Automated reservoir model prediction using ML/AI intergrating seismic, well log and production data

Patent metadata

Field	Value
Publication number	US-12353807-B2
Application number	US-202017016075-A
Country	US
Kind code	B2
Filing date	Sep 9, 2020
Priority date	Sep 9, 2020
Publication date	Jul 8, 2025
Grant date	Jul 8, 2025

Abstract

Methods and apparatus for generating one or more reservoir 3D models are provided. In one or more embodiments, a method can include training a first machine learning model to generate one or more integrated enhanced logs based, at least in part, on an integrated data set, wherein the integrated data set includes seismic data and well log data; generating one or more integrated enhanced logs from the first machine learning model; grouping the one or more integrated enhanced logs into an ensemble of integrated enhanced logs to form a static reservoir 3D model of a subterranean reservoir; inputting additional data to the first machine learning model to produce one or more updated integrated enhanced logs; and grouping the one or more updated integrated enhanced logs into an ensemble of updated integrated enhanced logs to form an updated 3D model.

First claim

The invention claimed is:

1. A method comprising:
training a first machine learning model using a first neural network to generate one or more integrated enhanced logs based, at least in part, on an integrated data set, wherein the integrated data set includes 2D, 3D, and 4D seismic attributes derived from a phase, a frequency, and an amplitude of seismic signal data, the 2D, 3D, and 4D seismic attributes matched to well log data by bringing at least some of the well log data and the seismic signal data used to generate the 2D, 3D, and 4D seismic attributes to a same sampling rate, the well log data including core data, production data, and drilling data;
generating the one or more integrated enhanced logs from the first machine learning model;
grouping the one or more integrated enhanced logs into an ensemble of integrated enhanced logs to form a static reservoir 3D model of a subterranean reservoir;
inputting additional data to the first machine learning model to produce one or more updated integrated enhanced logs;
grouping the one or more updated integrated enhanced logs into an ensemble of updated integrated enhanced logs to form an updated 3D model; and
training a second machine learning model using a second neural network to generate a dynamic reservoir 3D model, wherein the updated 3D model and a set of dynamic modeling data are used as input for training the second machine learning model, wherein the dynamic modeling data includes data used to predict flow properties of the subterranean reservoir.

2. The method of claim 1 , wherein training the first machine learning model to generate the updated 3D model and training the second machine learning model to generate the dynamic reservoir 3D model occurs concurrently and in real time.

3. The method of claim 1 , further comprising:
applying seismic enhancement to the seismic signal data to provide enhanced seismic data, wherein the first machine learning model is based on the enhanced seismic data.

4. The method of claim 1 , wherein the additional data is real-time data.

5. The method of claim 1 , wherein the one or more integrated enhanced logs are machine learning generated logs of 2D properties of the subterranean reservoir.

6. One or more non-transitory machine-readable media comprising program code for generating one or more reservoir 3D models, the program code to:
train a first machine learning model to generate one or more integrated enhanced logs based, at least in part, on an integrated data set, wherein the integrated data set includes 2D, 3D, and 4D seismic attributes derived from a phase, a frequency, and an amplitude of seismic signal data, the 2D, 3D, and 4D seismic attributes matched to well log data by bringing at least some of the well log data and the seismic signal data used to generate the 2D, 3D, and 4D seismic attributes to a same sampling rate, the well log data including-that includes core data, production data, and drilling data;
generate the one or more integrated enhanced logs from the first machine learning model;
group the one or more integrated enhanced logs into an ensemble of integrated enhanced logs to form a static reservoir 3D model of a subterranean reservoir;
input additional data to the first machine learning model to produce one or more updated integrated enhanced logs; and
group the one or more updated integrated enhanced logs into an ensemble of updated integrated enhanced logs to form an updated 3D model; and
training a second machine learning model using a second neural network to generate a dynamic reservoir 3D model, wherein the updated 3D model and a set of dynamic modeling data are used as input for training the second machine learning model, wherein the dynamic modeling data includes data used to predict flow properties of the subterranean reservoir.

7. The one or more non-transitory machine-readable media of claim 6 , further comprising program code to:
train the first machine learning model to generate the updated 3D model and train the second machine learning model to generate the dynamic reservoir 3D model concurrently and in real time.

8. The one or more non-transitory machine-readable media of claim 6 , further comprising program code to:
apply seismic enhancement to the seismic signal data to provide enhanced seismic data, wherein the first machine learning model is based on the enhanced seismic data.

9. The one or more non-transitory machine-readable media of claim 6 , wherein the additional data is real-time data.

10. The one or more non-transitory machine-readable media of claim 6 , wherein the one or more integrated enhanced logs are machine learning generated logs of 2D properties of the subterranean reservoir.

11. An apparatus comprising:
a processor; and
a machine-readable medium having program code executable by the processor to cause the apparatus to,
train a first machine learning model to generate one or more integrated enhanced logs based, at least in part, on an integrated data set, wherein the integrated data set includes 2D, 3D, and 4D seismic attributes derived from a phase, a frequency, and an amplitude of seismic signal data, the 2D, 3D, and 4D seismic attributes matched to well log data by bringing at least some of the well log data and the seismic signal data used to generate the 2D, 3D, and 4D seismic attributes to a same sampling rate, the well log data including core data, production data, and drilling data;
generate the one or more integrated enhanced logs from the first machine learning model; group the one or more integrated enhanced logs into an ensemble of integrated enhanced logs to form a static reservoir 3D model of a subterranean reservoir;
input additional data to the first machine learning model to produce one or more updated integrated enhanced logs;
group the one or more updated integrated enhanced logs into an ensemble of updated integrated enhanced logs to form an updated 3D model; and
training a second machine learning model using a second neural network to generate a dynamic reservoir 3D model, wherein the updated 3D model and a set of dynamic modeling data are used as input for training the second machine learning model, wherein the dynamic modeling data includes data used to predict flow properties of the subterranean reservoir.

12. The apparatus of claim 11 , further comprising program code to:
train the first machine learning model to generate the updated 3D model and train the second machine learning model to generate the dynamic reservoir 3D model concurrently and in real time.

13. The apparatus of claim 12 , further comprising a user interface, wherein at least one of the static reservoir 3D model, the updated 3D model, and the dynamic reservoir 3D model is visualized via the user interface.

14. The apparatus of claim 11 , further comprising program code to:
apply seismic enhancement to the seismic signal data to provide enhanced seismic data, wherein the first machine learning model is based on the enhanced seismic data.

15. The apparatus of claim 11 , wherein the one or more integrated enhanced logs are machine learning generated logs of 2D properties of the subterranean reservoir.

Assignees

• Landmark Graphics Corp

Inventors

• Vallabhaneni Sridharan
• Roy Samiran
• Chaki Soumi
• Mandapaka Bhaskar Jogi Venkata
• Pakalapati Rajeev
• Srivastav Shreshth
• Priyadarshy Satyam

Classifications

• G01V1/302 Primary

  in 3D data cubes · CPC title

• G06F2113/08

  Fluids · CPC title

• G01V2210/614

  Synthetically generated data · CPC title

• G01V1/282

  Application of seismic models, synthetic seismograms · CPC title

• G01V1/44

  using generators and receivers in the same well (G01V1/52 takes precedence) · CPC title