System and method for delineating spatially dependent objects, such as hydrocarbon accumulations from seismic data

• US 6,236,942 B1
• Filed: 08/30/1999
• Issued: 05/22/2001
• Est. Priority Date: 09/15/1998
• Status: Expired due to Term

• Current Assignee: Bush Seismic Technologies, LLC
• Original Assignee: Scientific Predication Inc.
• Inventors: Bush, Ronald R.
• Primary Examiner(s): McElheny, Jr., Donald E.
• Application Number:
  US09/385,345
• Time in Patent Office: 631 Days
• Field of Search: 702/11, 702/12, 702/13, 702/14, 703/10, 703/5, 706/928, 706/929
• US Class Current: 702/14
• CPC Class Codes: G01V 1/30 Analysis G01V1/50 takes precedence
• CPC Class Codes: G01V 3/38 Processing data, e.g. for analysis, for interpretation, for correction
• CPC Class Codes: Y10S 706/929 Geological, e.g. seismology

Abstract

A neural network based system, method, and process for the automated delineation of spatially dependent objects is disclosed. The method is applicable to objects such as hydrocarbon accumulations, aeromagnetic profiles, astronomical clusters, weather clusters, objects from radar, sonar, seismic and infrared returns, etc. One of the novelties in the present invention is that the method can be utilized whether or not known data is available to provide traditional training sets. The output consists of a classification of the input data into clearly delineated accumulations, clusters, objects, etc. that have various types and properties. A preferred but non-exclusive application of the present invention is the automated delineation of hydrocarbon accumulations and sub-regions within the accumulations with various properties, in an oil and gas field, prior to the commencement of drilling operations.

Claims

1. A method for the automated prediction of locations of hydrocarbon producing areas and non-producing areas directly from seismic data gathered in an area comprising the steps of:

developing a neural network using seismic training data relating to one or more hydrocarbon producing areas and seismic training data relating to one or more hydrocarbon non-producing areas; and

applying the neural network to at least a portion of the seismic data to generate predictions of locations of hydrocarbon producing areas and hydrocarbon non-producing areas of the area.

2. The method of claim 1, further comprising the steps of:

developing the neural network to distinguish sub-regions within hydrocarbon producing areas; and

applying the neural network to at least a portion of the seismic data to distinguish sub-regions within the hydrocarbon producing areas.

3. The method of claim 2, wherein one of the sub-regions distinguished is a gas cap.

4. The method of claim 1, wherein the automated prediction of locations is performed in real-time as the seismic data is gathered.

5. The method of claim 4, wherein the seismic data is gathered using Vibroseis.

6. A method of delineating hydrocarbon accumulations from seismic data gathered in an area comprising the steps of:

developing a neural network within a conceptual sliding window to distinguish hydrocarbon producing areas and hydrocarbon non-producing areas, wherein the neural network is developed without using historical data; and

applying the neural network to at least a portion of the seismic data to distinguish hydrocarbon producing areas and hydrocarbon non-producing areas.

7. The method of claim 6, wherein the conceptual sliding window is provided with an “

In”

portion and an “

Out”

portion.

8. The method of claim 7, wherein the step of developing a neural network further comprises the steps of:

associating a first portion of the seismic data with the “

Out”

portion of the sliding window, wherein the first portion of the seismic data is assumed to be from a hydrocarbon non-producing area;

associating a second portion of the seismic data with the “

In”

portion of the sliding window, wherein the second portion of the seismic data is assumed to be from a hydrocarbon producing area;

using the associated data as inputs to the neural network;

training and testing the neural network using the first and second portions of the seismic data; and

determining whether the assumptions about the first and second portions of the seismic data were accurate, and if not, repeating the above steps using different portions of the seismic data.

9. The method of claim 8, wherein the step of determining whether the assumptions about the first and second portions of the seismic data were accurate further comprises the step of calculating a variance.

10. The method of claim 6, wherein the hydrocarbon producing areas and hydrocarbon non-producing areas are distinguished in real-time as the seismic data is gathered.

11. The method of claim 10, wherein the seismic data is gathered using Vibroseis.

12. A method of predicting locations of mineral producing areas and non-producing areas from data relating to a given area comprising the steps of:

developing a neural network to distinguish mineral producing areas and mineral non-producing areas; and

applying the neural network to at least a portion of the data to generate predictions of locations of mineral producing areas and non-producing areas in the given area.

13. The method of claim 12, wherein the data is aeromagnetic data.

14. The method of claim 12, wherein the data is seismic data.

15. A method of producing hydrocarbon products from an oil and/or gas field comprising the steps of:

gathering seismic data in the Oil and/or gas field;

developing a neural network using seismic training data relating to one or more hydrocarbon producing areas and seismic training data relating to one or more hydrocarbon non-producing areas;

applying the neural network to at least a portion of the seismic data to generate predictions of locations of hydrocarbon producing areas and hydrocarbon non-producing areas of the oil and/or gas field; and

extracting hydrocarbons from the oil and/or gas field in one or more predicted locations.

16. An automated method of predicting locations of hydrocarbon producing areas and non-producing areas in a field directly from seismic data acquired in the field comprising the steps of:

providing a neural network;

training the neural network to recognize hydrocarbon producing areas and non-producing areas using a first and second training sets of seismic data, wherein the first training set of seismic data corresponds to a hydrocarbon producing area and the second training set of seismic data corresponds to a hydrocarbon non-producing area;

applying the neural network to at least a portion of the seismic data to identify locations of hydrocarbon producing areas and non-producing areas in the field.

17. The method of claim 16, wherein data in the first and second training sets is comprised of historical data.

18. The method of claim 16, wherein the first training set of seismic data is comprised of seismic data taken from an area known to be a hydrocarbon producing area.

19. The method of claim 18, wherein the second training set of seismic data is comprised of seismic data taken from an area known to be a hydrocarbon non-producing area.

20. The method of claim 16, wherein data in the first training set is comprised of seismic data taken from an area assumed to be in a hydrocarbon producing area.

21. The method of claim 20, wherein data in the second training set is comprised of seismic data taken from an area assumed to be in a hydrocarbon non-producing area.

22. The method of claim 16, wherein the neural network is trained using the steps of:

providing a conceptual sliding window having a first portion and a second portion;

associating the first portion of the conceptual sliding window with a first area in the field;

selecting the first training set from seismic data in the first area;

assigning a first classification to the data in the first training set, wherein the first classification represents a hydrocarbon producing area;

associating the second portion of the conceptual sliding window with a second area in the field;

selecting the second training set from seismic data in the second area;

assigning a second classification to the data in the second training set, wherein the second classification represents a hydrocarbon non-producing area; and

training the neural network using the first and second training sets without using historical data.

23. The method of claim 22, further comprising the step of determining whether the neural network is adequately trained prior to applying the neural network to the seismic data.

24. The method of claim 23, wherein the determination of whether the neural network is adequately trained is accomplished by monitoring one or more variances as the neural network converges.

25. The method of claim 23, wherein if it is determined that the neural network is not adequately trained, the step of training the neural network further comprises the steps of:

associating the first portion of the conceptual sliding window with a third area in the field;

selecting the first training set from seismic data in the third area;

assigning the first classification to the data in the first training set;

associating the second portion of the conceptual sliding window with a fourth area in the field;

selecting the second training set from seismic data in the fourth area;

assigning the second classification to the data in the second training set; and

training the neural network using the first and second training sets.

26. The method of claim 25, further comprising the steps of determining whether the neural network is adequately trained prior to applying the neural network to the seismic data and repeating the training steps while associating the first and second portions of the conceptual sliding window with different areas.

27. The method of claim 16, further comprising the steps of:

providing a second neural network;

training the second neural network using training data selected from output data of the initial neural network;

applying the second neural network to at least a portion of the seismic data; and

determining the accuracy of the initial network by comparing the results of the initial neural network and the second neural network.

28. A method for predicting production levels of a hydrocarbon reservoir in a hydrocarbon field comprising the steps of:

gathering seismic data in the field;

gathering wellbore production data from the at least one wellbore;

developing a neural network using at least a portion of the seismic data and at least a portion of the wellbore production data as training sets; and

generating predictions of hydrocarbon production levels at contemplated well sites using the trained neural network.

29. The method of claim 28, wherein the wellbore production data includes data gathered from a plurality of wellbores.

30. The method of claim 28, further comprising the step of gathering bottom hole pressure data from at least one wellbore in the reservoir.

31. The method of claim 28, wherein the step of generating predictions further comprises the step of applying the trained neural network to at least a portion of the seismic data and to at least a portion of the wellbore production data.