVS

successi on

for structure A

successi on

find structure

A

take data

select point

analyze response

put

take

make VRML

take data from library

find structure from GIS

select point to which strong motion is input

run simulation program

take results of simulation program

put data to simulation program

visualize results of simulation program

object of mediator for structure A

INHERETANCE OF MEDIAOTR OBJECTINHERETANCE OF MEDIAOTR OBJECT

MethodsProgram Structure of Mediator

Computational Science Approach:seek to automatic coding of mediator program by taking full advantage of inheritance

MEDIATOR MAKER AS AIMEDIATOR MAKER AS AIMediator Maker

artificial intelligence of automatically generating a mediator for a given simulation programmediator maker analyzes source code of a simulation program, and constructs a code of a corresponding mediator

It is possible to make mediator maker by taking advantage of

common structure of source code (FEM-based)usage of common key variables (such as number of nodes, elements, etc.)

source codes

digested code

code for mediator

analysis of source codesubroutinecommon memory

construction of mediatoranalysis of I/Oanalysis of loop andconditionprediction of variable

0.5

3 2 31

8

1

5

5

10 15

11.7x1.7

5x2.5

10

1x1

10 8 14

20

MODELS OF INFRASTRUCUTRES IN VIRTUAL CITY 3 MODELS OF INFRASTRUCUTRES IN VIRTUAL CITY 3

infrastructuresgas pipe lineconcrete piersteel pierground mold

particular material properties & mechanism for earthquake resistance

reliable analysis methods used in ERD are plugged in IES

NON-LINEAR RESPONSE OF INFRASTRUCTURES IN VIRTUAL CITY 3 NON-LINEAR RESPONSE OF INFRASTRUCTURES IN VIRTUAL CITY 3

steel pier ground mold

concrete pier

plugged-in simulation programs work in IES

RESPONSE OF VIRTUAL CITY 3RESPONSE OF VIRTUAL CITY 3

unified visualization will contribute to form common recognition of possible earthquake disaster

all numerical analysis of structures arenon-linearbased on earthquake resistant design

SNAPSHOT OF RESPONSE AND DAMAGESNAPSHOT OF RESPONSE AND DAMAGE

SIMULATION OF ACTIONSIMULATION OF ACTION

Evacuation Simulationphysical simulation model based on flow, potential, etc.multi-agent simulation intelligent agents

Damage and Loss Simulationmicro: evaluation of damage and repair costmacro: based on past record of damagerisk analysis/management

need to estimate panic in evacuation process

prediction of damage and loss of future society

ACTION SIMULATION: EVACUATION SIMULATION USING MAS ACTION SIMULATION: EVACUATION SIMULATION USING MAS

Evacuation Simulationpredict evacuation process in underground towns, high-rise buildings, department stores, schools, etc.use intelligent agents to mimic people in panic state

underground subway station4 floors3 elevators

MULTI-AGENT SIMULATIONMULTI-AGENT SIMULATION

Utilization of Intelligent Agentinput internal variables (max. speed, intelligence, memory)

external stateoutput see surrounding environment, agent, etc.

think judge most suitable pathmove go to next position

AgentAbilityThoughtSee(): …Wall, …Corridor, …Agent Think(): …Destination, …Direction,

…SpeedMove(): …Direction, …Check

AbilityMaximumSpeedVisibilityIntelligence

DirectionSpeedPath

Thought

MODELING FOR DYNAMIC STRUCTUREIS & EVACUATION ANALYSIS MODELING FOR DYNAMIC STRUCTUREIS & EVACUATION ANALYSIS

modeling strategystructure data for each floordata for escalator

evacuation pathcorridors determined from structure dataExit determined from escalator

data structure is arranged to be used for structure analysis and for evacuation analysis

RESULTS OF SIMULATION (1)RESULTS OF SIMULATION (1)dynamic structure analysis

modelstructure & groundsimulationMMAM

structure response analysis is made by accounting for soil-structure interaction effects

RESULTS OF SIMULATION (2)RESULTS OF SIMULATION (2)green less intelligent agent

- slow thinkgray more intelligent agent

- quick think

multi-agent analysis is possible for structures to which dynamic structure analysis is made.

lower floor

escalator& stairs

upper floor

no damage

damage

DIGITAL KOBEDIGITAL KOBE

700 [m

]

500 [m]

buildings 1262wooden, RC/SRC

bridge for highway14 RC piers

Kobe City, Higashi-Nada District500[m]x700[m]

SIMULATION RESULTS (1)SIMULATION RESULTS (1)

SIMULATION RESULTS (2): THREE EARTHAUEK SCENARIOS SIMULATION RESULTS (2): THREE EARTHAUEK SCENARIOS

Case 1

Case 2

Case 3

stress concentration

Case 1 Case 2

Case 3

differences in earthquake disaster due to difference in earthquake scenario

FULL CITY SIMULATIONFULL CITY SIMULATION

DIGITAL BUNKYO CITYDIGITAL BUNKYO CITYground structure model 700[m]ｘ700[m]ｘ40[m]3 layers

building modelsaround 2000 MDOF system

data sourceGIS of boring data of Tokyo Metro-policecommercially available GIS for structure data

STRONG GROUND MOTION DISTRIBUIONSTRONG GROUND MOTION DISTRIBUION

Expected Kanto Earthquake Recorded Hanshin-Awaji Earthquake

STRONG GROUND MOTION DISTRIBUTIONSTRONG GROUND MOTION DISTRIBUTION

Expected Kanto Earthquake Recorded Hanshin-Awaji Earthquake

VIBRATION OF BUILDINGSVIBRATION OF BUILDINGS

VIBRATION OF BUILDINGSVIBRATION OF BUILDINGS

-8000 -7500 -7000-33000

-32800

-32600

-32400

-32200

-32000

-31800

-31600

EVACUATION OF RESIDENCESEVACUATION OF RESIDENCES

road network model

multi-agent simulationsimulation for various situations is possible.

SIMULATION OF TOKYO STATIONSIMULATION OF TOKYO STATION

strong ground motion

EXAMPLE 1EXAMPLE 1

structure-wise damage evaluation

EXAMPLE 2EXAMPLE 2

TOKYO DOWNTOWNTOKYO DOWNTOWN

scenario 1

EXAMPLE 1EXAMPLE 1

strong ground motion

シナリオ1

EXAMPLE 2EXAMPLE 2

structure response

CONCLUDING REMARKSCONCLUDING REMARKS

Potentials of IESGeneration of new information on natural disaster risk and hazard by applying numerical simulation to data stored in GISImprove residents awareness on natural disaster risk and hazards and form common recognition of them among community

Future TasksAdvancement of structure data stored in GISnew exclusive sensing of buildings and structuresAdvancement of natural disaster preparedness, mitigation and recoverysignificant need for integration of technologies for natural disasters