SYSTEM AND METHOD FOR AUTOMATING EXPLORATION OR PRODUCTION OF SUBTERRANEAN RESOURCE

• US 20110155462A1
• Filed: 07/23/2009
• Published: 06/30/2011
• Est. Priority Date: 07/23/2008
• Status: Active Grant

• Current Assignee: Schlumberger Technology Corporation (Schlumberger NV)
• Original Assignee: Schlumberger Technology Corporation (Schlumberger NV)
• Inventors: Barrow, Harry, Du Castel, Bertrand
• Granted Patent: US 8,606,734 B2
• Time in Patent Office: Days
• US Class Current: 175/24
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

A method and system for automating exploration and production of subterranean resources, using data collected from exploration and production activities, including drilling rig data and/or the like is described. The method and system include the use of a knowledge representation containing representation of uncertainty in the exploration and production operations.

Claims

1. A method for implementation of automated systems that uses an explicit representation of both general and specific knowledge and experience of some aspect of oilfield operations, together with software that accesses the representation and uses it to construct one or more software components of the automated system.

3. A method for performing oilfield operations that incorporates both claims 1 and 2.

2. A method for performing oilfield operations that uses an explicit representation of both general and specific knowledge and experience of some aspect of oilfield operations, in which uncertainty is explicitly represented.

4. A method for operating a computer to use knowledge and experiences in a petroleum exploration/recovery/processing/conveyance/extraction/enhanced recovery domain to automate a task associated with a petroleum operation, comprising:

representing knowledge and experience of the domain in an ontology including representations of possible states of the petroleum exploration operation, associating probabilities with a first subset of the states, and associating conditional probabilities mapping a second subset of the states to a third subset of states;

extracting a Bayesian belief network from the first, second and third subsets of states of the ontology; and

inputting data obtained from the petroleum exploration operation into the Bayesian belief network to provide an interpretation of the input data.

5. The method for operating a computer to use knowledge and experiences in a petroleum exploration/recovery/processing domain to automate a task associated with a petroleum exploration operation of claim 4, wherein the ontology is organized as classes and individual members of those classes and the step of representing knowledge and experience of the domain in an ontology further comprises:

representing states in the first subset as individuals of a first class in the ontology and wherein each individual member of the first class has a value from a set of values and assigning each value a probability associated with the likelihood the individual member has that particular value.

6. The method for operating a computer to use knowledge and experiences in a petroleum exploration/recovery/processing domain to automate a task associated with a petroleum exploration operation of claim 4, wherein the ontology is organized as classes and individual members of those classes and the step of representing knowledge and experience of the domain in an ontology further comprises:

representing states in the second subset as individuals of a second class in the ontology and wherein each individual member of the second class has a value from a set of states and associating with each such state a probability associated with the likelihood such state is indicative of a particular state in the third subset of states.

7. A method for operating petroleum exploration and production equipment comprising:

retrieving information organized in an ontology describing relationships and states involving the operation of the equipment;

translating the retrieved information into a functional program; and

executing on a computer according the instructions of the functional program to receive sensor data obtained from the equipment into the functional program; and

process the sensor data to perform a task related to the operating of the equipment.

8. The method of claim 7, wherein the ontology is a representation of petroleum exploration and production knowledge and experiences are related to the equipment and is organized hierarchically in classes and sub-classes including classes for actions that may be performed by the equipment, states that the equipment may be in, and events that may occur in the operation of the equipment.

9. The method of claim 8, wherein the ontology contains classes corresponding to particular operations of the equipment and contains descriptions of probabilities for transitioning from one class corresponding to a first particular operation to a class corresponding to a second particular operation.

10. The method of claim 9, wherein the retrieved information further comprises the descriptions of probabilities for transitioning from one operation to another operation and the step of translating the retrieved information into a functional program comprises producing a finite state machine having states corresponding to particular operations and transitions corresponding to possible transitions between the particular operations.

11. The method of claim 10, wherein the ontology further comprises descriptions of operating environments in which the equipment may be deployed and wherein the probabilities related to the transitioning from one operation to another operation is dependent on the environment in which the equipment is deployed.

12. The method of claim 10, wherein the step of translating further comprises constructing a probability matrix containing, for each particular operation the probability of a transition to any other possible operation.

13. The method of claim 10, wherein the step of translating comprises:

constructing functional program code to cause a computer executing the functional program to spawn a thread for each possible operation given observed sensor data, and kill any currently active thread corresponding to an operation not possible in light of observed sensor data.

14. The method of claim 10, wherein the step of translating comprises:

constructing functional program code to cause a computer executing the functional program to create an object corresponding to each possible operation given observed sensor data, and delete any currently active object corresponding to an operation not possible in light of observed sensor data.

15. The method of claim 8, wherein the ontology contains classes corresponding to particular operations of the equipment and contains descriptions of probabilities for particular events occurring given the equipment being in a state of performing a particular operation.

16. The method of claim 15, wherein the retrieved information further comprises the descriptions of probabilities for particular events occurring given the equipment being in a state of performing a particular operation and the step of translating the retrieved information into a functional program comprises producing a finite state machine having states corresponding to particular operations and transitions corresponding to the particular events.

17. The method of claim 7, wherein the step of translating further comprises constructing a probability matrix containing, for each particular operation the probability of any particular observation.

18. The method of claim 7 wherein the step of retrieving information comprises retrieving probabilities described in the ontology and the step of translating comprises using the retrieved probabilities to produce a Bayesian belief network.

19. The method of claim 8 wherein the equipment is well-drilling equipment, and wherein the functional program is a program for interpreting the sensor data to determine the activities occurring in a drilling operation.

20. A program generator for producing a functional program for use in controlling an oilfield machine having a device for performing an oilfield operation, a sensor for collecting data related to the oilfield operation, a control computer connected to the sensor and having a storage medium, the program generator comprising:

a computer comprising;

a storage device containing instructions to direct the operation of the computer, wherein the instructions include;

an ontology containing classes describing actions of the device, events effecting the device, and possible states of the device and classes describing uncertainties in transitions between states of the device and of events occurring in given states of the device;

program generation instructions operable to produce a functional program to be executed by the control computer comprising instructions to cause the control computer to receive sensor data from the device and to perform a control task, the program generation instructions comprising instructions to cause the computer to retrieve a description of the device including actions the device may perform, events effecting the device, and possible states of the device from the ontology and to use the retrieved description of the device from the ontology to create a functional program that receives sensor data and produces a result.

21. The program generator of claim 20, wherein the produced result is an interpretation of observed sensor data in terms of actions performed by the device.

22. The program generator of claim 20, wherein the produced result is a control signal to an actuator controlling the operation of the device.

23. A system for automating drilling comprising:

drilling machinery controlled by an actuator;

a sensor for collecting data related to a drilling operation performed by the drilling machinery;

a first computer comprising;

a storage medium containing therein an ontology containing a description of the drilling machinery including classes describing actions of the drilling machinery, events effecting the drilling machinery, and possible states of the drilling machinery;

program generation instructions operable to produce a functional program to be executed by a control computer comprising instructions to cause the control computer to receive sensor data and to perform a control task, the program generation instructions comprising instructions to cause the first computer to;

retrieve a description of the drilling machinery including actions the drilling machinery may perform, events effecting the drilling machinery, and possible states of the drilling machinery from the ontology; and

use the retrieved description of the drilling machinery from the ontology to create a functional program that receives sensor data and produces a signal to control the actuator;

a second computer connected to the sensor and the actuator and having a storage medium into which the functional program receives sensor data and produces a signal to control the actuator is loaded thereby causing the second computer when executing the functional program to be operable to receive the sensor data, process the sensor data using the functional program and produce a signal to control the drilling machinery by causing the actuator to perform an action.

24. A method for automating exploration or production of subterranean resources, comprising:

representing a first body of knowledge relevant to exploration or production of subterranean resources in an ontology;

representing a second body of knowledge relevant to exploration or production of subterranean resources in the ontology;

producing a first interpretation product from the first body of knowledge the first interpretation product consisting of a first set of indexed values;

producing a second interpretation product from the second body of knowledge the second interpretation product consisting of a second set of indexed values;

correlating the first and second interpretation products to produce rules reflecting how the first set of indexed values effect the second set of indexed values; and

using the rules to automate at least one task associated with exploration or production of subterranean resources.

25. The method of claim 24, wherein the first body of knowledge is a Bayesian Belief Network containing conditional probabilities linking exploration or production problems to observable physical quantities, and wherein the second body of knowledge is a stochastic grammar having rules useful to interpret a data stream in terms of activities performed during an exploration or production operation.

26. The method of claim 25, wherein:

the first interpretation product is produced by a first interpretation program derived from a representation of the Bayesian Belief Network useable by a Bayesian Belief Network interpreter from the ontology representation of the Bayesian Belief Network; and

the second interpretation product is produced by a second interpretation program derived from an interpretation program for exploration or production equipment data from the stochastic grammar wherein the interpretation program computes first probabilities of states reflecting activities from transitions from states and second probabilities of states reflecting activities from the equipment data, and combining the combining the first probabilities with the second probabilities to produce the second interpretation results.

27. The method of claim 24, wherein the at least one automated task is an operational task in operation of a subterranean exploration or production equipment.

28. The method of claim 24, wherein the at least one automated task is an automated training operation for the operation of a subterranean exploration or production equipment.

29. The method of claim 24, wherein the subterranean exploration or production is drilling for petroleum resources, wherein the first body of knowledge is a Bayesian Belief Network linking drilling problems to observable physical conditions, and wherein the second body of knowledge is a stochastic grammar useful to interpret a drilling states data channel as activities of a driller;

wherein the correlation is a correlation of probabilities assigned to various drilling problems to probabilities of activities of the driller, thereby producing rules indicative of how a driller reacts to particular drilling problems given particular conditions.

30. The method of claim 29 further comprises:

receiving a data stream;

interpreting the data stream to compute probabilities of drilling problems;

applying the rules indicative of how a driller reacts to the interpretation of the data stream to derive a suggested action.

31. The method of claim 30, further comprising:

suggesting the suggested action to a driller.

32. The method of claim 30 further comprising:

using the suggested action to control the operation of drilling equipment.

33. The method of claim 30 further comprising:

comparing the suggested action to an action suggested by a student driller and using the comparison to provide feedback to the student.