CORRELATION AND ANNOTATION OF TIME SERIES DATA SEQUENCES TO EXTRACTED OR EXISTING DISCRETE DATA

• US 20160239756A1
• Filed: 10/10/2013
• Published: 08/18/2016
• Est. Priority Date: 10/10/2013
• Status: Abandoned Application

• Current Assignee: GE Intelligent Platforms Incorporated (Emerson Electric Company)
• Original Assignee: GE Intelligent Platforms Incorporated (Emerson Electric Company)
• Inventors: AGGOUR, Kareem Sherif, BOWMAN, Ward Linnscott, COURTNEY, Brian Scott, MATHUR, Sunil, WILLIAMS, Jenny Marie Weisenberg
• Application Number:
  US15/028,044
• Publication Number:
  US 20160239756A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G05B 23/0221 Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
• CPC Class Codes: G05B 23/0229 knowledge based, e.g. expert systems; genetic algorithms
• CPC Class Codes: G05B 23/0283 Predictive maintenance, e.g. involving the monitoring of a system and, based on the monitoring results, taking decisions on the maintenance schedule of the monitored system; Estimating remaining useful life [RUL] preventive maintenance, i.e. planning maintenance according to the available resources without monitoring the system G06Q10/06
• CPC Class Codes: G06F 16/24568 Data stream processing; Continuous queries
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: H04L 41/0654 using network fault recovery ring fault isolation or reconfiguration in loop networks without recovery actions by a network management system H04L12/437
• CPC Class Codes: H04L 41/142 using statistical or mathematical methods
• CPC Class Codes: H04L 41/147 for predicting network behaviour
• CPC Class Codes: H04L 41/16 using machine learning or artificial intelligence
• CPC Class Codes: Y04S 40/00 Systems for electrical power generation, transmission, distribution or end-user application management characterised by the use of communication or information technologies, or communication or information technology specific aspects supporting them

Abstract

A system for predicting events by associating time series data with other types of non-time series data can include a processor configured to receive a data stream including time series data transmitted from a sensor configured to measure an operating parameter of a component being monitored. The processor identifies sequences of interest in the time series data having predictive value. The processor compares the real-time data stream to a set of known historical patterns that act as effective leading indicators of different alarms and events. The processor extracts any identified sequences of interest from the time series data as an extracted event. The processor quantifies the relationship between the data of the extracted event and the known historical pattern by calculating a confidence level to denote a probability of occurrence of the event by comparing how closely the new time series data matches the data patterns associated with known events.

Claims

1. A method for associating time series data to template time series data patterns to detect current or predict future events, the method comprising:

receiving at a processor a data stream transmitted from a sensor configured to measure an operating parameter of a component being monitored, wherein the data stream comprises at least time series data;

analyzing the data stream to identify a sequence of interest in the time series data having predictive values by matching patterns in the time series data to a template pattern of a known event;

extracting the identified sequence of interest from the time series data;

marking the identified sequence of interest as an extracted event;

storing the extracted event in a database to indicate a possible detection or prediction of an occurrence of an event;

specifying a relationship between the time series data of the extracted event and the template pattern of the known event in quantifiable terms; and

quantifying the relationship between the time series data of the extracted event and the known event by calculating a confidence level to denote a statistical probability of occurrence of the event by comparing data patterns of the extracted event with data of the template pattern associated with the known event.

2. The method of claim 1, wherein the occurrence of an event indicates an alarm event.

3. The method of claim 1, wherein the occurrence of an event indicates a failure event.

4. The method of claim 1, wherein the template pattern of the known event represents data associated with an alarm event.

5. The method of claim 1, wherein the template pattern of the known event represents data associated with a failure event.

6. The method of claim 1, further comprising acquiring metadata from content of the time series data of the extracted event, wherein the metadata contains information explaining details of the content of the time series data related to the extracted event;

and

storing in the database the metadata associated with the extracted event.

7. The method of claim 6, further comprising:

performing a predictive analysis on the data stream to determine operational problems of the component and to identify the sequence of interest to predict future events associated with the component based on the extracted event data and the metadata stored in the database.

8. The method of claim 1, further comprising a predictive system configured to take preventative actions on the component when the probability of occurrence indicates a failure event has occurred.

9. The method of claim 1, further comprising a predictive system configured to take preventative actions on the component when the probability of occurrence indicates an alarm event has occurred.

10. The method of claim 1, wherein the sequence of interest identifies at least one of a situation that needs preventative actions to avoid an actual failure in the future.

11. The method of claim 1, further comprising taking preventative action on a failing component based on the confidence level generated by a predictive system.

12. A system for associating time series data to template time series data patterns to detect current predict future events, the system comprising:

at least one processing unit and at least one database;

a plurality of sensors in communication with the at least one processing unit;

wherein the at least one processing unit is configured to;

receive at a processor a data stream transmitted from a sensor configured to measure an operating parameter of a component being monitored, wherein the data stream comprises at least time series data;

analyze the data stream to identify a sequence of interest in the time series data having predictive values by matching patterns in the time series data to a template pattern of a known event;

extract the identified sequence of interest from the time series data;

mark the identified sequence of interest as an extracted event;

store the extracted event in a database to indicate a possible detection or prediction of an occurrence of an event;

specify a relationship between the time series data of the extracted event and the template pattern of the known event in quantifiable terms; and

quantify the relationship between the time series data of the extracted event and the known event by calculating a confidence level to denote a statistical probability of occurrence of the event by comparing data patterns of the extracted event with data of the template pattern associated with the known event.

13. The system of claim 12, wherein the occurrence of an event indicates an alarm event.

14. The system of claim 12, wherein the occurrence of an event indicates a failure event.

15. The system of claim 12, wherein the processing unit is configured to acquire metadata from content of the time series data of the extracted event, wherein the metadata contains information explaining details of the content of the time series data related to the extracted event;

and

store in a database the metadata associated with the extracted event.

16. The system of claim 12, wherein the processing unit is configured to perform a predictive analysis on the data stream to determine operational problems of the component and to identify the sequence of interest to predict future events associated with the component based on the extracted event data and the metadata stored in the database.

17. The system of claim 12, further comprising a predictive system configured to take preventative actions on the component when the probability of occurrence indicates a failure event has occurred.

18. The system of claim 12, further comprising a predictive system configured to take preventative actions on the component when the probability of occurrence indicates an alarm event has occurred.

19. A method for associating time series data to pre-existing discrete data to predict future events, the method comprising:

receiving at a processor a data stream transmitted from a sensor configured to measure an operating parameter of a component being monitored, wherein the data stream comprises at least time series data;

correlating relevant time series data to pre-existing event data to detect extracted time series events;

identifying each occurrence of the relevant time series data in the data stream, extracting the identified relevant time series data and marking the relevant time series data as an extracted time series event;

inspecting event data that chronologically follows the relevant time series sequence for each occurrence of the extracted time series event to identify positive cases and negative cases to calculate a measure of predictive power of the time series sequence;

training a prediction algorithm using training samples to identify the positive cases and ignore the negative cases of the relevant time series sequence;

storing time series data patterns for the relevant time series sequences having a high predictive value; and

performing data mining on historical data within a database to create new templates for the time series sequences having high predictive value.

20. The method of claim 19, further comprising analyzing incoming new time series data arriving in the data stream to determine pattern matches with the time series templates to predict an occurrence of an event.

21. The method of claim 20, further comprising assessing a likelihood of occurrence of the event by determining whether the new time series data is a strong match to the pattern of one or more templates.

22. The method of claim 21, further comprising taking preventative actions on the component to prevent a future event when the likelihood of the occurrence of the event is a strong match.

23. The method of claim 19, wherein the relevant time series data indicates an alarm event.

24. The method of claim 19, wherein the relevant time series data indicates a failure event.

25. The method claim 19, wherein an identification of a positive case indicates a predicted event occurs following the time series sequence and an identification of a negative case indicates the predicted event does not occur following the time series sequence.

26. The method of claim 19, wherein the prediction algorithm comprises a genetic algorithm.

27. A system for associating time series data to pre-existing discrete data to predict future events, the system comprising:

at least one processing unit and at least one database;

a plurality of sensors in communication with the at least one processing unit;

wherein the at least one processing unit is configured to;

receive at a processor a data stream transmitted from a sensor configured to measure an operating parameter of a component being monitored, wherein the data stream comprises at least time series data;

correlate relevant time series data to pre-existing event data to detect extracted time series events;

identify each occurrence of the relevant time series data in the data stream, extract the identified relevant time series data and mark the relevant time series data as an extracted time series event;

inspect event data that chronologically follows the relevant time series sequence for each occurrence of the extracted time series event to identify positive cases and negative cases to calculate a measure of predictive power of the time series sequence;

train a prediction algorithm using training samples to identify the positive cases and ignore the negative cases of the relevant time series sequence;

store time series data patterns for the relevant time series sequences having a high predictive value; and

perform data mining on historical data within a database to create new templates for the time series sequences having high predictive value.

28. The system of claim 27, wherein the processing unit is configured to analyze incoming new time series data arriving in the data stream to determine pattern matches with the time series templates to predict an occurrence of an event.

29. The system of claim 28, wherein the processing unit is configured to assess a likelihood of occurrence of the event by determining whether the new time series data is a strong match to the pattern of one or more templates.

30. The system of claim 29, further comprises a predictive system configured to take preventative actions on the component to prevent a future event when the likelihood of the occurrence of the event is a strong match.