Bearing anomaly detection and location

• US 6,999,884 B2
• Filed: 12/15/2003
• Issued: 02/14/2006
• Est. Priority Date: 01/10/2003
• Status: Active Grant

• Current Assignee: Rolls-Royce PLC (Rolls-Royce Holdings Plc)
• Original Assignee: Oxford Biosignals (Us) Incorporated
• Inventors: Anuzis, Paul, King, Stephen Peter, Astley, Kenneth Richard, King, Dennis Maxwell
• Primary Examiner(s): Ngheim, Micheal
• Assistant Examiner(s): LE, JOHN H
• Application Number:
  US10/734,152
• Publication Number:
  US 20050119840A1
• Time in Patent Office: 792 Days
• Field of Search: 702/34, 702/42, 702/56, 702/57, 702/190, 736/60, 706/20, 700/71, 340/683
• US Class Current: 702/56
• CPC Class Codes: G01M 13/045 Acoustic or vibration analysis
• CPC Class Codes: G01N 2291/2696 Wheels, Gears, Bearings

Abstract

Novel tracked orders (i.e., tracked orders that are not present in “healthy” machinery) are useful for locating bearing anomalies. Accordingly, a method for locating bearing anomalies in machinery is provided that includes receiving vibration measurements acquired from the machinery, analyzing the vibration measurements to identify novel tracked orders indicative of bearing anomalies, and ascertaining the location of a bearing anomaly by relating a novel tracked order thus-identified to one or more further tracked orders. Thus, the novel tracked order does not merely indicate the occurrence of a bearing anomaly, but, in combination with the one or more further tracked orders, allows the bearing anomaly to be traced to a particular position.

Claims

1. A method for detecting bearing anomalies in machinery, which comprises performing at each of a plurality of times the steps of:

constructing a condition signature from a plurality of condition indicators including (a) a plurality of vibration measurements acquired from the machinery or (b) one or more vibration measurements and one or more performance parameter measurements acquired from the machinery;

predicting a normal signature corresponding to the condition signature for the machinery without bearing anomalies;

comparing the condition signature with the normal signature; and

registering a bearing anomaly if the condition signature differs from the normal signature by more than a predetermined threshold.

2. A method according to claim 1, wherein the normal signature is predicted from a model defining one or more inter-dependencies between the condition indicators.

3. A method according to claim 2, wherein the model is a learnt model.

4. A method according to claim 2, wherein the model comprises a matrix with one or more non-zero off-diagonal terms to define said inter-dependencies.

5. A method according to claim 4, wherein the step of comparing the condition signature with the normal signature involves calculating a value for the normalized innovations squared.

6. A method according to claim 2, wherein the model comprises a neural network.

7. A method according to claim 6, wherein the step of comparing the condition signature with the normal signature involves calculating a prediction error.

8. A method according to claim 1, wherein said times define successive intervals of at most 1 sec duration.

9. A method according to claim 1, wherein the machinery comprises a gas turbine engine.

10. A data processing system for detecting bearing anomalies in machinery, comprising:

data acquisition devices for acquiring a plurality of condition indicators from the machinery at each of a plurality of times, the condition indicators including (a) a plurality of vibration measurements or (b) one or more vibration measurements and one or more performance parameter measurements;

a processor for constructing a condition signature from said vibration measurements and for predicting a normal signature corresponding to the condition signature for the machinery without bearing anomalies;

a comparator for comparing the condition signature with the normal signature; and

a register for registering a bearing anomaly if the comparator indicates that the condition signature differs from the normal signature by more than a predetermined threshold.