Académique Documents
Professionnel Documents
Culture Documents
Sebastiano Foti
Email: sebastiano.foti@polito.it
(ITALY) www.soilmech.polito.it/people/foti_sebastiano
Surface wave methods
u ( x, t ) = A e i ( kx t ) + B e i ( kx +t )
u = B1 sin (kx + t ) + B2 sin (kx t ) + B3 cos(kx + t ) + B4 cos(kx t )
k = 2
Symbol Quantity Dimensions SI Unit
A Amplitude various various T = 2
Radial frequency [1/time] [rad/s]
f Frequency (ciclic) [cicles/time] [Hz=1/s] = kVB
Wavelength [length] [m] VB
k Wavenumber [1/length] [1/m] =
f
V Phase velocity [length /time] [m/s]
2f
T Period [time] [s] VB =
k
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
Rayleigh Waves
z/R
VR
R =
f Animation courtesy of Prof. Braile
VR
R =
(after Richart et al., 1970)
f
% of total
W ave Type
energy
R ayleigh 67
Shear 26
(After Woods, 1968) C om pression 7
Geometric Dispersion
?
Wavelength
VR = f
VS2> VS1
VS3> VS2
Z Z Frequency f
INVERSE PROBLEM
Controlled or
impact
source frequency components
Low frequency vertical geophones
1 2 3 n
VS1
VS2
VS3
Processing
VR
Experimental dispersion curve: Phase velocity
of Rayleigh waves vs frequency
Inversion
VS G0
Variations of Shear Wave velocities with depth
G0 = VS2
Small Strain Stiffness profile (G0 vs depth)
Z
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
Processing
VR
Experimental dispersion curve: Phase velocity
of Rayleigh waves vs frequency
Inversion
VS G0
Variations of Shear Wave velocities with depth
G0 = VS2
Small Strain Stiffness profile (G0 vs depth)
Z
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
1 2 3 n
Multistation:
Active methods
Two-station (SASW)
Passive methods
Spatial Array:
Spatial Autocorrelation
(SPAC, ESAC), f-k spectra
(FDBF, MLM, Music), ...
1 2 3 n
Linear array (ReMi) ?
X X
Testing
Equipment
Impact Sources
14
hammer
13 sledge hammer
w eight drop
autopow er spectrum (log)
12
11
10
7
0 50 100 150
frequency, Hz
Minibang
Controlled Sources
Electromechanical
shaker
Transducers
Geophones
(Doyle, 1995)
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
Acquisition
Geophones
2Hz
4.5Hz
2Hz 10Hz
Seismographs
Channel 1
+ 18
A/D 10
- 16
Mplx. S&H Conv.
+
-
Channel n
Seismographs
Main characteristics:
Recording window - usually (0.25 - 2 s).
Sampling rate typically (25 - 1000 s).
amplification of the signal (A.G.C. - automatic gain control).
dynamic ratio between max measurable value and sensitivity,
expressed in number of bit (18 24 bits).
stacking during acquisition to improve S/N ratio
Displaying of acquired signals
trigger
noise level test
memory
Weight and power
(Courtesy L.V. Socco)
Seismographs
Signal Analysers
Home made
cheaper
specific software
not easy
Trigger
Processing
VR
Experimental dispersion curve: Phase velocity
of Rayleigh waves vs frequency
Inversion
VS G0
Variations of Shear Wave velocities with depth
G0 = VS2
Small Strain Stiffness profile (G0 vs depth)
Z
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
x=D
Harmonic Source
D X
x=D+X
Usually D=X
Phase Velocity
VR () =X/ t
t
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
Fourier Transform
T=1/f
geophone 1 output
1 1
0 0
-1 -1
-2 -2
0 1 2 3 4 1.8 1.82 1.84 1.86 1.88 1.9
5 5
x 10 x 10
Impulsive, Sinusoidal or 2 2
geophone 2 output
1 1
Random Noise Source 0 0
-2 -2
0 1 2 3 4 1.8 1.82 1.84 1.86 1.88 1.9
time, s time, s
200
Im(G12 ( ))
CPS Phase
12 ( ) = tan 1
Re(G12 ( ))
(a)
0
-200
0 50 100 150
1
G12 ( ) G12 ( )
Coherence
0.5 (b) 12
2
( ) =
G11 ( ) G22 ( )
0
150 50 100 150
10
APS 1
G11 ( ) = Y1 ( ) Y1 ( )
10 (c)
10
5
10
15
0 50 100 150
10
G22 ( ) = Y2 ( ) Y2 ( )
APS 2
10 (d)
10
5
10
0 50 100 150
frequency, Hz
D
< < 2D
3
Common receiver
midpoint array
Heavy sources are used with larger spacing to obtain low
frequency (long wavelength) information
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
Dispersion curve
700
700
sledge-hammer(D=3m)
sledge-hammer(D=3m)
sledge-hammer(D=6m)
sledge-hammer(D=6m)
600 weight-drop(12m)
weight-drop(12m)
600
weight-drop(18m)
weight-drop(18m)
weight-drop(30m)
m/s
velocity, m/s
500
500
phase velocity,
400
400
phase
300
300
200
200
100
100
100
00
0 10
10
10 20
20
20 30
30
30 40
40
40 50
50
50 60
60 70
70 80
80 90
90 100
100
frequency, Hz
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
Phase unwrapping
200
0
CPS phase, degrees
-200
-400
-600
-800
wrapped
unwrapped
-1000
0 50 100 150
frequency, Hz
Transform-based Methods: fk
Hz
frequency,Hz
frequency,
time (s)
2D FFT
slowness, s/m
time (s)
p +
FFT
receiver offset (m) frequency, Hz
MAXIMA
frequency, Hz
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
Transform-based Methods
slowness 500
Equivalent 300
procedures
200
100
0 10 20 30 40 50 60 70
frequency, Hz
(Foti, 2000)
fk analysis results
1 2 3 24 1 2 3 24
1m 1m 1m 3m 3m 3m
700
sledge-hammer (1m)
600
weight-drop (3m)
phase velocity, m/s
500
400
300
200
100
0 10 20 30 40 50 60 70
frequency, Hz
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
700
SASW
fk analysis
600
phase velocity, m/s
500
400
300
200
100
0 20 40 60 80 100
frequency, Hz
Processing
VR
Experimental dispersion curve: Phase velocity
of Rayleigh waves vs frequency
Inversion
VS G0
Variations of Shear Wave velocities with depth
G0 = VS2
Small Strain Stiffness profile (G0 vs depth)
Z
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
particle motion
Approximate Inversion (SSRM) 0
1
depth/wavelength
0.8
0.5 0.6
V/Vs
S-wave
0.4
Mapping rule 0.2
R-wave
1
experimental estimated 0 0.5
dispersion stiffness 1.5 Poisson Ratio
curve profile Example with experimental data
VR VS Vs (m/s)
VR* 1.1 VR*
100 200 300 400 500 600 700
0
*R
5
3
*R 10
Depth (m)
15
20
25
R Depth 30
35
FORWARD PROBLEM
Wavelength
?
VR = f
VS2> VS1
VS3> VS2
Z Z Frequency f
INVERSE PROBLEM
H1 1 G1 1 460
H2 2 G2 2 440
300
0 50 100 150
frequency, Hz
Considering an active source: mode superposition
For simple stratigraphies (stiffness increasing with depth) the fundamental mode is
dominant and mode superposition can be neglected
Usual assumptions
H3 =? Vs3=? 400
Vs=? 300
200
Usually i and i are fixed 100
and Hi and Gi (or VSi) are 0 10 20 30 40 50 60 70
Inversion algorithms
Inversion Strategies
5
Dispersion curve fitting
10 700
experimental
Depth (m)
600
25 300
200
30
100
CHT 0 10 20 30 40 50 60 70
35
frequency, Hz
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
5 10 10
Depth (m)
Depth (m)
Depth (m)
15 15
10 20 20
Down Hole
25 Cross Hole 25 Cross Hole
SASW-fk
SASW-fk SASW-fk
15 30 30
Vs (m/s) Vs (m/s) Vs (m/s)
0 400 800 0 400 800 1200 0 400 800 1200 1600
0 0 0
Pontremoli Pontremoli Pontremoli
site 1 site 2 site 3
5 5 5
Depth (m)
Depth (m)
Depth (m)
10 10 10
Vs (m/s) 700
0 400 800 VR 620experimental
ms
max = = 70m z max 35m
0 H1 =? Vs1=? 600 f f max
8.5 Hz
numerical
Vs=? 300
15
200
Usually i and i are fixed
20
and Hi and Gi (or Vsi) are 100
0 10 20 30 40 50 60 70
25 theCross Hole
unknowns
SASW-fk frequency, Hz
30
Active+Passive - SW Tests
Passive Active
Processing
VR
Passive Active
Inversion
VS
Active
Passive
Z
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
1100 DHT
SW
1050
SW
1000
950 SW
900
850
800
750
700
VS [m/s]
25 20
Profondit, z (m)
Profondit, z (m)
30
25
35
30
40
45 35
50 40
55 LAquila 45 LAquila
60 Roio Piano Pianola
50
65
70 55
UBC EC8
30
VS ,30 =
hi
V
i =1.. N S ,i
(Louie, 2001)
ReMi (Refraction Microtremors)
= Passive Surface Wave Tests with linear arrays
Seismograph or Signal Analyzer
Microtermors arrays
Experimental data at La Salle test site
2D array ReMi linear array
apparent
true
v=f
Be careful it is not
the bedrock!
Spatial resolution
A-priori hypothesis
Non-uniqueness
Higher modes
1D model pseudo 2D
Spatial resolution
A-priori hypothesis
Non-uniqueness
Higher modes
1D model pseudo 2D
Soil Model
In standard practice i and i (or VPi) are fixed a-priori while Hi and VSi=
(Gi/ i) are the unknowns (2n-1) [Stokoe et al., 1984]
Soil Density 1.2 2.0 1.8 2.3 Weight of water filling the voids
Experimental Data
0
0 200 400 600 800 1000 Hp#1 Water table from P-wave refraction
starting profile Hp#2 No water table
inversion #1
inversion #2 Hp#3 Water table deeper than Hp #1
5
inversion #3
cross-hole test
700
experimental
Depth (m)
10
inversion #1
600 inversion #2
400
20
300
200
25
100
0 10 20 30 40 50 60 70
30 frequency, Hz
Spatial resolution
A-priori hypothesis
Non-uniqueness
Higher modes
1D model pseudo 2D
Inverse methods
Experimental data
Model
fit
parameters
Numerical simulation
(forward problem) (solution of the inverse problem)
Solution non-uniqueness
(equivance of several possible solutions with respect to the experimental data)
0.5
0.4
0.6
top
acceleration (g)
0.3
0.5
0.2
0.4
0.3 outcrop
0.1
0
0.2
-0.1
0.1
0
-0.2
-0.1
-0.3
-0.2
-0.4
-0.3
-0.5
-0.4
-0.6
0 20 40 60 80
-0.5
te mpo (s) -0.6
time (s) 0 20 40
tempo (s)
60 80
time (s)
Site characterization:
Shear wave velocity model
1D Vs profile
2D/3D Vs models to simulated
S-wave
complex situation (e.g. valley
edges)
Consequences of non-uniqueness
Soil Profile
Experimental Data Local Site Response
UBC EC8
30
VS ,30 =
hi
V
i =1.. N S ,i
sand
?
claystone
Top soil
sand
claystone
sand
claystone
Spatial resolution
A-priori hypothesis
Non-uniqueness
Higher modes
1D model pseudo 2D
440
H1 1 G1 1
H2 2 G2 2 400
380
H3 3 G3 3 360
4 G4 4
340
320
300
0 50 100 150
frequency, Hz
Higher modes can be often retrieved but are difficult to be included in the
inversion because they can hardly be numbered.
Even when a single continuous curve is retrieved and assumed to be the
fundamental mode, higher modes can be present and this can drive the
inversion into sever pitfalls.
Higher modes contain further information can therefore contribute to better
constraints the results.
Geocongress 2014 23rd February 2014 SEBASTIANO FOTI POLITECNICO DI TORINO
Surface wave methods
f(Hz)
t(s)
k(1/m)
x(m)
Seismogram fk spectrum
determinant inversion
* synthetic data determinant
synthetic profile inversion
determinant inversion
theoretical modal curves
Subsoil category A* A
Vs
Vs
Vp
experimental - no caprock
350
250
150 100
0 20 40 60 80 100 120 140 10 20 30 40 50 60 70 80 90 100
4
caprock
depth [m]
depth [m]
no caprock
6
6
SASW
SCPT1 left
8
9 SCPT1 right
SCPT2 left
10 SCPT2 right
Spatial resolution
A-priori hypothesis
Non-uniqueness
Higher modes
1D model pseudo 2D
Pseudo-2D (3D)
Local approximation of submerged structure with 1D profiles
\
VS1
VS2
VS3
Independent Inversions
1 2 3 1 2 3 1 2 3 1 2 3
VR VR VR VR
f f f f
1 2 3 1 2 3 1 2 3 1 2 3
VR VR VR VR
f f f f
Misfit
Vs [m/s]
A-priori information
3D VS model
VS (m/s)
1000
800
600
400
200
(Piatti et al., 2013)