Method for automating the development and execution of diagnostic reasoning software in products and processes
First Claim
1. A method for automated diagnosis of faults in a system containing reparable parts comprising the steps of:
- selecting a set of faults representing all known failures which can occur among a given plurality of reparable parts contained in the system, wherein each fault is characterized by symptom data comprising expected passing or failing results for tests that can be applied to a plurality of test locations in the system;
selecting a plurality of test locations in the system at which given tests can be applied and respective outputs can be acquired as actual passing or failing results for those tests;
mapping the set of faults to their characteristic symptom data of expected passing or failing test results by generating a fault/symptom matrix defined by the set of faults as matrix row indices, by the plurality of test locations as matrix column indices, and by the expected passing and failing results for the plurality of test locations as matrix elements representing the symptom data for each of the respective faults;
performing tests at at least some of the plurality of test locations of the system, then acquiring the outputs of the performed tests as actual passing or failing test results;
correlating the actual passing or failing test results to the expected test results comprising the symptom data for the faults defined in the fault/symptom matrix; and
creating a suspect list of faults having symptom data corresponding to the actual passing or failing test results of the performed tests;
wherein said method is implemented by programmed means of a computer processing unit, said fault/symptom matrix is stored as a knowledge base of the computer processing unit, and the outputs of the tests performed on the System being diagnosed are provided as inputs to the computer processing unit;
wherein said method implemented by the programmed means of the computer processing unit is optimized for run-time efficiency by preprocessing the steps of selecting the set of faults and plurality of test locations and generating the fault/symptom matrix and storing it as a preprocessed, deterministic part of the knowledge base beforehand, then dynamically processing the steps of acquiring the outputs of the performed tests, correlating the actual test results to the stored fault/symptom matrix, and creating the suspect list of faults at run time.
3 Assignments
0 Petitions
Accused Products
Abstract
A method for automated diagnosis of faults in a system containing repairable parts is performed by selecting a set of faults representing all known failures which can occur among the parts of the system, characterized by symptom data representing the expected passing or failing results for tests applied at selected test locations in the system, generating a fault/symptom matrix of the set of faults mapped to the expected passing and failing results for the selected test locations, then performing actual tests one or more test locations and correlating the actual passing or failing test results to the fault/symptom matrix in order to identify a suspect list of faults. Additional tests may be performed until the suspect list cannot be reduced further. For efficiency, the tests are selected according to which have most diagnostic significance. The design data for the parts of the system are captured and the fault/symptom matrix is preprocessed for diagnostic efficiency and speed during run time. The diagnostic method can be implemented on a single microchip and embedded in a system for on-line diagnosis during system operation, for equipment testing, or even operator training.
-
Citations
18 Claims
-
1. A method for automated diagnosis of faults in a system containing reparable parts comprising the steps of:
-
selecting a set of faults representing all known failures which can occur among a given plurality of reparable parts contained in the system, wherein each fault is characterized by symptom data comprising expected passing or failing results for tests that can be applied to a plurality of test locations in the system; selecting a plurality of test locations in the system at which given tests can be applied and respective outputs can be acquired as actual passing or failing results for those tests; mapping the set of faults to their characteristic symptom data of expected passing or failing test results by generating a fault/symptom matrix defined by the set of faults as matrix row indices, by the plurality of test locations as matrix column indices, and by the expected passing and failing results for the plurality of test locations as matrix elements representing the symptom data for each of the respective faults; performing tests at at least some of the plurality of test locations of the system, then acquiring the outputs of the performed tests as actual passing or failing test results; correlating the actual passing or failing test results to the expected test results comprising the symptom data for the faults defined in the fault/symptom matrix; and creating a suspect list of faults having symptom data corresponding to the actual passing or failing test results of the performed tests; wherein said method is implemented by programmed means of a computer processing unit, said fault/symptom matrix is stored as a knowledge base of the computer processing unit, and the outputs of the tests performed on the System being diagnosed are provided as inputs to the computer processing unit; wherein said method implemented by the programmed means of the computer processing unit is optimized for run-time efficiency by preprocessing the steps of selecting the set of faults and plurality of test locations and generating the fault/symptom matrix and storing it as a preprocessed, deterministic part of the knowledge base beforehand, then dynamically processing the steps of acquiring the outputs of the performed tests, correlating the actual test results to the stored fault/symptom matrix, and creating the suspect list of faults at run time. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18)
-
Specification