Hybrid optimization of fault detection and interpretation

Patent metadata

Field	Value
Publication number	US-11269100-B2
Application number	US-201716610031-A
Country	US
Kind code	B2
Filing date	Dec 21, 2017
Priority date	Aug 18, 2017
Publication date	Mar 8, 2022
Grant date	Mar 8, 2022

Abstract

A method includes receiving a training selection of a first set of faults located in a first subset of a seismic dataset for a subsurface geologic formation, detecting a second set of faults in the seismic dataset based on fault interpretation operations using a first set of interpretation parameters, and determining a difference between the first set of faults and the second set of faults. The method also includes generating a second set of interpretation parameters for the fault interpretation operations based on the difference between the first set of faults and the second set of faults, and determining a feature of the subsurface geologic formation based on fault interpretation operations using the second set of interpretation parameters.

First claim

What is claimed is:

1. A method comprising:
obtaining a first labeled seismic dataset for a subsurface geologic formation;
iteratively determining fault interpretation parameters for a fault interpretation algorithm to tune the fault interpretation algorithm, wherein each iteration comprises,
obtaining a second labeled seismic dataset from running the fault interpretation algorithm on an unlabeled seismic dataset for the subsurface geologic formation using fault interpretation parameters determined for the current iteration;
determining a quantified difference between the first labeled seismic dataset and the second labeled seismic dataset of the current iteration; and
based, at least in part, on the quantified difference between the first labeled seismic dataset and the second labeled seismic dataset of the current iteration, determining fault interpretation parameters for a next iteration with stochastic optimization.

2. The method of claim 1 , wherein the stochastic optimization is a Bayesian optimization.

3. The method of claim 1 , wherein the seismic dataset is a three-dimensional seismic dataset.

4. The method of claim 1 , wherein iteratively determining the fault interpretation parameters for the fault interpretation algorithm further comprises iteratively determining the fault interpretation parameters with the stochastic optimization until the quantified difference is minimized.

5. The method of claim 1 further comprising configuring a second fault interpretation algorithm with the determined fault interpretation parameters corresponding to a minimized quantified difference.

6. The method of claim 1 further comprising running the fault interpretation algorithm on a second unlabeled seismic dataset with the fault interpretation parameters determined in the iteration corresponding to a minimal one of the quantified differences.

7. The method of claim 1 , wherein the quantified difference is based on an at least one of a Euclidean norm, a Manhattan norm, and a Minkowski norm.

8. The method of claim 1 , wherein iteratively determining the fault interpretation parameters for the fault interpretation algorithm comprises running the fault interpretation until reaching an iteration step threshold.

9. One or more non-transitory machine-readable media comprising program code, the program code to:
obtain a labeled version of a first subset of a seismic dataset for a subsurface geologic formation;
iteratively determine fault interpretation parameters to tune a fault interpretation algorithm, wherein each iteration comprises,
running the fault interpretation algorithm on the first subset of the seismic dataset with fault interpretation parameters for a current iteration to obtain fault predictions of the current iteration;
determine a quantified difference between the labeled version of the first subset of the seismic dataset and the fault predictions of the current iteration;
determine fault interpretation parameters for a next iteration with stochastic optimization based on the quantified difference; and

with the tuned fault interpretation algorithm, obtain fault predictions for at least a second subset of the seismic dataset of the subsurface geologic formation.

10. The one or more non-transitory machine-readable media of claim 9 , wherein the stochastic optimization is a Bayesian optimization.

11. The one or more non-transitory machine-readable media of claim 9 , wherein the seismic dataset is a three-dimensional seismic dataset.

12. The one or more non-transitory machine-readable media of claim 9 , further comprising program code to iteratively determine the fault interpretation parameters for the fault interpretation algorithm with the stochastic optimization until the quantified difference is minimized.

13. The one or more non-transitory machine-readable media of claim 9 , wherein the quantified difference is based on at least one of a Euclidean norm, a Manhattan norm, and a Minkowski norm.

14. The one or more non-transitory machine-readable media of claim 9 , wherein the program code to iteratively determine the fault interpretation parameters for the fault interpretation algorithm with the stochastic optimization comprises the program code to iteratively determine the fault interpretation parameters until an iteration step threshold is reached.

15. A system comprising:
a processor; and
a machine-readable medium having program code executable by the processor to cause the processor to:
obtain a labeled version of a first subset of a seismic dataset for a subsurface geologic formation;
iteratively determine fault interpretation parameters to tune a fault interpretation algorithm, wherein each iteration comprises,
running the fault interpretation algorithm on the first subset of the seismic dataset with fault interpretation parameters for a current iteration to obtain fault predictions of the current iteration;
determine a quantified difference between the labeled version of the first subset of the seismic dataset and the fault predictions of the current iteration;
determine fault interpretation parameters for a next iteration with stochastic optimization based on the quantified difference; and

with the tuned fault interpretation algorithm, obtain fault predictions for at least a second subset of the seismic dataset of the subsurface geologic.

16. The system of claim 15 , wherein the stochastic optimization is a Bayesian optimization.

17. The system of claim 15 , wherein the seismic dataset is a three-dimensional seismic dataset.

18. The system of claim 15 , wherein the quantified difference is based on an at least one of a Euclidean norm, a Manhattan norm, and a Minkowski norm.

19. The system of claim 15 , further comprising program code to iteratively determine the fault interpretation parameters for the fault interpretation algorithm with the stochastic optimization until the quantified difference is minimized.

20. The system of claim 15 , wherein iteratively determining the fault interpretation parameters for the fault interpretation algorithm further comprises the program code to iteratively determine the fault interpretation parameters with stochastic optimization until an iteration step threshold is reached.

Assignees

• Landmark Graphics Corp

Inventors

• Mao Youli
• Pandya Raja Vikram
• Mandapaka Bhaskar
• Rangarajan Keshava Prasad
• Madasu Srinath
• Priyadarshy Satyam
• Dev Ashwani

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06F18/211

  Selection of the most significant subset of features · CPC title

• G01V1/30 Primary

  Analysis (G01V1/50 takes precedence) · CPC title

• G01V2210/643

  Horizon tracking · CPC title

• G01V1/301 Primary

  for determining seismic cross-sections or geostructures · CPC title