Method of identifying abnormal behavior in a fleet of vehicles

• US 6,732,063 B2
• Filed: 02/01/2002
• Issued: 05/04/2004
• Est. Priority Date: 02/01/2002
• Status: Expired due to Fees

• Current Assignee: National Research Council Canada
• Original Assignee: National Research Council Canada
• Inventors: OʼBrien, Chris, Famili, A., Letourneau, Sylvain
• Primary Examiner(s): WACHSMAN, HAL D
• Application Number:
  US10/060,388
• Publication Number:
  US 20030149550A1
• Time in Patent Office: 823 Days
• Field of Search: 702/188, 702/33-36, 702/56, 702/113-115, 702/159, 702/182-185, 702/187, 702/193, 702/FOR.155, 702/FOR-170, 702/FOR-123, 702/FOR.134, 702/FOR.135, 702/FOR.137, 702/FOR.139, 701/29, 701/33, 701/35, 701/99, 701/14, 701/30, 701/34, 701/100, 701/31, 244/1.R, 340/500, 340/945, 340/971, 340/973, 340/286.01, 706/905, 706/913
• US Class Current: 702/188
• CPC Class Codes: G05B 23/0235 based on a comparison with predetermined threshold or range, e.g. "classical methods", carried out during normal operation; threshold adaptation or choice; when or how to compare with the threshold

Abstract

A method is disclosed for monitoring a complex system, such as a fleet of aircraft, having multiple sub-systems described by a plurality of operating parameters. Data pertaining to the operating parameters is continually generated during operation of the vehicles. The data is normalized to take into account variability factor and stored in a central database. New incoming data from the sub-systems is continually compared with the stored data to identify abnormalities. The invention is applicable to the monitoring of a fleet of aircraft.

Claims

1. A method of monitoring a complex system having multiple sub-systems described by a plurality of operating parameters, comprising:

continually generating data pertaining to said operating parameters during operation of said system;

storing said data in a central database;

defining a window of samples over which said data is to be analyzed;

normalizing said data to take into account variability factors and introduce a weighting factor to define thresholds dependent on the performance of individual monitored components and the performance of said components across the system;

storing said defined thresholds for said defined window; and

continually comparing new incoming data from said sub-systems with said stored defined thresholds to identify abnormalities in the system.

2. A method as claimed in claim 1, wherein said window represents a specific time period.

3. A method as claimed in claim 1, wherein each said threshold represents the expected marginal value of a given parameter.

4. A method as claimed in claim 3, wherein each said threshold is determined according to the formula:

5. A method as claimed in claim 4, wherein an alert A is generated when an abnormality is identified.

6. A method as claimed in claim 5, wherein said alert A is a function of:

A=ƒ

(p, ws, wo, nt, ac)

where;

p=parameter of interest

ws=window size, (number of cases)

wo=window overlap,

nt=current normalized threshold,

ac=alert condition

and wherein an alert condition ac represents a percent of excedence from a current normalized threshold.

7. A method as claimed in claim 1, wherein each said window is partitioned into smaller windows of equal size, and the data in said smaller windows is compared to identify trends.

8. A method as claimed in claim 1, wherein an alert A is generated when an abnormality is generated, and said alert is a function of:

9. A method as claimed in claim 1, wherein said sub-systems are located on a fleet of vehicles transferring data pertaining to said operating parameters to said central database.

10. A method as claimed in claim 9, wherein said vehicles are aircraft connected to said central database by an air-to-ground link.