Problems in Solid Mechanics

A Symposium in Honor of H.D. Bui

Symi, Greece, July 23-28, 2006

INVERSE
INVERSEANALYSIS
ANALYSISMETHODS
METHODSOF
OF
IDENTIFYING
IDENTIFYINGCRUSTAL
CRUSTALCHARACTERISTICS
CHARACTERISTICS
USING
USINGGPS
GPSARRYA
ARRYADATA
DATA

M. HORI (Earthquake Research Institute, University of Tokyo)

Contents
1. Stress inversion method: find equilibrating stress using
measured strain and partial information on stress-strain relation
2. Elasticity inversion method: find local elasticity using densely
measured displacement
 GPS
GPS NETWORK
NETWORK AND
AND ITS
ITS DATA
DATA

40Ê 40Ê

GPS stations displacement rate

5cm/yr
30Ê 30Ê
130Ê 140Ê 130Ê 140Ê

2
NEED
NEED FOR
FOR LOCAL
LOCAL STRESS
STRESS PREDICTION
PREDICTION
‹ Material Test of Next ‹ Earthquake Prediction
Generation

GPS data

strain

σ σ

ε ε
local ε−σ relation regional stress
3
IS
IS STRESS
STRESS INVERSION
INVERSION POSSIBLE?
POSSIBLE?
‹3D State Most Difficult

‹2D State Possible?

3 UNKNOWNS 2 EQUATIONS

σ11 equilibrium in x1-direction

σ22 equilibrium in x2-direction

σ12
1 CONDITION
STRAIN

4
 STRESS
STRESS INVERSION
INVERSION
Airy partial information: σ11 + σ 22 = f ( ε ij )
⎧σ11 = a , 22 f ( ε ij ) = κ ( ε11 + ε 22 )
⎪
⎨σ 22 = a ,11 ⇓
⎪σ = −a
⎩ 12 ,12 σ11 + σ 22 = κ( ε11 + ε 22 )

2D bulk modulus?
self-equilibrating stress

B.V.P.
Poisson equation measured strain

G.E. a ,11 + a , 22 = κ(ε11 + ε 22 )

B.C. n1a ,1 + n 2 a , 2 = −n1r2 + n 2 r1
resultant forces

5
EXTENSION
EXTENSION TO
TO OTHER
OTHER DEFORMATION
DEFORMATION STATE
STATE
‹Dynamic State
⎧σ11,1 + σ12,2 = ρ&u&1
⎪
⎨σ12,1 + σ 22,2 = ρ&u& 2
⎪σ + σ = f (ε)
⎩ 11 22

‹Finite Deformation State:

⎧∑ (∂X k ∂x j )σ1j, k = 0
⎪ j, k
⎪
⎨∑ (∂X k ∂x j )σ 2j, k = 0 (σ ij = σ ij (X), x i = x i (X ) )
⎪ j, k
⎪⎩σ11 + σ 22 = f (ε)

6
NUMERICAL
NUMERICAL SIMULATION
SIMULATION
‹Conditions
– elasto-plastic material with unknown yield function
– prediction of stress and stress-strain relation

distributed force
displacement field is
used as input data

1.0
FEM computation with
20x48 elements
2.0
7
RESULTS
RESULTS OF
OF INVERSION
INVERSION

σ12 σ12
distribution of σ12

sample surface
exact predicted

σ1 −σ3 σ1 −σ3
principle stress-
plastic strain rate

σ2 −σ3 σ2 −σ3

good agreement
yield locus
exact predicted
8
 MODEL
MODEL EXPERIMENT
EXPERIMENT

7cm
Riedel shears
8cm
Torsional Shearing
examples of image

video recorder
CCD camera

servomotor

max. shear strain

Experiment Apparatus 9
 OVERALL
OVERALL STRESS-STRAIN
STRESS-STRAIN RELATION
RELATION
stress (kgf/m3)

stress (kgf/m3)
time (sec)
rotation (radian)

strain

will be different from local relations?

time (sec)

10
LOCAL
LOCAL STRESS-STRAIN
STRESS-STRAIN RELATIONS
RELATIONS

σ11 [kPa] σ12 [kPa]

40
20 D
A D
20
10 A

0
0
-10
0 1 2 ε11 -0.5 0 ε12
normal stress-strain shear stress-strain

A: far from crack

no common relations?
D: near crack
11
RESULTS
RESULTS OF
OF INVERSION
INVERSION

max. shear stress (kPa)

30 A D
B E
C
20
yielding

10

0 0.1 0.2 0.3 0.4

max. shear strain

common elasto-plastic relations?

12
APPLICATION
APPLICATIONTO
TOGPS
GPSARRYA
ARRYADATA
DATA
‹verification of numerical analysis method
– check numerical stability of solving boundary value problem
– check dependency of parameters

‹application of stress inversion method

– geophysical interpretation of analysis results
– critical examination of assumption of plane state

‹development of crust deformation monitor

– automatic processing of GPS array data

13
CONVERGENCE
CONVERGENCE

σ
(hydrostatic stress)

Δ: resolution of strain distribution (degree)

14
EFFECT
EFFECT OF
OF REFERENCE
REFERENCE

τ
(max. shear stress)

15
COMPARISON
COMPARISON OF
OF STRESS
STRESS WITH
WITH STRAIN
STRAIN

σ (hydrostatic stress) τ (max. shear stress)

ε γ
(vol. strain) (max. shear strain)

16
REGIONAL
REGIONAL CONSTITUTIVE
CONSTITUTIVE RELATIONS
RELATIONS

τ/γ φ−θ
regional stiffness regional anisotropy
(τ, γ: max. shear stress and strain) (φ, γ: principle stress and strain)

regional heterogeneity and anisotropy

17
CHANGE
CHANGE IN
IN INVARIANT
INVARIANT

1st invariant 2nd invariant

18
CHANGE
CHANGE IN
IN REGIONAL
REGIONAL STATE
STATE
50 50

1st invariant ε
σ 2nd invariant
25
25

ε 0 γ
σ 1 101 201 301
τ 0
-25 1 101 201 301

A γ
τ
-50 -25

C 50
ε
σ
50

25
25

B 0
1 101 201 301
Y

0
-25 1 101 201 301

-50
B -25
γ
τ

A 50 50
ε
σ
25
25

0
1 101 201 301
0
-25 1 101 201 301

-50
C -25
γ
τ

19
REGIONAL
REGIONAL STRESS
STRESS AND
AND STRAIN
STRAIN

1st invariant 50
σ−ε
2nd invariant τ−γ
75

25 50

ε−γ 0
γ−τ 25
-50 -25 0 25 50

-25 0

A
-25 0 25 50 75

-50 -25

C 50
τ−γ
75
σ−ε

25 50

B 0 25
-50 -25 0 25 50

B
-25
-25 0 25 50 75

-50 -25

A
50
75
σ−ε t-g

25
50

0
25
-50 -25 0 25 50

-25
0

-50 C -25

-25
0 25 50 75

20
GPS
GPS DATA
DATA DURING
DURING 1998-1999
1998-1999
‹GPS Data
– no spatial filtering to get rid of measurement noise
– linear interpolation between two GPS station

‹More Sophisticated Treatment of BVP

– FEM with triangle element
– weak form
– regionally averaged field quantities

21
APPLICATION
APPLICATION TO
TO GPS
GPS NETWORK
NETWORK DATA
DATA
‹BVP in Rate Form and Weak Form
strain rate
G.E. a&,11 +a&,22 = κ(ε&11 +ε&22)
B.C. n1a&,1 +n2a&,2 = −n1r&2 +n2r&1

∫ ϕ a&
,1 ,1 + ϕ, 2 a& , 2 − κ(ϕ,1u& 1 + ϕ, 2 u& 2 )dS = 0

displacement rate
‹Computation of Average Quantities
1
ε& 11 = ∫ n1u& 1dL average stress rate computed by
Ω ∂Ω
using 1st-order derivative
1
σ& 11 = ∫ n 2 a& , 2 dL
Ω ∂Ω
22
 GPS
GPS NETWORK
NETWORK AND
AND ITS
ITS DATA
DATA

40Ê 40Ê

GPS stations displacement rate

5cm/yr
30Ê 30Ê
130Ê 140Ê 130Ê 140Ê

23
 REGIONAL
REGIONAL STRAIN
STRAIN RATE
RATE

6e-07 6e-07

40Ê 40Ê
4e-07

2e-07 4e-07

-2e-07 2e-07

-4e-07

volumetric -6e-07
Strain
max. shear 0
Strain

30Ê 30Ê
130Ê 140Ê 130Ê 140Ê

24
 REGIONAL
REGIONAL STRESS
STRESS RATE
RATE

6e-07 6e-07

40Ê 40Ê
4e-07

2e-07 4e-07

-2e-07 2e-07

-4e-07

volumetric -6e-07
x κ [Pa]
max. shear 0
x κ [Pa]

30Ê 30Ê
130Ê 140Ê 130Ê 140Ê

25
 COMPARISON
COMPARISON WITH
WITH SEISMIC
SEISMIC EVENTS?
EVENTS?

6e-07

40Ê 40Ê
4e-07

2e-07

-2e-07

-4e-07

volumetric -6e-07
x κ [Pa]
>20[km], >M3
30Ê
30Ê
130Ê 140Ê 130Ê 140Ê

26
REGIONAL
REGIONAL CONSTITUTIVE
CONSTITUTIVE RELATIONS
RELATIONS

find κ(x) s.t.

τ&(x; κ) = κ(x)γ& (x)
1.20

40
τ& , γ& maximum shear 1.15

κ regional stiffness 1.10

1.05
κ is originally used to relate σ & ε
through σ = κ ε. 1.00

0.95

not too far from known 0.90

geological structure
30
130 140

27
DRAWBACKS
DRAWBACKS OF
OF STRESS
STRESS INVERSION
INVERSION
‹Need to Know One Constitutive Relation
– bulk stress and bulk strain
– isotropy assumption

‹Need to Know Boundary Traction/Resultant Force

– assumption of uniform stress
– fast decrease of non-uniform boundary traction

‹Difficulty in Understanding Plane-Stress-State Model

another analysis method needed?

28
DRAWBACKS
DRAWBACKS OF
OF ELASTICITY
ELASTICITY INVERSION
INVERSION
‹Sensitive to Displacement Error
– need to make fine discretization of target body
– need to have some strong modes of deformation

Why is it so?
‹no mistakes in mathematics
‹poor understanding of physics

29
PHYSICAL
PHYSICALPROCESS
PROCESSAND
ANDMEASUREMENT
MEASUREMENT

GIS data
‹ May not be good to pose an inverse problem from data
data
to characteristics, because a path from characteristics

in v
to data has physical process and measurement

ers
ea
na
l ys
measurement

is
X

unknown mechanical property

characteristics

response source
physical process
crust deformation unknown plate slip/
underground deformation

30
ESTIMATION
ESTIMATION AND
AND INVERSION
INVERSION

‹ estimation: from data to response, i.e.,

data estimate function for displacement from data which are
measured discretely
‹ inversion: from response to characteristics, i.e.,
find most suitable characteristics for physical process
even though source is not known
estimation

measurement

e rsion characteristics
inv

response source
physical process

31
BLOCK
BLOCK IN
IN CONTINUUM
CONTINUUM
nodes at which displacement is measured

inner node
body B x1
boundary node
x2

block S
principal of elasticity inversion
k1 k2
displacement at all nodes given
Δ1 k 2
= no need to consider interaction
Δ1 Δ2 Δ 2 k1
of S with outside region

32
IDENTIFICATION
IDENTIFICATIONOF
OFDISPLACEMENT
DISPLACEMENTMODE
MODE

node

concentrated force

block 1 hexagonal block

block 2 block 3

material sample

material sample test

apply several BC’s, and measure displacement at nodes of a hexagonal block.

1. identify displacement modes (a characteristic set of nodal displacement)
2. identify local elasticity
33
 IDENTIFICATION
IDENTIFICATIONOF
OFDISPLACEMENT
DISPLACEMENTMODE
MODE

displacement mode elastic parameters

1 c1111
2 c1122 1
1 3 c1112
0.05
4 c2222
12 2 5 c2212
constraint 6 c1212
block 2

11 0.00 3
6 2
1 2
g g

10 -0.05 4
-0. 15
block 3

9 5

5 block 1 0.00 3

8 6 unconstraint
7

0.15
4

34
ELASTICITY
ELASTICITYINVERSION
INVERSIONMETHOD
METHOD
body B
local field variables allows
Taylor series expansion

block Ω
nodes at which displacement is measured

BASIC PROCEDURES OF INVERSION

1. use displacement data to determine Tayler series

expansion coefficients of displacement

2. use equilibrium equation to estimate elastic

parameters by expanding stress in Talker series uniform Poisson's ratio is assumed

35
DETERMINATION
DETERMINATIONOF
OFDISPLACEMENT
DISPLACEMENTCOEFFICIENT
COEFFICIENT
1. Taylor Expansion: {a ip }
P
u i (x) = ∑ a ip f p (x) {a ip } = {u i , u i ,1 , u i , 2 , 12 u i , 2 , u i ,11 , 12 u i ,12 , u i , 22 ,L}
p =1
{f p } = {1, x 1 , x 2 , x 12 , x 1 x 2 , x 22 ,L}

n
2. Displacement Data:{u i }
P
u = ∑ f pn a ip
n
i
p =1

3. Solution of Matrix Equation f pn = f p (x n )

1 ⎛ n α⎞ α
{λ , ϕ }
A P−A
a ip = ∑ α ⎜ ∑ u i ϕ n ⎟ψ p + ∑ b βi ψ βp α α
, ψ αp : SVD of fpn
α =1 λ ⎝ n
n
⎠ β =1

fully determined undetermined

36
ESTIMATION
ESTIMATIONOF
OFPOISSON RATIOνν
POISSONRATIO

1. Elasticity Tensor Expressed in Terms of Poisson Ratio

c ijkl = c ijkl
0
+ νc1ijkl

2. Equation of Equilibrium and Its Taylor Expansion

b i (x) = (c ijkl u k ,l (x)) , j = ∑ b ip f p (x) = 0
p

3. Coefficient of Expansion: bip=0 for 0th Order (p=1)

⎡ a 14 ⎤
⎢ ⎥
⎡ b11 ⎤ ⎛ ⎡2 0 0 12 1 0⎤ ⎡0 0 0 12 − 1 0⎤ ⎞ ⎢a 24 ⎥
⎢b ⎥ = ⎜⎜ ⎢0 1 1 0 0 2⎥ + ν ⎢0 − 1 1 0 0 0⎥ ⎟⎟ ⎢ M ⎥
⎣ 21 ⎦ ⎝ ⎣ ⎦ ⎣ ⎦⎠
⎢ ⎥
2 2

⎣a 26 ⎦
linear equation of ν is derived
37
 NUMERICAL
NUMERICALSIMULATION
SIMULATION(1)
(1)
0.010 0.010
error error

2nd order 2nd order

3rd order 3rd order
0.005 0.005

0.000 0.000
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5

-0.005 -0.005

-0.010 -0.010
ν ν=0.25 ν ν=0.35
ν=0.25 frequency
ν=0.35
80

60 measurement 200
40 expansion of displacement 3rd order
20
expansion of equilibrium 0th or 1st order
0
0.40
0.25 0.30 0.35
0.20 ν
0.10 0.15
38
NUMERICAL
NUMERICALSIMULATION
SIMULATION(2)
(2)
0.012

1. Measurement Error: {e } n
i 0th
0.008
P
u in = ∑ f pn a ip + e in |e'| ν
exact
=0.35
p =1
0.004

ν
exact
=0.25

2. Find ν such that 0.000

0.00 0.10 0.20 0.30 0.40 0.50
ν

3.586920

minimize | e |2 = ∑ (e in ) 2 1st
subjected to b ip (ν) = 0 3.586915

|e|

exact
ν =0.25
3.586910
measurement 200
expansion of displacement 3rd order 3.586905
0.00 0.10 0.20 0.30 0.40 0.50

expansion of equilibrium 0th or 1st order ν

39
APPLICATION
APPLICATIONOF
OFLOCAL
LOCALGPS
GPSDATA
DATA
pre-WTE
Velocity field
Pre-Tottori Earthquake
35.5˚

35˚

34.5˚

34˚
3cm/year

132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

GPS array
post-WTE
Velocity field
2000 Western Tottori Earthquake (MJMA=7.3) 35.5˚
Post-Tottori Earthquake

examine change in deformation and elasticity

before and after this earthquake 35˚

‹ 53 GPS observation points

‹ 82 triangular elements 34.5˚

‹ GPS data obtained from 1997 to 2002

‹ annual and biannual sinusoidal variations 34˚
3cm/year

excluded 132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

40
STRAIN
STRAINRATE
RATE
pre-WTE post-WTE
Dilatational strain rate Dilatational strain rate
97.05.24 -- 00.08.06 00.11.04 -- 02.07.27
35.5˚ 35.5˚
6e-07 6e-07

4e-07 4e-07

35˚ 2e-07 35˚ 2e-07

dilatational 0 0

34.5˚ -2e-07 34.5˚ -2e-07

-4e-07 -4e-07

-6e-07 -6e-07
34˚ [Strain/Year] 34˚ [Strain/Year]

132.5˚ 133˚ 133.5˚ 134˚ 134.5˚ 132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

Maximum shear strain rate Maximum shear strain rate

97.05.24 -- 00.08.06 00.11.04 -- 02.07.27
35.5˚ 35.5˚
1e-06 1e-06

8e-07 8e-07

35˚ 35˚
6e-07 6e-07

shear
4e-07 4e-07
34.5˚ 34.5˚

2e-07 2e-07

0 0
34˚ [Strain/Year] 34˚ [Strain/Year]

132.5˚ 133˚ 133.5˚ 134˚ 134.5˚ 132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

41
STRESS
STRESSRATE
RATE
pre-WTE post-WTE
Dilatational stress increment Dilatational stress increment
97.05.24 -- 00.08.06 00.11.04 -- 02.07.27
35.5˚ 35.5˚
120 120

100 100

80 80

60 60
35˚ 40 35˚ 40

dilatational
20 20

0 0

-20 -20

34.5˚ -40 34.5˚ -40

-60 -60

-80 -80

-100 -100

-120 -120
34˚ [KPa/Year] 34˚ [KPa/Year]

132.5˚ 133˚ 133.5˚ 134˚ 134.5˚ 132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

Maximum shear stress increment Maximum shear stress increment

97.05.24 -- 00.08.06 00.11.04 -- 02.07.27
35.5˚ 35.5˚
120 120

100 100

35˚ 80 35˚ 80

shear 60 60

34.5˚ 40 34.5˚ 40

20 20

0 0
34˚ [KPa/Year] 34˚ [KPa/Year]

132.5˚ 133˚ 133.5˚ 134˚ 134.5˚ 132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

42
PRINICPLE
PRINICPLEAXIS
AXIS
pre-WTE post-WTE
Principal axes of strain rate Principal axes of strain rate
97.05.24 -- 00.08.06 00.11.04 -- 02.07.27
35.5˚ 35.5˚

35˚ 35˚

strain
34.5˚ 34.5˚

(extensional) (extensional)
34˚ 34˚
(contractional) (contractional)
0.5 [Micro Strain/Year] 0.5 [Micro Strain/Year]
132.5˚ 133˚ 133.5˚ 134˚ 134.5˚ 132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

Principal axes of stress rate Principal axes of stress rate

97.05.24 -- 00.08.06 00.11.04 -- 02.07.27
35.5˚ 35.5˚

35˚ 35˚

stress
34.5˚ 34.5˚

(extensional) (extensional)
34˚ 34˚
(contractional) (contractional)
100 [KPa/Year] 100 [KPa/Year]
132.5˚ 133˚ 133.5˚ 134˚ 134.5˚ 132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

43
POISSON
POISSONRATIO
RATIO
pre-WTE post-WTE
Poisson’s ratio Poisson’s ratio
Pre-Tottori Earthquake Mt. Daisen Post-Tottori Earthquake Mt. Daisen
35.5˚ 35.5˚
0.4 0.4

Naka-Umi Naka-Umi

35˚ 35˚
0.3 0.3

34.5˚ 34.5˚

0.2 0.2

34˚ 34˚

132.5˚ 133˚ 133.5˚ 134˚ 134.5˚ 132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

Change of Poisson’s ratio

Mt. Daisen
35.5˚
20

Naka-Umi

10
35˚

difference 0 ν has been reduced near the source

34.5˚
-10
faults. The comparison with other
analyses are being made.
-20
34˚
[%]
132.5˚ 133˚ 133.5˚ 134˚ 134.5˚

44
CONCLUDING
CONCLUDINGREMARKS
REMARKS
‹Two inverse analysis methods
– stress inversion
find Airy’s stress function by solving Poisson’s equation
– elasticity inversion
find elastic parameters by estimating displacement expansion
coefficients

‹Development of new inversion is needed for geophysics

where experiments cannot be made.

‹Application
– small material samples used for bio-mechanics
– geomaterials
– new image analysis with higher spatial resolution
45