Generating low frequency models for seismic waveform inversion in formation regions with limited control wells

Patent metadata

Field	Value
Publication number	US-11965996-B2
Application number	US-202217682284-A
Country	US
Kind code	B2
Filing date	Feb 28, 2022
Priority date	Feb 28, 2022
Publication date	Apr 23, 2024
Grant date	Apr 23, 2024

Abstract

The systems and methods described in this specification relate to generating a low frequency model of a subterranean formation for performing a seismic inversion. The systems and methods receive seismic data for a first region of the subterranean formation and well log data of one or more wells located at the first region. The systems and methods determine one or more relative layer attributes of the first region, one or more first input values for a machine learning model, and one or more second input values for the machine learning model. The systems and methods generate, a first relative low frequency model for the first region, and extrapolate, by executing the machine learning model by the processor, the first relative low frequency model to a second region of the subterranean formation.

First claim

What is claimed is:

1. A method of generating a low frequency model of a subterranean formation for performing a seismic waveform inversion, the method comprising:
receiving, at a processor from a first seismometer, seismic data for a first region of the subterranean formation, the seismic data representing a propagation of seismic waves within the first region;
receiving, at the processor from a well log instrument, well log data of one or more wells located at the first region;
determining, by the processor based on the received seismic data of the first region, one or more relative layer attributes of the first region, the one or more relative layer attributes representing one or more relative elastic properties of one or more formation layers of the first region;
determining, by the processor based on the one or more relative layer attributes of the first region, one or more first input values for a machine learning model;
determining, by the processor based on the received well log data, one or more second input values for the machine learning model;
generating, based on the one or more first input values and the one or more second input values, by executing the machine learning model by the processor, a relative low frequency model for the first region;
extrapolating, by executing the machine learning model by the processor, the first relative low frequency model to a second region of the subterranean formation, the second region being distinct from the first region;
updating, based on the extrapolating, by executing the machine learning model by the processor, the relative low frequency model to span the first region and the second region;
scaling, by the processor based on the well log data of the first region, the relative low frequency model;
generating, by the processor based on the scaling, a scaled low frequency model representing one or more scaled elastic properties of the one or more formation layers of the first region and one or more formation layers of the second region; and
performing, by the processor based on the scaled low frequency model, the seismic waveform inversion.

2. The method of claim 1 , wherein extrapolating the relative low frequency model to the second region comprises recursively extrapolating the relative low frequency model to one or more additional regions of the subterranean formation.

3. The method of claim 2 , wherein updating the relative low frequency model to span the first region and the second region comprises recursively updating, based on the recursive extrapolation, the relative low frequency model to span the first region, the second region, and the one or more additional regions.

4. The method of claim 1 , further comprising:
receiving, at the processor from a second seismometer, seismic data of the second region, the seismic data representing a propagation of seismic waves within the second region;
determining, by the processor based on the received seismic data of the second region, one or more relative layer attributes of the second region, the one or more relative layer attributes of the second region representing one or more relative elastic properties of the one or more formation layers of the second region; and
determining, by the processor based on the one or more relative layer attributes of the second region, one or more third input values for the machine learning model.

5. The method of claim 4 , wherein updating the relative low frequency model to span the first region and the second region comprises updating, based on the relative low frequency model and the one or more third input values, by executing the machine learning model by the processor, the relative low frequency model to span the first region and the second region.

6. The method of claim 5 , wherein the second region is void of wells.

7. The method of claim 5 , further comprising measuring, by the second seismometer, the seismic data of the second region.

8. The method of claim 7 , wherein the machine learning model is independent of well log data from the second region.

9. The method of claim 1 , wherein determining the one or more relative layer attributes of the first region comprises transforming, by the processor, the seismic data of the first region by a −90 degree phase shift prior to determining the one or more relative layer attributes.

10. The method of claim 1 , wherein the scaled low frequency model represents at least one of a density, a velocity, and an impedance for the one or more formation layers of the first region and the one or more formation layers of the second region.

11. The method of claim 1 , wherein updating the relative low frequency model to span the first region and the second region comprises updating the second relative low frequency model to span the first region and the second region in entirety.

12. The method of claim 1 , further comprising:
measuring, by the first seismometer, the seismic data at the first region; and
measuring, by the well log instrument, the well log data at the first region.

13. The method of claim 1 , further comprising:
determining, by the processor based on the seismic inversion, one or more well sites;
drilling, by a drill, one or more wellbores at each of the one or more well sites; and
extracting, by a pump, hydrocarbons from the one or more wellbores at the one or more well sites.

14. A system of generating a low frequency model of a subterranean formation for performing a seismic waveform inversion, the system comprising:
a first seismometer operable to measure seismic data of a first region of the subterranean formation, the seismic data representing a propagation of seismic waves within the first region;
a well log instrument operable to measure well log data of one or more wells located at the first region;
a computer storing computer instructions that, when executed by a processor of the computer, cause the processor to perform operations comprising:
receiving, from the first seismometer, the seismic data for the first region;
receiving, from the well log instrument, the well log data of the one or more wells located at the first region;
determining, based on the received seismic data of the first region, one or more relative layer attributes of the first region, the one or more relative layer attributes representing one or more relative elastic properties of one or more formation layers of the first region;
determining, based on the one or more relative layer attributes of the first region, one or more first input values for a machine learning model;
determining, based on the received well log data, one or more second input values for the machine learning model;
generating, based on the one or more first input values and the one or more second input values, by executing the machine learning model, a relative low frequency model for the first region;
extrapolating, by executing the machine learning model, the relative low frequency model to a second region of the subterranean formation, the second region being distinct from the first region;
updating, based on the extrapolating, by executing the machine learning model, the relative low frequency model to span the first region and the second region;
scaling, based on the well log data of the first region, the relative low frequency model;
generating, based on the scaling, a scaled low frequency model representing one or more scaled elastic properties of the one or more formation layers of the first region and one or more formation layers of the second region; and
performing, based on the scaled low frequency model, the seismic waveform inversion.

Assignees

• Saudi Arabian Oil Co

Inventors

• Daghistani Ahmed W

Classifications

• G01V1/282 Primary

  Application of seismic models, synthetic seismograms · CPC title

• G01V1/306

  for determining physical properties of the subsurface, e.g. impedance, porosity or attenuation profiles · CPC title

• G01V2210/6169

  using well-logging · CPC title

• G01V2210/614

  Synthetically generated data · CPC title

• G01V1/303

  for determining velocity profiles or travel times · CPC title