Method for data filtering and anomoly detection

• US 6,584,434 B1
• Filed: 04/24/2000
• Issued: 06/24/2003
• Est. Priority Date: 04/24/2000
• Status: Expired due to Term

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Schick, Louis Andrew, Sanborn, Stephen Duane, Catharine, Douglas Ancona
• Primary Examiner(s): Barlow, John
• Assistant Examiner(s): BHAT, ADITYA S
• Application Number:
  US09/556,987
• Time in Patent Office: 1,156 Days
• Field of Search: 702/190, 379/112, 370/286, 364/555, 364/572, 364/550, 364/900, 385/12, 235/455, 846/03
• US Class Current: 702/190
• CPC Class Codes: F23N 2223/04 Memory
• CPC Class Codes: F23N 2223/06 Sampling
• CPC Class Codes: F23N 2241/20 Gas turbines
• CPC Class Codes: F23N 5/003 using detectors sensitive to combustion gas properties F23N5/02, F23N5/18 - F23N5/26 take precedence
• CPC Class Codes: F23N 5/10 using thermocouples
• 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

Abstract

A method for low pass filtering data used in change-detect compression of data collected from a system includes buffering the data from said system. Rolling averages of the buffered data are calculated where the calculation of the rolling averages low pass filters the data. Change-detect compression is performed on the rolling averaged data, and the compressed data are archived. The archived data are transmitted to a central location and received. The received data are archived at the central location.

Claims

1. A method for low pass filtering data used in change-detect compression of said data collected from a system under test, said method comprising the steps of:

buffering said data from said system under test at a frequency of about one hertz;

calculating rolling averages of the buffered data wherein said step of calculating said rolling averages low pass filters the buffered data;

performing change-detect compression on the rolling averaged data wherein said step of performing said change-detect compression calculates a change-delta and compares said calculated change-delta with a predetermined change-delta for each of said rolling averaged data; and

archiving the compressed data.

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

transmitting the archived data points to a central location;

receiving the transmitted data at said central location; and

archiving the received data at said central location.

3. The method of claim 1, wherein said step of archiving the compressed data comprises archiving the compressed data a predetermined amount of time.

4. The method of claim 1, wherein said step of calculating said rolling average comprises continuously calculating said rolling average of said data.

5. The method of claim 1, further comprising the step of over-writing the archived data after said step of transmitting the archived data.

6. The method of claim 1, further comprising the step of deleting the archived data after said step of transmitting the archived data.

7. The method of claim 1, wherein said step of archiving the compressed data comprises archiving the compressed data points in a dynamic memory location.

8. The method of claim 1, wherein said step of archiving the compressed data comprises archiving the compressed data on a magnetic medium.

9. The method of claim 1, wherein said step of archiving the received data at said central location comprises archiving the received data on a magnetic medium.

10. A method for filtering and determining anomalies of corrected data from a system under test, said method comprising the steps of:

buffering said data from said system under test;

calculating rolling averages of the buffered data wherein said step of calculating said rolling averages low pass filters the buffered data wherein said step of performing said change-detect compression calculates a change-delta and compares said calculated change-delta with a predetermined change-delta for each of said rolling averaged data;

performing change-detect compression on the rolling averaged data;

archiving the compressed data;

transmitting the archived data points to a central location;

receiving the transmitted data at said central location;

archiving the received data at said central location

gathering the archived data from said central location;

filtering said gathered data into at least one subset, each of said at least one subset being differentiated by mode;

correcting said at least one subset of the gathered data;

calculating distributive statistics for said each of said at least one subset; and

identifying long-term anomalies in said at least one subset.

11. The method of claim 10, wherein said mode comprises an operating condition of said system under test for which constant values are expected.

12. The method of claim 10, wherein said mode comprises a known deterministic function of operating said system under test.