# New advances in three-dimensional controlled-sourceelectromagnetic inversion

## Abstract

New techniques for improving both the computational andimaging performance of the three dimensional (3D) electromagnetic inverseproblem are presented. A non-linear conjugate gradient algorithm is theframework of the inversion scheme. Full wave equation modelling forcontrolled sources is utilized for data simulation along with anefficient gradient computation approach for the model update. Improvingthe modelling efficiency of the 3D finite difference method involves theseparation of the potentially large modelling mesh, defining the set ofmodel parameters, from the computational finite difference meshes usedfor field simulation. Grid spacings and thus overall grid sizes can bereduced and optimized according to source frequencies and source-receiveroffsets of a given input data set. Further computational efficiency isobtained by combining different levels of parallelization. While theparallel scheme allows for an arbitrarily large number of parallel tasks,the relative amount of message passing is kept constant. Imageenhancement is achieved by model parameter transformation functions,which enforce bounded conductivity parameters and thus prevent parameterovershoots. Further, a remedy for treating distorted data within theinversion process is presented. Data distortions simulated here includepositioning errors and a highly conductive overburden, hiding the desiredtarget signal. The methods are demonstrated using both synthetic andfield data.

- Authors:

- Publication Date:

- Research Org.:
- Ernest Orlando Lawrence Berkeley NationalLaboratory, Berkeley, CA (US)

- Sponsoring Org.:
- USDOE Director. Office of Science. Basic EnergySciences

- OSTI Identifier:
- 925536

- Report Number(s):
- LBNL-63010

Journal ID: ISSN 0956-540X; R&D Project: 468113; BnR: KC0303010; TRN: US200809%%793

- DOE Contract Number:
- DE-AC02-05CH11231

- Resource Type:
- Journal Article

- Resource Relation:
- Journal Name: Geophysical Journal International; Journal Volume: 172; Journal Issue: 2; Related Information: Journal Publication Date: 02/2008

- Country of Publication:
- United States

- Language:
- English

- Subject:
- 54; ALGORITHMS; EFFICIENCY; FINITE DIFFERENCE METHOD; OVERBURDEN; PERFORMANCE; POSITIONING; SIMULATION; TARGETS; TRANSFORMATIONS; WAVE EQUATIONS

### Citation Formats

```
Commer, Michael, and Newman, Gregory A.
```*New advances in three-dimensional controlled-sourceelectromagnetic inversion*. United States: N. p., 2007.
Web.

```
Commer, Michael, & Newman, Gregory A.
```*New advances in three-dimensional controlled-sourceelectromagnetic inversion*. United States.

```
Commer, Michael, and Newman, Gregory A. Sat .
"New advances in three-dimensional controlled-sourceelectromagnetic inversion". United States.
doi:. https://www.osti.gov/servlets/purl/925536.
```

```
@article{osti_925536,
```

title = {New advances in three-dimensional controlled-sourceelectromagnetic inversion},

author = {Commer, Michael and Newman, Gregory A.},

abstractNote = {New techniques for improving both the computational andimaging performance of the three dimensional (3D) electromagnetic inverseproblem are presented. A non-linear conjugate gradient algorithm is theframework of the inversion scheme. Full wave equation modelling forcontrolled sources is utilized for data simulation along with anefficient gradient computation approach for the model update. Improvingthe modelling efficiency of the 3D finite difference method involves theseparation of the potentially large modelling mesh, defining the set ofmodel parameters, from the computational finite difference meshes usedfor field simulation. Grid spacings and thus overall grid sizes can bereduced and optimized according to source frequencies and source-receiveroffsets of a given input data set. Further computational efficiency isobtained by combining different levels of parallelization. While theparallel scheme allows for an arbitrarily large number of parallel tasks,the relative amount of message passing is kept constant. Imageenhancement is achieved by model parameter transformation functions,which enforce bounded conductivity parameters and thus prevent parameterovershoots. Further, a remedy for treating distorted data within theinversion process is presented. Data distortions simulated here includepositioning errors and a highly conductive overburden, hiding the desiredtarget signal. The methods are demonstrated using both synthetic andfield data.},

doi = {},

journal = {Geophysical Journal International},

number = 2,

volume = 172,

place = {United States},

year = {Sat May 19 00:00:00 EDT 2007},

month = {Sat May 19 00:00:00 EDT 2007}

}