Parameter independent detection of rotating machinery faults

• US 8,544,331 B2
• Filed: 12/04/2009
• Issued: 10/01/2013
• Est. Priority Date: 12/04/2008
• Status: Active Grant

• Current Assignee: University of Ottawa
• Original Assignee: University of Ottawa
• Inventors: Liang, Ming, Bozchalooi, Iman Soltani
• Primary Examiner(s): Macchiarolo, Peter
• Assistant Examiner(s): MILLER, ROSE MARY
• Application Number:
  US12/631,528
• Publication Number:
  US 20100139403A1
• Time in Patent Office: 1,397 Days
• Field of Search: 736/60, 736/59, 735/93
• US Class Current: 73/660
• CPC Class Codes: G01N 2291/2696 Wheels, Gears, Bearings
• CPC Class Codes: G01N 29/14 using acoustic emission techniques echo of particles G01N29/046; measuring mechanical vibrations or acoustic waves in solids in general G01H1/00
• CPC Class Codes: G01N 29/4454 Signal recognition, e.g. specific values or portions, signal events, signatures
• CPC Class Codes: G01N 29/46 by spectral analysis, e.g. Fourier analysis or wavelet analysis spectral signal processing per se G06F17/14

Abstract

A parameter-free method to analyze sensor signals incorporates two or more of frequency demodulation, amplitude demodulation and phase demodulation of the raw signal data. The resulting signal is transformed to a frequency domain, and target fault characteristics from the demodulated signal are identified. The method is used to detect faults in bearings, gears and other mechanical components.

Claims

1. A method for detecting the existence of one or more faults of one or more mechanical components of an operating machine or a mechanical system from raw signal data comprising one or more target fault characteristic signals, one or more interference signals and intrinsic noise signals, said one or more target fault characteristic signals featuring periodically occurring frequency, phase and amplitude modulations, said method comprising:

operating the machine or mechanical system including the one or more mechanical components, at least one of the one or more mechanical components being susceptible to developing one or more mechanical faults, said one or more mechanical faults each having a target fault characteristic signal, said target fault characteristic signal being created by a fault in said at least one mechanical component;

obtaining said raw signal data with one or more sensors positioned proximate to the machine;

transforming said raw signal data using one or more transformations which simultaneously incorporate any combination of two or more of frequency, phase and amplitude demodulations of said raw signal data into a new set of data;

transforming said new set of data to a frequency domain;

identifying a target fault characteristic signature from the frequency domain selected from the group consisting of an impulsive target signal, a non-impulsive target signal, a combination of multiple impulsive target signals, a combination of multiple non-impulsive target signals, and a combination of multiple impulsive and non-impulsive target signals, said target fault characteristic signature corresponding to said target fault characteristic signal;

generating an alert corresponding to the identified target fault characteristic signature detected in the machine or mechanical system.

2. The method as defined in claim 1 wherein obtaining said raw signal data includes obtaining said one or more interference signals having a frequency range which is lower than that of any resonance excited by the one or more mechanical faults.

3. The method as defined in claim 1 further comprising removing said intrinsic noise signals from said raw signal data source to transforming the raw data signal.

4. The method as defined in claim 3 wherein removing said intrinsic noise signals includes performing a thresholding noise reduction or removal technique.

5. The continuous use of the method as defined in claim 3 to monitor conditions of the machine or mechanical system comprising the one or more mechanical components.

6. The method as defined in claim 1 wherein transforming said raw signal data includes using a Teager Energy Operator.

7. The continuous use of the method as defined in claim 6 to monitor conditions of the machine or mechanical system comprising the one or more mechanical components.

8. The method as defined in claim 1 wherein obtaining said raw signal data with the one or more sensors includes obtaining said raw signal data from a sensor including one or more of source signal is obtained from an acoustic sensor configured to measure an acoustic signal of the machine or mechanical system and a vibration sensor configured to measure vibrations of the machine or mechanical system.

9. The method as defined in claim 8 wherein said vibration sensor is an accelerometer or laser vibrometer.

10. The method as defined in claim 8 wherein said acoustic sensor is an acoustic emission sensor or a microphone.

11. The method as defined in claim 1 wherein operating the machine or mechanical system including the one or more mechanical components includes operating a machine or mechanical system including a mechanical component chosen from the group consisting of one or more bearings, gears, journal bearings, slider cranks, cams, shafts, springs, wheels, fans, turbines, rotors, disks, impellers, propellers and dampers.

12. The continuous use of the method as defined in claim 1 to monitor conditions of the machine or mechanical system comprising the one or more mechanical components.

13. The method as defined in claim 1 further comprising changing operation of the machine or mechanical system.

14. The method as defined in claim 1 further comprising identifying a mechanical component corresponding to the target fault characteristic signature.