Electromagnetic response data inversion using singular value decomposition

Patent metadata

Field	Value
Publication number	US-11061159-B2
Application number	US-201816164881-A
Country	US
Kind code	B2
Filing date	Oct 19, 2018
Priority date	Oct 20, 2017
Publication date	Jul 13, 2021
Grant date	Jul 13, 2021

Abstract

Inversion of enhanced-sensitivity controlled source electromagnetic data can include approximating a background response from measured controlled source electromagnetic (CSEM) response data. The approximation can include performing a first inversion of the CSEM response data using a largest singular value in a diagonal of a matrix associated with the CSEM response data to create a first resistivity model of a subsurface of a subterranean formation and iteratively performing subsequent inversions while increasing an amount of singular values in the diagonal to obtain modeled CSEM response data to create a second resistivity model of the subsurface of the subterranean formation. Inversion of enhanced-sensitivity controlled source electromagnetic data can further include storing results of the first inversion and the iterative subsequent inversions producing a resistivity map based on the first and the second resistivity models.

First claim

What is claimed is:

1. In a computer-implemented process for generating a resistivity map corresponding to geological features located beneath the surface of the earth (the “subsurface”), wherein the process is of the kind that applies inversion to measured, controlled-source electromagnetic (“CSEM”) response data to produce a resistivity model of the subsurface, and wherein the measured CSEM response data corresponds to measurements made by one or more electromagnetic receivers during a marine CSEM survey, the specific improvement of performing computer-implemented steps comprising:
generating, based on the measured CSEM response data, initial approximated response data corresponding to an initial level of complexity that is lower than that of the measured CSEM response data;
performing an initial inversion that uses as inputs the initial approximated response data and an initial resistivity start model and produces, as an output, an initial resistivity output model corresponding to the geological features; and
performing one or more subsequent inversions, wherein each of the subsequent inversions follows a previous inversion that produces a previous resistivity output model, and wherein performing each of the subsequent inversions comprises: (a) generating, based on the measured CSEM response data, subsequent approximated response data corresponding to a higher level of complexity than that of approximated response data used in the previous inversion; and (b) performing the subsequent inversion such that it uses as inputs the subsequent approximated response data and the previous resistivity output model and produces, as an output, a subsequent resistivity output model;
thereby improving the robustness of the process for generating the resistivity map and, correspondingly, the accuracy of the subsequent resistivity output model.

2. A process according to claim 1 :
further comprising representing the measured CSEM response data with a data structure that corresponds to a matrix; and
wherein generating the initial approximated response data comprises generating an initial approximation of the matrix.

3. A process according to claim 2 , wherein:
generating the initial approximation of the matrix comprises generating a reduced-rank approximation of the matrix.

4. A process according to claim 3 , wherein:
generating the reduced-rank approximation of the matrix comprises generating a singular value decomposition UΣV* of the matrix, wherein all but the largest k singular values of the matrix are set to zero in Σ, and wherein k is less than the rank of the matrix.

5. A process according to claim 4 , wherein k is equal to 1.

6. A process according to claim 4 , wherein:
generating the subsequent approximated response data comprises generating a singular value decomposition UΣV* of the matrix wherein all but the largest n singular values of the matrix are set to zero in Σ, and wherein n is greater than k.

7. A process according to claim 6 , wherein, for each subsequent inversion:
when generating the subsequent approximated response data, the number of non-zero singular values in Σ is one greater than the number of non-zero singular values used to generate approximated response data for the previous inversion.

8. One or more tangible, non-volatile, computer-readable media having instructions stored therein that, when executed on one or more computing devices, cause the computing devices to perform a process for generating a resistivity map that corresponds to geological features located beneath the surface of the earth, wherein the process comprises:
accessing measured, controlled-source electromagnetic (“C SEM”) response data, wherein the measured CSEM response data corresponds to measurements made by one or more electromagnetic receivers during a marine CSEM survey;
generating, based on the measured CSEM response data, initial approximated response data corresponding to an initial level of complexity that is lower than that of the measured CSEM response data;
performing an initial inversion that uses as inputs the initial approximated response data and an initial resistivity start model and produces, as an output, an initial resistivity output model corresponding to the geological features; and
performing one or more subsequent inversions, wherein each of the subsequent inversions follows a previous inversion that produces a previous resistivity output model, and wherein performing each of the subsequent inversions comprises: (a) generating, based on the measured CSEM response data, subsequent approximated response data corresponding to a higher level of complexity than that of approximated response data used in the previous inversion; and (b) performing the subsequent inversion such that it uses as inputs the subsequent approximated response data and the previous resistivity output model and produces, as an output, a subsequent resistivity output model.

9. Media according to claim 8 :
further comprising representing the measured CSEM response data with a data structure that corresponds to a matrix; and
wherein generating the initial approximated response data comprises generating an initial approximation of the matrix.

10. Media according to claim 9 , wherein:
generating the initial approximation of the matrix comprises generating a reduced-rank approximation of the matrix.

11. Media according to claim 10 , wherein:
generating the reduced-rank approximation of the matrix comprises generating a singular value decomposition UΣV* of the matrix, wherein all but the largest k singular values of the matrix are set to zero in Σ, and wherein k is less than the rank of the matrix.

12. Media according to claim 11 , wherein k is equal to 1.

13. Media according to claim 11 , wherein:
generating the subsequent approximated response data comprises generating a singular value decomposition UΣV* of the matrix wherein all but the largest n singular values of the matrix are set to zero in Σ, and wherein n is greater than k.

14. Media according to claim 13 , wherein, for each subsequent inversion:
when generating the subsequent approximated response data, the number of non-zero singular values in Σ is one greater than the number of non-zero singular values used to generate approximated response data for the previous inversion.

15. A method of manufacturing a geophysical data product containing resistivity indications corresponding to geological features located beneath the surface of the earth, by performing a computer-implemented process that comprises:
accessing measured, controlled-source electromagnetic (“C SEM”) response data, wherein the measured CSEM response data corresponds to measurements made by one or more electromagnetic receivers during a marine CSEM survey;
generating, based on the measured CSEM response data, initial approximated response data corresponding to an initial level of complexity that is lower than that of the measured CSEM response data;
performing an initial inversion that uses as inputs the initial approximated response data and an initial resistivity start model and produces, as an output, an initial resistivity output model corresponding to the geological features; and
performing one or more subsequent inversions, wherein each of the subsequent inversions follows a previous inversion that produces a previous resistivity output model, and wherein performing each of the subsequent inversions comprises: (a) generating, based on the measured CSEM response data, subsequent approximated response data corresponding to a higher level of complexity than that of approximated response data used

Assignees

• Pgs Geophysical As

Inventors

• Tshering Tashi
• Bekara Maiza
• Mckay Allan
• Vesterås Eivind Rödnes

Classifications

• G01V2003/086

  Processing · CPC title

• G01V3/083 Primary

  Controlled source electromagnetic [CSEM] surveying · CPC title