System for gas turbine health monitoring data fusion

• US 7,769,507 B2
• Filed: 08/26/2004
• Issued: 08/03/2010
• Est. Priority Date: 08/26/2004
• Status: Active Grant

• Current Assignee: Rtx Corporation
• Original Assignee: United Technologies Corporation (Raytheon Technologies Corporation d/b/a RTX)
• Inventors: Wood, C. Bruce, Volponi, Allan J.
• Primary Examiner(s): Tran; Dalena
• Application Number:
  US10/926,464
• Publication Number:
  US 20060047403A1
• Time in Patent Office: 2,168 Days
• Field of Search: 701/27, 701/29, 701/106, 700 47- 49, 700/291, 706/2, 706 15- 16, 706/20, 706/905, 706/913, 702 57- 59, 702 64- 65, 702/81, 702/84
• US Class Current: 701/29
• CPC Class Codes: F01D 17/02 Arrangement of sensing elements
• CPC Class Codes: F05D 2260/80 Diagnostics
• CPC Class Codes: F05D 2270/707 fuzzy logic
• CPC Class Codes: F05D 2270/709 with neural networks
• 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

An apparatus for assessing health of a device comprising a data alignment module for receiving a plurality of sensory outputs and outputting a synchronized data stream, an analysis module for receiving the synchronized data stream and outputting at least one device health feature, and a high level diagnostic feature information fusion module for receiving the at least one device health feature and outputting a device health assessment.

Claims

1. An apparatus for assessing health of a device comprising:

a plurality of sensors for acquiring data at different sampling rates and for generating a plurality of sensory outputs at said different sampling rates;

a data alignment module for receiving said plurality of sensory outputs at different sampling rates a beginning of a window duration, processing said plurality of sensory outputs to produce a synchronized data stream, and outputting said synchronized data stream as time synchronized data at an end of said window duration;

an analysis module for receiving said synchronized data stream and outputting at least one device health feature;

a high level diagnostic feature information fusion module for receiving said at least one device health feature and transforming said at least one device health feature in real time into a device health assessment; and

a fault isolation reasoner module for combining said engine health assessment with at least one additional data source to produce a recommended maintenance action.

2. The apparatus of claim 1 wherein said device is an engine.

3. The apparatus of claim 1 wherein said at least one additional data source is selected from the group consisting of maintainer observations, pilot observations, engine maintenance history, and Full Authority Digital Engine Control (FADEC) fault codes.

4. The apparatus of claim 1 wherein said sensory outputs are selected from the group consisting of altitude, mach, ambient temperature, ambient pressure, temperature, pressure, gas flow speeds, oil temperatures, oil pressures, and oil quantities.

5. The apparatus of claim 1 wherein said sensors are selected from the group consisting of accelerometers, oil debris monitors, oil condition monitors, electrostatic debris monitors, acoustic monitors, and eddy current sensors.

6. The apparatus of claim 1 additionally comprising at least one feature extraction unit for extracting a feature from at least one of said sensory outputs.

7. The apparatus of claim 1 wherein said analysis module is adapted to utilize models selected from the group consisting of State Variable Engine Models with Kalman observers, empirical engine models, physics based lubrication system models, empirical based lubrication system models, and engine component life models.

8. The apparatus of claim 1 wherein said analysis module is adapted to utilize constructs selected from the group consisting of Kalman Filters, Artificial Neural Networks, and Fuzzy Logic based systems.

9. The apparatus of claim 1 wherein said high level diagnostic feature information fusion module is adapted to utilize Bayesian Belief Networks (BBN), and Fuzzy Belief Networks (FBN).

10. A method for assessing health of a device comprising the steps of:

providing a plurality of sensors for acquiring data at different sampling rates and for generating a plurality of sensory outputs at said different sampling rates;

providing a data alignment module;

inputting said plurality of sensory outputs into said data alignment module and outputting a time synchronized data stream from said data alignment module;

inputting said time synchronized data stream into an analysis module and outputting at least one device health feature from said analysis module;

inputting said at least one device health feature into a high level diagnostic feature information fusion module and transforming said at least one device health feature in real time into a device health assessment; and

combining said engine health assessment with at least one additional data source in a fault isolation reasoner module to produce and output a recommended maintenance action from said fault isolation reasoner module

wherein said step of inputting said plurality of sensory outputs comprises the steps of

commencing to receive said plurality of sensory outputs at a beginning of a window duration;

sampling at least one of said plurality of sensory outputs to produce a sample data stream in said data alignment module; and

said outputting a time synchronized data stream from said data alignment module step comprising outputting said time synchronized data stream at an end of said window duration.

11. The method of claim 10 wherein said step of outputting said at least one device health feature comprises outputting at least one engine health feature.

12. The method of claim 10 wherein said step of commencing to receive said plurality of sensory outputs comprises commencing to receive said plurality of sensory outputs at said beginning of said window duration and continuing to receive said plurality of sensory outputs for between 33 ms and 100 ms.

13. The method of claim 10 wherein said step of commencing to receive said plurality of sensory outputs comprises commencing to receive said plurality of sensory outputs at said beginning of said window duration and continuing to receive said plurality of sensory outputs for between 50 ms and 66 ms.

14. The method of claim 10 wherein said inputting said plurality of sensory outputs comprises inputting said plurality of sensory outputs selected from the group consisting of altitude, mach, ambient temperature, ambient pressure, temperature, pressure, gas flow speeds, oil temperatures, oil pressures, and oil quantities.

15. The method of claim 10 wherein said inputting said plurality of sensory outputs comprises inputting said plurality of sensory outputs from a plurality of sensors selected from the group consisting of accelerometers, oil debris monitors, oil condition monitors, electrostatic debris monitors, acoustic monitors, and eddy current sensors.

16. The method of claim 10 further comprising performing analysis of said time synchronized data stream in said analysis module by utilizing at least one model selected from the group consisting of State Variable Engine Models with Kalman observers, empirical engine models, physics based lubrication system models, empirical based lubrication system models, and engine component life models.

17. The apparatus of claim 10 further comprising performing analysis of said time synchronized data stream in said analysis module by utilizing at least one construct selected from the group consisting of Kalman Filters, Artificial Neural Networks, and Fuzzy Logic based systems.

18. An apparatus for assessing the health of a device, comprising:

a plurality of sensors for acquiring data at different sampling rates and for generating a plurality of sensory outputs at said different sampling rates;

a means for aligning data received from said plurality of sensors to produce a time synchronized data stream;

a means for analyzing said time synchronized data stream to produce at least one device health feature;

a means for fusing high level diagnostic feature information with said at least one device health feature to transform said at least one device health feature in real time into a device health assessment; and

a means for combining said engine health assessment with at least one additional data source to produce a recommended maintenance action,

wherein said means for aligning data further comprises;

a means for commencing to receive a plurality of said sensory outputs at a beginning of a window;

a means for sampling at least one of said plurality of sensory outputs to produce a sample data stream; and

a means for outputting said time synchronized data stream comprising said sampled data stream at an end of said window duration.