Learning Expected Operational Behavior Of Machines From Generic Definitions And Past Behavior

• US 20150149134A1
• Filed: 11/27/2013
• Published: 05/28/2015
• Est. Priority Date: 11/27/2013
• Status: Active Grant

• Current Assignee: Falkonry Inc.
• Original Assignee: Falkonry Inc.
• Inventors: MEHTA, NIKUNJ R., BOSE, PRASANTA
• Granted Patent: US 10,409,926 B2
• Time in Patent Office: Days
• US Class Current: 703/6
• CPC Class Codes: G05B 23/0227 Qualitative history assessment, whereby the type of data acted upon, e.g. waveforms, images or patterns, is not relevant, e.g. rule based assessment; if-then decisions
• CPC Class Codes: G05B 23/0245 based on a qualitative model, e.g. rule based; if-then decisions
• CPC Class Codes: G06F 30/20 Design optimisation, verification or simulation optimisation, verification or simulation of circuit designs G06F30/30

Abstract

In an embodiment, a data processing method comprises storing one or more generic machine operating definitions, wherein each of the generic machine operating definitions describes expected operational behavior of one or more types of machines during one or more operating states; analyzing operating data that describes past operation of a plurality of machines of a plurality of types; based at least in part on the operating data and the one or more generic machine operating definitions, generating and storing one or more machine operating models that describe expected operational behavior corresponding to a plurality of operating states of the plurality of machines; wherein the one or more machine operating models comprise a plurality of data patterns, wherein each of the data patterns is associated with a different set of one or more operating states of one or more machines; wherein the method is performed by one or more computing devices.

Claims

1. A method comprising:

storing one or more generic machine operating definitions, wherein each of the generic machine operating definitions describes expected operational behavior of one or more types of machines during one or more operating states;

analyzing past operating data that describes past operation of a plurality of machines of a plurality of types;

based at least in part on the past operating data and the one or more generic machine operating definitions, generating and storing one or more machine operating models that describe expected operational behavior corresponding to a plurality of operating states of the plurality of machines;

wherein the one or more machine operating models comprise a plurality of data patterns, wherein each of the data patterns is associated with a different set of one or more operating states of one or more machines;

wherein the method is performed by one or more computing devices.

2. The method of claim 1, comprising indexing the one or more machine operating models based at least in part on frequencies and magnitudes of physical parameters defining the data patterns.

3. The method of claim 1, wherein the operating data comprises a plurality of machine operating measurements obtained from a plurality of sensors on the plurality of machines.

4. The method of claim 1, comprising detecting that current operational behavior of one or more particular machines of the plurality of machines matches a particular data pattern of the plurality of data patterns, wherein the particular data pattern is associated with a particular set of one or more operating states of the one or more particular machines;

causing displaying an indication that the one or more particular machines are in a particular set of one or more operating states.

5. The method of claim 1, comprising detecting that current operational behavior of one or more particular machines of the plurality of machines matches a particular data pattern that is associated with a particular set of one or more operating states of the one or more particular machines;

estimating a future set of one or more operating states of one or more of the plurality of machines.

6. The method of claim 1, comprising, using the one or more models, causing generating and displaying specifications of preventive action that is estimated to avoid reaching a possible future operating state that is damaging to the asset or remedial action that is estimated to change a current operating state.

7. The method of claim 1, wherein the machines are in a system of a plurality of systems;

comprising receiving and storing the operating data using a cloud storage service that stores data for the plurality of systems;

generating the one or more models using a cloud computing service that computes models for the plurality of systems; and

using the one or more generic definitions to generate different models for different systems of the plurality of systems.

8. The method of claim 1, wherein the one or more generic machine operating definitions comprise at least a first generic machine operating definition and a second generic machine operating definition, wherein the first generic machine operating definition describes expected operational behavior of one or more first types of machines during one or more first operating states, and wherein the second generic machine operating definition describes expected operational behavior of one or more second types of machines during one or more second operating states, wherein the one or more first types differ from the one or more second types by model or vendor.

9. The method of claim 1, comprising:

continuously receiving new operating data from the plurality of machines;

continuously adding old operating data to an updated collection of past machine operating data; and

continuously modifying the one or more machine operating models based on the updated collection of past machine operating data.

10. The method of claim 1, wherein each association between data patterns of the plurality of data patterns and the different sets of one or more operating states comprises an expected likelihood value, and wherein a particular data pattern of the plurality of data patterns is associated with a first set of one or more operating states with a first expected likelihood value, and wherein the particular data pattern is associated with a second set of one or more operating states with a second expected likelihood value, wherein the first set is different than the second set, and wherein the first expected likelihood value is different than the second expected likelihood value.

11. The method of claim 1, wherein a first data pattern of the plurality of data patterns is associated with a first set of one or more operating states of one or more particular machines, and wherein a second data pattern of the plurality of data patterns is associated with a second set of one or more operating states of the one or more particular machines;

wherein the first data pattern has a different temporal feature than the second data pattern; and

wherein the one or more machine operating models allow the expected operational behavior to describe the one or more particular machines as concurrently in both the first set of one or more operating states and the second set of one or more operating states.

12. The method of claim 1, wherein, for a machine, a first data pattern of the plurality of data patterns is defined based on a first subset of relevant parameters of a plurality of parameters related to the plurality of machines, and wherein a second data pattern of the plurality of data patterns is based on a second subset of relevant parameters of the plurality of parameters.

13. The method of claim 1, wherein generating one or more machine operating models is based at least in part on one or more physical dependencies between different machines of the plurality of machines;

comprising determining the one or more physical dependencies from a stored metamodel that describes relationships between different machines of the plurality of machines.

14. The method of claim 1, wherein the plurality of operating states include two or more of:

starting, running, shutting down, tripping, failing, a certain level of degradation, operating for a prolonged amount of time at some level of performance, or shutting down for a prolonged amount of time.

15. The method of claim 1, wherein a particular generic machine operating definition of the one or more generic machine operating definitions describes an expected output of one or more particular types of machines during a particular operating state.

16. The method of claim 1, comprising establishing a probabilistic association between a particular data pattern of the plurality of data patterns and a particular set of one or more operating states based at least in part on at least two conflicting sets of previously received label inputs that labeled the particular data pattern as the particular set of one or more operating states.

17. One or more non-transitory computer-readable storage media storing instructions which, when executed, cause:

storing one or more generic machine operating definitions, wherein each of the generic machine operating definitions describes expected operational behavior of one or more types of machines during one or more operating states;

analyzing past operating data that describes past operation of a plurality of machines of a plurality of types;

based at least in part on the past operating data and the one or more generic machine operating definitions, generating and storing one or more machine operating models that describe expected operational behavior corresponding to a plurality of operating states of the plurality of machines;

wherein the one or more machine operating models comprise a plurality of data patterns, wherein each of the data patterns is associated with a different set of one or more operating states of one or more machines.

18. The one or more non-transitory computer-readable storage media of claim 17, wherein the instruction, when executed, cause indexing the one or more machine operating models based at least in part on frequencies and magnitudes of physical parameters defining the data patterns.

19. The one or more non-transitory computer-readable storage media of claim 17, wherein the instruction, when executed, cause:

continuously receiving new operating data from the plurality of machines;

continuously adding old operating data to an updated collection of past machine operating data; and

continuously modifying the one or more machine operating models based on the updated collection of past machine operating data.

20. The one or more non-transitory computer-readable storage media of claim 17, wherein generating one or more machine operating models is based at least in part on one or more physical dependencies between different machines of the plurality of machines;

wherein the instructions, when executed, cause determining the one or more physical dependencies from a stored metamodel that describes relationships between different machines of the plurality of machines.