Down-Hole Geophysical Testing for Rock Sockets Dennis R. Hiltunen

Down-Hole Geophysical Testing for
Rock Sockets
Dennis R. Hiltunen
Pengxiang Jiang
University of Florida
FDOT GRIP
August 17, 2012
Ft. McCoy
■ Surface array: 31
geophones at 1-m spacing
■ Borehole array: geophones
4-18 m at 1-m spacing
■ Sledgehammer source:
1,2,3,4,5,6,8,10,12,15 m each
side of well
Ft. McCoy
60
Observed
Estimated
50
-4
-6
-8
Depth (m)
Arrival time (ms)
40
30
20
-10
-12
Observed
Estimated
-14
10
-16
0
-18
5
-10
0
5
10
15
Distance (m)
20
25
30
10
15
20
Arrival time (ms)
25
30
35
Ft. McCoy
-2
1800
-4
1600
-6
1400
1200
-8
1000
-10
800
-12
600
-14
P_wave velocity (m/s)
Depth (m)
0
400
-16
-18
0
5
10
15
20
25
30
Distance (m)
d)
-2
55
50
-4
45
40
Depth (m)
-6
35
-8
30
25
-10
20
-12
15
10
-14
5
0
-16
-18
0
5
10
15
Distance (m)
20
25
30
Coefficient of Variation (%)
0
Workplan
■ Task 1: Review Literature and State of Practice
 Single-well borehole logging devices
■ Task 2: Computer and Synthetic Model Studies
 Borehole instrumentation, full waveform analysis
 Preliminary array design: geometry, instrumentation
■ Task 3: Array Experiments
 Test/modify preliminary array design
 In-house instrumentation
■ Task 4: Design Borehole Tool
 Mechanical design of array, analysis software
■ Task 5: Report
Chabot (2003)
 University of Calgary
 Sonic logging tool in
fluid-filled borehole
 Seismic reflectionstyle processing
 Full waveform
analysis of body
waves (P and S)
 No surface waves
Kalinski (1998)
 University of Texas
(Stokoe)
 SASW along axis of
borehole (1-D)
 Concrete, rock, and
soil
 Geometry-induced
dispersion
Normalized magnitude
Full Waveform: FD vs. FEM
0
300 m/s
5
10
200 m/s
1
25
400 m/s
30
35
0
20
40
X, m
60
80
FDM
FEM
0.5
0
-0.5
-1
20
Normalized magnitude
Y, m
15
Receiver @ 30m
0
0.2
0.4
0.6
0.8
1
Receiver @ 60m
1
FDM
FEM
0.5
0
-0.5
-1
0
0.2
0.4
0.6
time in sec
0.8
1
Full Waveform: LVL at TAMU
SV-Wave Velocity (m/s)
0
100
200
300
400
0
500
Standard Penetration
Resistance
(SPT N-Value)
0
2
12
15
18
4
15
14
Tan Sand
6
13
17
Tan Silty Fine
Sand
Tan Silty Fine
Sand
6
Tan Sandy Clay
10
10
60
12
70
Dark Gray Clay
w/Gravel
52
12
14
62
16
4
8
20
14
2
Gravel
16
8
Depth (m)
Soil Profile
0 10 20 30 40 50 60 70 80 90
Dark Gray Clay
16
18
18
Crosshole test
20
22
20
Full wave inversion of
3 layers + half space
Full wave inversion of
4 layers + half space
22
24
24
EOB 15.2m
Full Waveform: Seg 2D at TAMU
FWI: Layer
 2D finite difference
model
 Simulated
annealing and
genetic algorithm
 Surface array
 Array and source
design
FWI: Block
FWI: Block in Gradient
FWI: 5x5 Grid, 0.5 m
Forward Model Questions
■ Original premise: Can we slice a radial plane
from the borehole, put a source and array of
receivers along the vertical borehole axis edge,
and then model with a 2D, plane-strain, flatground model?
■ Under investigation: Or, do the actual
surroundings of the borehole geometry
significantly influence the wavefield and the
waveforms collected along the borehole axis?
Geometry-Induced Dispersion
2D, Flat-Ground Models
 Plane-strain: FD,
Plaxis, Abaqus
 Axisymmetric: Plaxis,
Abaqus
 Waveforms similar for
all three codes
 Movies
2.5D, Axisymmetric, Borehole Model
 Two models: Plaxis,
Abaqus
 Ring load
 Waveforms similar for
both codes
3D Borehole Model




Abaqus
Ring load
Point load
Movies
3D Borehole Model, 50 cm Radius
 Code: Abaqus
 Configuration: R=10 m, r=50
cm, L=20 m
 Properties: Vs=1000 m/s
 Array: 20 @ 0.25 m along
borehole wall
 Source: point, triangular,
center of cylinder, S=0.25 m
3D Borehole Model, 5 cm Radius
 Code: Abaqus
 Configuration: R=10 m, r=5
cm, L=20 m
 Properties: Vs=1000 m/s
 Array: 20 @ 0.25 m along
borehole wall
 Source: point, triangular,
center of cylinder, S=0.25 m
Compare: Waveforms
2D
50 cm
2.5D
5 cm
Compare: Waveforms at x=middle
Compare: Dispersion Images
2D
50 cm
2.5D
5 cm
Inversion with 2D, Plane-Strain,
Flat-Ground Model: 3D Flat Ground





Code: Abaqus synthetic
Configuration: 20 m cube
Properties: Vs=1000 m/s
Array: 10 @ 0.5 m
Source: point, triangular,
S=0.5 m
 Inversion: 6 m x 6 m, 2 layers
+half-space
 Waveforms not well matched,
but return about 1000 m/s
Inversion with 2D, Plane-Strain,
Flat-Ground Model: 50 cm Radius
 Code: Abaqus synthetic
 Configuration: R=10 m, r=50
cm, L=20 m
 Properties Vs=1000 m/s
 Array: 10 @ 0.5 m along
borehole wall
 Source: point, triangular, center
of cylinder, S=0.5 m
 Inversion: 6 m x 6 m, 2 layers
+half-space
 Waveforms not well matched,
return stiffer than 1000 m/s
Forward Model Questions
■ Summary: appears that borehole influence is
significant, based on wavefield movies,
dispersion curves, waveform comparisons, and
inversions assuming 2D plane-strain, flatground model
■ Future: original premise does not appear
feasible, require a forward model that includes
borehole geometry, e.g., build a FD model, use
existing FEM code
Two Solutions
■ Have built two inversion models that use
borehole model in Abaqus to invert waveforms
collected along vertical wall of borehole
■ First inversion model uses simulated annealing
■ Second inversion model uses linearized, local
inversion following recent work of Tran
■ 2.5D a few seconds, 3D an hour for each
forward
■ Good for a ring experiment, 2.5D and 3D similar
Abaqus FWI: Flat-Ground, 5x5 Grid
9
True model
x 10
9
0.5
1
8
1.5
7
2
9
Inverted model
x 10
9
0.5
1
8
1.5
7
2
6
2.5
6
2.5
3
5
3.5
4
4
3
5
3.5
4
4
3
4.5
2
5
5.5
3
4.5
1
2
3
4
5
1
2
5
5.5
1
2
3
4
5
1
Abaqus FWI: 2.5D Borehole, 5x5 Grid
9
True model
x 10
5
0.5
1
4.5
1.5
9
Inverted model
x 10
0.5
6
1
5.5
1.5
4
2
5
2
3.5
2.5
4.5
2.5
4
3
3
3
3.5
3.5
2.5
4
3.5
3
4
2.5
2
4.5
4.5
5
5.5
2
1.5
5
1
2
3
4
5
1
5.5
1.5
1
1
2
3
4
5
Array Experiments
■ Test multi-sensor array concept and full
waveform inversion on experimental data
■ Use existing instrumentation
■ Flat surface and borehole
■ Do we use fluid-filled or dry hole?
■ How to do tests on a model with correct-scale
array and instruments? Effects of boundaries?
■ How to design/construct a model with
properties of Florida limerock?
Synthetic Limerock Specimen
data fit
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
data
fit
0
1
2
3
4
5
6
8
7
-4
x 10
Synthetic Limerock Specimen
 Preliminary results
 Vs = 935 m/s from FWI
 Free-free resonant
column tests
 Vp = 1500 m/s
 Poisson’s ratio = 0.2
Thank You!