New 3D parallel SGILD modeling and inversion Page: 17 of 20
This article is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
Xie et al, LBNL 42252, New Parallel SGILD Modeling and Inversion, September 1, 1998
The SGILD algorithm can be used for the electromagnetic and flow modeling
and inversion, see  and . In , new magnetic boundary and volume
integral equations for the moments of the resistivity, permittivity, and the
magnetic field are derived . A 2.5D SGILD electromagnetic code is tested
primarily using a synthetic and field data. The mean resistivity imaging and
standard deviations are presented. In Figure 2, 16 frequencies, 6 electric line
sources on the surface and 20 receivers in the vertical logging well are used to
make synthetic data with Gaussian noise, the maximum standard deviation
of the data is 5%. The high resolution imaging of the mean resistivity is
obtained. The total maximum standard deviation (TSTD) of the resistivity is
11.8%, The local standard deviation (LSTD) of the resistivity of the target in
left top corner (read) is 6%, The other local standard deviation of resistivity
in right lower corner (blue) is 18.6%, that is because the read block is in the
coverage area of the data site. The 2D mesh is 128x128, 64 x 30.5 CPU minites
in T3E and 58 iterations are used to obtained these moments imaging. The
optimization mean regularizing is 0.687456x10-6. Other resistivity imaging
from practical field data in the geothermal exploration is presented in . The
field data configuration includes 16 frequencies, 6 electric line sources on the
surface and 20 receivers in the vertical logging well. The maximum standard
deviation of the field data is 21%. A reasonable mean resistivity imaging is
obtained. The maximum standard deviation of the resistivity is 31.8%, The
local standard deviation of resistivity near the borehole area is 19%, which is
less than standard deviation of the field data. The second order mean term
is effective to improve the resolution of the mean resistivity imaging. The 2D
mesh is 256x256, optimization mean regularizing is 0.32934x10-1, the 64 x 3.8
CPU hours in T3E and 96 iterations were used to these moments imaging. The
parallel rate of the primary SGILD code is 70% 90%. The SGILD acoustic
velocity imaging and data configuration is presented in  and SGLID flow
permeability inversion is presented in .
The primary tests shown that the SGILD modeling and inversion is a high
resolution , robust stable, and high performance parallel imaging algorithm.
There are obvious improvements of resolution of imaging from the field data.
Actually, most of the conventional deterministic inversion approaches were
only used to obtain the zero order mean of the target parameters, but no
second order correction term and the standard deviation term. The SGILD
algorithm can be used to obtain the improved ensemble mean parameter with
second correction term, cross covarience between the parameter and field, and
Here’s what’s next.
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
Xie, G.; Li, J. & Majer, E. New 3D parallel SGILD modeling and inversion, article, September 1, 1998; Berkeley, California. (https://digital.library.unt.edu/ark:/67531/metadc710929/m1/17/: accessed May 25, 2019), University of North Texas Libraries, Digital Library, https://digital.library.unt.edu; crediting UNT Libraries Government Documents Department.