Computer system and method for defining and using a predictive model configured to predict asset failures

• US 10,754,721 B2
• Filed: 11/16/2018
• Issued: 08/25/2020
• Est. Priority Date: 12/01/2014
• Status: Active Grant

• Current Assignee: Uptake Technologies, Inc.
• Original Assignee: Uptake Technologies, Inc.
• Inventors: Horrell, Michael, Ciasulli, John, Zhong, Sheng, Kolb, Jason
• Primary Examiner(s): Duncan, Marc
• Assistant Examiner(s): Gibson, Jonathan D
• Application Number:
  US16/194,036
• Publication Number:
  US 20190087256A1
• Time in Patent Office: 648 Days
• CPC Class Codes: G01D 3/08 with provision for safeguarding the apparatus, e.g. against abnormal operation, against breakdown
• CPC Class Codes: G01M 99/005 Testing of complete machines, e.g. washing-machines or mobile phones testing of machine parts G01M13/00; testing of electric apparatus or components G01R31/50
• CPC Class Codes: G01M 99/008 by doing functionality tests
• CPC Class Codes: G05B 19/18 Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form G05B19/418 takes precedence
• CPC Class Codes: G05B 23/024 Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks
• CPC Class Codes: G05B 23/0254 based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks
• CPC Class Codes: G05B 23/0275 Fault isolation and identification, e.g. classify fault; estimate cause or root of failure
• CPC Class Codes: G06F 11/008 Reliability or availability analysis
• CPC Class Codes: G06F 11/0709 in a distributed system consisting of a plurality of standalone computer nodes, e.g. clusters, client-server systems
• CPC Class Codes: G06F 11/0721 within a central processing unit [CPU]
• CPC Class Codes: G06F 11/0751 Error or fault detection not based on redundancy power supply failures G06F1/30; network fault management H04L41/06
• CPC Class Codes: G06F 11/0754 by exceeding limits
• CPC Class Codes: G06F 11/0772 Means for error signaling, e.g. using interrupts, exception flags, dedicated error registers
• CPC Class Codes: G06F 11/0787 Storage of error reports, e.g. persistent data storage, storage using memory protection
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36
• CPC Class Codes: G06F 11/0793 Remedial or corrective actions recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40
• CPC Class Codes: G06F 11/2007 using redundant communication media
• CPC Class Codes: G06F 11/26 Functional testing
• CPC Class Codes: G06F 11/263 Generation of test inputs, e.g. test vectors, patterns or sequences ; with adaptation of the tested hardware for testability with external testers
• CPC Class Codes: G06F 2201/85 Active fault masking without idle spares active fault masking without idle spare hardware where processing functionality is redundant G06F11/2035
• CPC Class Codes: G06N 5/02 : Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/04 : Inference or reasoning models
• CPC Class Codes: G06N 7/01 : Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06Q 10/04 : Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/06312 : Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
• CPC Class Codes: G06Q 10/0633 : Workflow analysis
• CPC Class Codes: G06Q 10/067 : Enterprise or organisation modelling
• CPC Class Codes: G06Q 10/20 : Administration of product repair or maintenance
• CPC Class Codes: G06Q 50/04 : Manufacturing
• CPC Class Codes: G06Q 50/08 : Construction
• CPC Class Codes: G07C 5/008 : communicating information to a remotely located station transmission systems for measured values G08C
• CPC Class Codes: G07C 5/0808 : Diagnosing performance data testing of vehicles G01M17/00; testing of electrical installation on vehicles G01R31/005
• CPC Class Codes: G07C 5/0825 : using optical means
• CPC Class Codes: G08B 21/18 : Status alarms G08B21/02 takes precedence
• CPC Class Codes: H04L 45/22 : Alternate routing
• CPC Class Codes: Y02P 90/30 : Computing systems specially adapted for manufacturing

Abstract

Disclosed herein are systems, devices, and methods related to assets and asset operating conditions. In particular, examples involve defining and using a predictive model that is configured to output an indication of whether at least one failure type from the group of possible failure types is likely to occur at an asset within the given period of time in the future.

Claims

1. A computing system comprising:

a network interface configured to facilitate communication with a plurality of assets and a plurality of computing devices;

at least one processor;

a non-transitory computer-readable medium; and

program instructions stored on the non-transitory computer-readable medium that are executable by the at least one processor to cause the computing system to;

identify a group of abnormal-condition types associated with a group of possible failure types for assets of a given type;

based on the identified group of abnormal-condition types, identify a subset of historical operating data comprising (i) historical abnormal-condition data for a plurality of assets of the given type that indicates past occurrences of the identified group of abnormal-condition types at the plurality of assets and (ii) historical sensor data for the plurality of assets that indicates sensor measurements associated with the past occurrences of the identified group of abnormal-condition types at the plurality of assets;

apply a supervised machine learning technique to the identified subset of historical operating data to define a predictive model that is configured to (i) receive sensor data for an asset as input, (ii) for each of at least two failure types from the group of possible failure types, make a respective prediction of whether the failure type is likely to occur at the asset within a given period of time in the future, and (iii) based on the respective predictions, output an indication of whether at least one failure type from the group of possible failure types is likely to occur at the asset within the given period of time in the future;

receive sensor data indicating operating conditions of a given asset;

apply the predictive model to the received sensor data and thereby determine, for the given asset, an indication of whether at least one failure type from the group of possible failure types is likely to occur at the given asset within the given period of time in the future;

compare the indication for the given asset to threshold criteria and thereby make a determination that the indication satisfies the threshold criteria; and

responsive to the determination that the indication satisfies the threshold criteria, carry out a remedial action that comprises at least one of (i) automatically generating and sending, to a computing device associated with an individual responsible for overseeing the given asset, an alert, (ii) automatically generating and sending, to the given asset, an instruction for the given asset to modify its operation to account for the determination that the indication satisfies the threshold criteria, (iii) automatically generating and sending, to a repair facility, an instruction to repair the given asset, or (iv) automatically generating and sending, to a parts-ordering system, an instruction for the parts ordering system to order a given component of the given asset.

2. The computing system of claim 1, wherein the program instructions that are executable by the at least one processor to cause the computing system to identify the group of abnormal-condition types associated with the group of possible failure types for assets of the given type comprise program instructions stored on the non-transitory computer-readable medium that are executable by the at least one processor to cause the computing system to:

identify one or more sensors associated with assets of the given type; and

identify one or more abnormal-condition types corresponding to the one or more sensors.

3. The computing system of claim 2, wherein identifying the one or more sensors associated with assets of the given type comprises identifying the one or more sensors associated with assets of the given type based on historical operating data and at least one of historical repair data or sensor attributes.

4. The computing system of claim 1, wherein the group of possible failure types comprises one or more failure types that could render an asset of the given type inoperable when the one or more failure types occur.

5. The computing system of claim 1, wherein each failure type from the group of possible failure types corresponds to at least one abnormal-condition type from the identified group of abnormal-condition types.

6. The computing system of claim 1, wherein the program instructions that are executable by the at least one processor to cause the computing system to identify the group of abnormal-condition types associated with the group of possible failure types for assets of the given type comprise program instructions stored on the non-transitory computer-readable medium that are executable by the at least one processor to cause the computing system to identify the group of abnormal-condition types based on user input.

7. The computing system of claim 1, wherein the indication for the given asset comprises a single, aggregated metric that reflects whether at least one failure type from the group of possible failure types is likely to occur at the given asset within the given period of time in the future.

8. The computing system of claim 7, wherein the single, aggregated metric comprises one of (i) a probability that no failure type from the group of possible failure types will occur at the given asset within the given period of time in the future or (ii) a probability that at least one failure type from the group of possible failure types will occur at the given asset within the given period of time in the future.

9. The computing system of claim 1, wherein the indication for the given asset comprises a set of probability metrics that each reflect a probability that a respective failure type from the group of possible failure types will occur at the given asset within the given period of time in the future.

10. The computing system of claim 1, wherein the threshold criteria comprises a maximum probability of a failure type from the group of possible failure types occurring at the given asset within the given period of time in the future.

11. A non-transitory computer-readable medium having instructions stored thereon that are executable to cause a computing system to:

identify a group of abnormal-condition types associated with a group of possible failure types for assets of a given type;

based on the identified group of abnormal-condition types, identify a subset of historical operating data comprising (i) historical abnormal-condition data for a plurality of assets of the given type that indicates past occurrences of the identified group of abnormal-condition types at the plurality of assets and (ii) historical sensor data for the plurality of assets that indicates sensor measurements associated with the past occurrences of the identified group of abnormal-condition types at the plurality of assets;

apply a supervised machine learning technique to the identified subset of the historical operating data to define a predictive model that is configured to (i) receive sensor data for an asset as input, (ii) for each of at least two failure types from the group of possible failure types, make a respective prediction of whether the failure type is likely to occur at the asset within a given period of time in the future, and (iii) based on the respective predictions, output an indication of whether at least one failure type from the group of possible failure types is likely to occur at the asset within the given period of time in the future;

receive sensor data indicating operating conditions of a given asset;

apply the predictive model to the received sensor data and thereby determine, for the given asset, an indication of whether at least one failure type from the group of possible failure types is likely to occur at the given asset within the given period of time in the future;

compare the indication for the given asset to threshold criteria and thereby make a determination that the indication satisfies the threshold criteria; and

responsive to the determination that the indication satisfies the threshold criteria, carry out a remedial action that comprises at least one of (i) automatically generating and sending, to a computing device associated with an individual responsible for overseeing the given asset, an alert, (ii) automatically generating and sending, to the given asset, an instruction for the given asset to modify its operation to account for the determination that the indication satisfies the threshold criteria, (iii) automatically generating and sending, to a repair facility, an instruction to repair the given asset, or (iv) automatically generating and sending, to a parts-ordering system, an instruction for the parts ordering system to order a given component the given asset.

12. The non-transitory computer-readable medium of claim 11, wherein the instructions that are executable to cause the computing system to identify the group of abnormal-condition types associated with the group of possible failure types for assets of the given type comprise instructions stored on the non-transitory computer-readable medium that are eecutable to cause the computing system to:

identify one or more sensors associated with assets of the given type; and

identify one or more abnormal-condition types corresponding to the one or more sensors.

13. The non-transitory computer-readable medium of claim 12, wherein identifying the one or more sensors associated with assets of the given type comprises identifying the one or more sensors associated with assets of the given type based on historical operating data and at least one of historical repair data or sensor attributes.

14. The non-transitory computer-readable medium of claim 11, wherein each failure type from the group of possible failure types corresponds to at least one abnormal-condition type from the identified group of abnormal-condition types.

15. The non-transitory computer-readable medium of claim 11, wherein the instructions that are executable to cause the computing system to identify the group of abnormal-condition types associated with the group of possible failure types for assets of the given type comprise instructions stored on the non-transitory computer-readable medium that are executable to cause the computing system to identify the group of abnormal-condition types based on user input.

16. The non-transitory computer-readable medium of claim 11, wherein the indication for the given asset comprises one of (a) a single, aggregated metric that reflects whether at least one failure type from the group of possible failure types is likely to occur at the given asset within the given period of time in the future or (b) a set of probability metrics that each reflect a probability that a respective failure type from the group of possible failure types will occur at the given asset within the given period of time in the future.

17. A computer-implemented method, the method comprising:

identifying a group of abnormal-condition types associated with a group of possible failure types for assets of a given type;

based on the identified group of abnormal-condition types, identifying a subset of historical operating data comprising (i) historical abnormal-condition data for a plurality of assets of the given type that indicates past occurrences of the identified group of abnormal-condition types at the plurality of assets and (ii) historical sensor data for the plurality of assets that indicates sensor measurements associated with the past occurrences of the identified group of abnormal-condition types at the plurality of assets;

applying a supervised machine learning technique to the identified subset of the historical operating data to define a predictive model that is configured to (i) receive sensor data for an asset as input, (ii) for each of at least two failure types from the group of possible failure types, make a respective prediction of whether the failure type is likely to occur at the asset within a given period of time in the future, and (iii) based on the respective predictions, output an indication of whether at least one failure type from the group of possible failure types is likely to occur at the given asset within the given period of time in the future;

receiving sensor data indicating operating conditions of a given asset;

apply the predictive model to the received sensor data and thereby determine, for the given asset, an indication of whether at least one failure type from the group of possible failure types is likely to occur at the given asset within the given period of time in the future;

comparing the indication for the given asset to threshold criteria and thereby making a determination that the indication satisfies the threshold criteria; and

responsive to the determination that the indication satisfies the threshold criteria, carrying out a remedial action that comprises at least one of (i) automatically generating and sending, to a computing device associated with an individual responsible for overseeing the given asset, an alert, (ii) automatically generating and sending, to the given asset, an instruction for the given asset to modify its operation to account for the determination that the indication satisfies the threshold criteria, (iii) automatically generating and sending, to a repair facility, an instruction to repair the given asset, or (iv) automatically generating and sending, to a parts-ordering system, an instruction for the parts ordering system to order a given component of the given asset.

18. The computer-implemented method of claim 17, wherein identifying the group of abnormal-condition type associated with the group of possible failure types for assets of the given type comprises:

identifying one or more sensors associated with assets of the given type; and

identifying one or more abnormal-condition types corresponding to the one or more sensors.

19. The computer-implemented method of claim 17, wherein identifying the group of abnormal-condition types associated with the group of possible failure types for assets of the given type comprises identifying the group of abnormal-condition types based on user input.

20. The computer-implemented method of claim 17, wherein the indication for the given asset comprises one of (a) a single, aggregated metric that reflects whether at least one failure type from the group of possible failure types is likely to occur at the given asset within the given period of time in the future or (b) a set of probability metrics that each reflect a probability that a respective failure type from the group of possible failure types will occur at the given asset within the given period of time in the future.