Automatic Rotating-Machine Fault Diagnosis With Confidence Level Indication
First Claim
1. A method of automatic fault diagnosis performed by a processor on machine diagnostic data sensed from a machine, comprising:
- analyzing with said processor said machine diagnostic data to screen a plurality of potential faults of the machine;
determining a probability value for each one potential fault of the plurality of potential faults, wherein said probability is an indication of fault severity for said one potential fault;
for said each one potential fault, deriving a confidence level of the potential fault, wherein the confidence level for said one potential fault is based in part on the probability value determined for each other one of the plurality of faults.
2 Assignments
0 Petitions
Accused Products
Abstract
Automatic fault diagnosis is performed on vibration data sensed from a machine. A set of faults to screen for is identified from the machine configuration. For each fault there are characteristic symptoms. For each characteristic symptom, there is a corresponding indication used to diagnose the symptom. The indications are based on analyses of the current vibration data. The diagnosed symptoms have weights assigned according to a Bayesian network, and are used to derive a Bayesian probability for the fault. A fault having a Bayesian probability exceeding a threshold value is identified as being present in the machine. For each fault a confidence level is derived. The confidence level for a first fault is based on a similarity between characteristic symptoms for the first fault and characteristic symptoms for each one of the other faults being screened.
36 Citations
21 Claims
-
1. A method of automatic fault diagnosis performed by a processor on machine diagnostic data sensed from a machine, comprising:
-
analyzing with said processor said machine diagnostic data to screen a plurality of potential faults of the machine; determining a probability value for each one potential fault of the plurality of potential faults, wherein said probability is an indication of fault severity for said one potential fault; for said each one potential fault, deriving a confidence level of the potential fault, wherein the confidence level for said one potential fault is based in part on the probability value determined for each other one of the plurality of faults. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
-
-
9. A method of automatic fault diagnosis performed by a processor on vibration data sensed from a machine to diagnose one or more faults in the machine, the method comprising:
-
analyzing with said processor said vibration data to screen a plurality of potential faults in the machine; for a first fault among said plurality of potential faults, deriving by said processor a confidence level as to whether said first fault is present in the machine, wherein said deriving is based on a plurality of similarity values computed by the processor for said first fault, each one similarity value for said first fault being with respect to another fault among the plurality of potential faults, wherein said one similarity value for said first fault with respect to said another fault is based in part on an intersection of characteristic symptoms for said first fault and said another fault; and displaying a confidence indication for said first fault, wherein the confidence indication is based on the confidence level for said first fault. - View Dependent Claims (10, 11, 12, 13, 14, 15, 16, 17)
-
-
18. A method for determining a confidence level for a fault diagnosed from vibration data of a machine, comprising:
-
identifying one or more components of the machine by accessing configuration data corresponding to the machine; identifying a set of faults to be screened for automatically for the machine, wherein the set is determined based upon presence or absence of prescribed components among the identified one or more components; receiving vibration data for a plurality of test point locations on a machine; processing the received vibration data to resolve a plurality of indicators; for each one fault in the identified set of faults, identifying one or more characteristic symptoms of said one fault; for each one characteristic symptom of said one or more characteristic symptoms, identifying a corresponding indicator from among the plurality of indicators; for each one characteristic symptom of said one or more characteristic symptoms, testing said corresponding indicator as resolved and computing a probability of whether said one characteristic symptom is present in the machine based on said testing of said corresponding indicator; for each one fault being screened, determining a Bayesian probability for said one fault based on each computed probability of symptom presence for said one fault; for a first fault in the identified set of faults, deriving by said processor a confidence level, and wherein said deriving is based on a plurality of similarity values computed by the processor for said first fault, each one similarity value for said first fault being with respect to another fault among the plurality of potential faults, wherein said one similarity value for said first fault with respect to said another fault is based in part on an intersection of characteristic symptoms for said first fault and said another fault; and wherein for each one similarity value said deriving further comprises with respect to said another fault, weighting said one similarity value according to the determined Bayesian probability for said another fault; displaying a confidence indication for said first fault, wherein the confidence indication is based on the confidence level for said first fault; and displaying a fault severity for said first fault, wherein the fault severity is based on the Bayesian probability for said one diagnosed fault.
-
-
19. A data collection apparatus having vibration data sensed from a machine stored in memory, and having a processor configured to perform automatic fault diagnosis on the vibration data to screen the machine for a set of potential faults;
-
wherein said processor is configured to analyze said vibration data for each one fault of a plurality of machine faults to be screened for to determine a probability value for said one fault, said determined probability being an indication of severity for said one fault and being stored in said memory; wherein said processor is configured to derive, for a first fault among said plurality of machine faults, a confidence level for said first fault, said confidence level being stored in memory; wherein said processor is configured to derive said confidence level based on a plurality of similarity values computed by the processor for said first fault, each one similarity value for said first fault being with respect to another fault among the plurality of potential faults, wherein said one similarity value for said first fault with respect to said another fault is based in part on an intersection of characteristic symptoms for said first fault and said another fault; and wherein for each one similarity value said processor is configured to derive said confidence level further based on, with respect to said another fault, weighting said one similarity value according to the determined Bayesian probability for said another fault. - View Dependent Claims (20, 21)
-
Specification