Geoscience Reference
In-Depth Information
After a certain number of iterations, the final vector,
m
w
, is transformed back to
assume that the resistivity model is perfectly known,
and 5% noise was added to the electrical potential data.
The first time corresponds to the time of occurrence of
the source, and therefore, we locate a type I anomaly.
In Figure 5.12a, we observe that the source current
density responsible for the electrical potential distribu-
tion is located in the vicinity of the true location of the
m
using the original relation-
Ω
−
1
ship
w
. We found that eight iterations in the
compactness algorithm are usually a good strategy to
focus the inverted source current distribution.
We apply this algorithm to the electrical potential
recorded at the 8 receiver stations three times. We
m
=
m
Localization of type I source at time = 0.150 s
×10
-7
0
2
200
1
400
0
−1
600
−2
800
−3
: True source location
1000
0
200
400
600
800
1000
(a)
Distance (m)
Localization of type II sources
×10
-7
At time = 0.217 s
At time = 0.248 s
0
0
0.2
200
200
0.1
0
400
400
−0.1
600
600
−0.2
800
800
−0.3
1000
1000
0
200
400
600
800
1000
0
200
400
600
800
1000
(b)
Distance (m)
Distance (m)
Figure 5.12
Localization of the source current density distribution responsible for the electrical potential distribution recorded
at the surface stations (5% noise).
a)
Type I anomaly associated with the seismic source.
b)
Type II anomalies (seismoelectric
conversions) (iteration #8, data root mean square error: 3%). The horizontal line denotes the interface between L1 and L2 where
seismoelectric conversions take place. (
See insert for color representation of the figure
.)
