METHODS SYSTEMS AND APPARATUS FOR RANKING TESTS USED TO IDENTIFY FAULTS IN A SYSTEM

• US 20120323518A1
• Filed: 06/15/2011
• Published: 12/20/2012
• Est. Priority Date: 06/15/2011
• Status: Active Grant

• Current Assignee: Honeywell International Inc.
• Original Assignee: Honeywell International Inc.
• Inventors: Shao, Qingqiu Ginger, Lechtenberg, Kelly Jean
• Granted Patent: US 8,831,901 B2
• Time in Patent Office: Days
• US Class Current: 702/108
• CPC Class Codes: G06Q 10/20 Administration of product repair or maintenance

Abstract

Methods, systems and apparatus are provided for ranking tests of interest. A set of failure modes of interest and a set of tests of interest are determined A differentiation factor is then computed for each of the tests of interest, and each of the tests of interest can then be ranked based on their respective differentiation factors.

Claims

1. A method for ranking tests of interest, the method comprising:

determining a set of failure modes of interest;

determining a set of tests of interest;

computing a differentiation factor for each of the tests of interest; and

ranking each of the tests of interest based on respective differentiation factors.

2. A method according to claim 1, wherein ranking comprises:

ranking each of the tests of interest based on respective differentiation factors in a first order from a highest-valued differentiation factor to a lowest-valued differentiation factor.

3. A method according to claim 1, wherein the differentiation factor has a value that is between zero and one.

4. A method according to claim 3, wherein the differentiation factor is computed based on:

a separation factor; and

a ranking constant (c_r) that is a configurable value that is greater than or equal to 0 and less than or equal to 1.

5. A method according to claim 4, wherein the differentiation factor is further computed based on:

either an ambiguity factor or an uncertainty factor.

6. A method according to claim 5, wherein the differentiation factor is equal to a sum of:

the separation factor; and

a product of the ranking constant (c_r) and the ambiguity factor.

7. A method according to claim 6, wherein the ranking constant (c_r) is a configurable value used to scale weight of the ambiguity factor during computation of the differentiation factor.

8. A method according to claim 5, wherein the differentiation factor is equal to a sum of:

a product of the separation factor, and one minus the ranking constant (c_r); and

a product of the ranking constant (c_r), and one minus an uncertainty factor.

9. A method according to claim 8, wherein the ranking constant (c_r) is a configurable value used to scale weight of the uncertainty factor and weight of the separation factor during computation of the differentiation factor.

10. A method according to claim 5, wherein the uncertainty factor, the separation factor, and the ambiguity factor are based upon a number of possible outcomes of a test of interest, and

wherein a sum of the uncertainty factor, the separation factor, and the ambiguity factor is one.

11. A method according to claim 1, further comprising:

determining whether any of the differentiation factors are equal; and

when at least one of the differentiation factors is equal to another one of the differentiation factors;

computing, for each of the tests of interest that have a differentiation factor equal to that of one or more other tests of interest, a coverage factor; and

re-ranking each of the tests of interest based on respective differentiation factors, and then based on respective coverage factors for any tests of interest that have a differentiation factor that is equal to that of another test of interest so that tests of interest having equal differentiation factors are ranked from a highest-valued coverage factor to a lowest-valued coverage factor.

12. A method according to claim 11, wherein each of the tests of interest are ranked based on respective differentiation factors and then by respective coverage factors when any of the differentiation factors are equal, such that each of the tests having equal differentiation factors are ranked based on their respective coverage factors.

13. A method according to claim 11, wherein the respective coverage factors computed for each of the tests of interest that have equal differentiation factors are used to distinguish between and specify a ranking order of the tests of interest that have the same differentiation factors.

14. A method according to claim 11, wherein each of the respective coverage factors has a value that is between zero and infinity.

15. A method according to claim 11, wherein each of the respective coverage factors is an average number of test results associated with a failure mode of interest that is equal to:

a number of test results associated with a failure mode of interest (|R<sub>T→

FI</sub>|) divided by a number of failure modes of interest (|FI|).

16. A method according to claim 5, wherein the uncertainty factor has a value that is between zero and one that corresponds to a probability that a test will provide no information to differentiate between the likelihoods of the failure modes of interest;

and

wherein the uncertainty factor is a sum of;

(a) a probability the test will not affect the likelihood of any failure modes of interest;

(b) a probability the test will increase the likelihood of all failure modes of interest; and

(c) a probability the test will decrease the likelihood of all failure modes of interest; and

wherein the uncertainty factor is calculated by dividing a first sum by a total number of outcomes, wherein the first sum is a sum of (a) a number of outcomes of the test which will not affect the likelihood of a failure mode of interest, (b) a number of outcomes of the test which will increase all failure modes of interest, and (c) a number of outcomes that will decrease the likelihood of all failure modes of interest.

17. A method according to claim 16, wherein the separation factor has a value that is between zero and one that corresponds to a probability that a test will separate out one and only one of the failure modes of interest from the other failure modes of interest;

and

wherein the separation factor is a sum of;

(a) a probability that the test will increase the likelihood of only one particular failure mode of interest; and

(b) a probability that the test will decrease the likelihood of all but one failure mode of interest and does not also increase the likelihood of the only one particular failure mode of interest; and

wherein the separation factor is calculated by dividing a second sum by the total number of outcomes, wherein the second sum is a sum of (a) a number of outcomes of the test which will increase the likelihood of the only one particular failure mode of interest, and (b) a number of outcomes which will decrease the likelihood of all but one failure mode of interest but not increase the likelihood of one failure mode of interest.

18. A method according to claim 17, wherein the ambiguity factor has a value that is between zero and one that corresponds to a probability that a test will provide some information to differentiate between the failure modes of interest, but not separate out a single failure mode of interest;

and

wherein the ambiguity factor is a sum of;

(a) a probability that the test will increase the likelihood of more than one failure mode of interest, but not increase the likelihood of all failure modes of interest; and

(b) a probability the test will decrease the likelihood of at least one failure mode of interest, but not decrease the likelihood of all but one failure mode of interest, and not decrease the likelihood of all failure modes of interest, and not increase the likelihood of any failure modes of interest; and

wherein the ambiguity factor is calculated by dividing a number of outcomes which is not counted towards either the uncertainty factor or the ambiguity factor by the total number of outcomes.

19. A method according to claim 1, wherein each test result in a set of test results of a test (R_T) is either associated with or not associated with a particular failure mode of interest, and wherein the set of test results includes a subset comprising one or more test results which are associated with the one or more failure modes of interest,

wherein the set of failure modes of interest (FI) is a subset of all failure modes in a failure mode ambiguity group (FM_FC) of a fault condition, wherein the set of failure modes of interest (FI) comprises any failure mode (fm) in the failure mode ambiguity group (FM_FC) that is significant based on an isolation constant (c), wherein the isolation constant (c) is a configurable value used to decide if a fault condition can be isolated to a failure mode (fm) or group of corrective actions, and wherein the set of failure modes of interest (FI) include:

a primary failure mode in the failure mode ambiguity group (FM_FC) that has a greatest individual likelihood (fm_max), and other failure modes in the failure mode ambiguity group that have an individual likelihood that is great enough to prevent the fault condition from isolating to a primary failure mode.

20. A method according to claim 19, wherein determining a set of tests of interest, comprises:

determining the set of tests of interest, based on the set of failure modes (FM), wherein the tests of interest include any test which has at least one result that is associated with at least one failure mode (fm) in the failure mode ambiguity group of the fault condition.