# US 10,310,929 B2 · Diagnosing combinations of failures in a system > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,310,929 B2 - Document ID: US10310929B2 - Status: Granted - Kind: B2 - Application number: 14/784,805 - Filed: 2014-04-22 - Published: 2016-03-24 - Granted: 2019-06-04 - Earliest claimed: 2013-04-19 - Assignees: BAE SYSTEMS plc - Applicants: BAE SYSTEMS plc - Inventors: Richard Lee Bovey; Erdem Turker Senalp - Web page: https://patentgrep.com/patent/US10310929B2 ## CPC - G06F 11/079: {Root cause analysis, i.e. error or fault diagnosis (in a hardware test environment G06F11/22; in a software test environment G06F11/36)} - G05B 23/0278: {Qualitative, e.g. if-then rules; Fuzzy logic; Lookup tables; Symptomatic search; FMEA} - G06F 11/0751: {Error or fault detection not based on redundancy (power supply failures G06F1/30; network fault management H04L41/06)} - G06F 11/0787: {Storage of error reports, e.g. persistent data storage, storage using memory protection} - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G) ## Abstract A system/method of diagnosing combinations of failures in a system includes receiving symptom data (116) including information relating to observed or detected symptoms in a system. The system/method generates (D4a, D4b, D5) failure data (118) including information relating to at least one most probable failures in the system based on the symptom data, and processes (D9) the failure data and the symptom data using an L-best inference (e.g. a Ranked Algorithm (RA)) technique in order to generate failure set data (120), the failure set data including information relating to at least one most probable combination of the failures that explain the symptoms. ## Claims ### Claim 1 A method of diagnosing combinations of failures in a system, the method including: receiving symptom data including information relating to observed or detected symptoms in a system; generating failure data including information relating to at least two most probable failures in the system based on the symptom data; processing the failure data and the symptom data using an L-rank inference class algorithm in order to generate failure set data, the failure set data including information relating to combinations of the at least two most probable failures that explain the symptoms; and selecting an algorithm for use in generating the failure data, wherein a comparison based on a number of present symptoms in the symptom data and a user-configurable value is used to select the algorithm for use in generating the failure data, and if the number of present symptoms is greater than a user-configurable value then an approximate marginal inference algorithm is selected; otherwise, an exact marginal inference algorithm is selected. ### Claim 2 (depends on 1) The method according to claim 1, wherein processing the failure and symptom data using the L-rank inference class algorithm includes using a Ranked Algorithm (RA) algorithm to produce a probability value associated with each said combination of the at least two most probable failures in the failure set data. ### Claim 3 (depends on 2) The method according to claim 2, further including ordering the combinations of the at least two most probable failures in the failure set data in accordance with the associated probability values. ### Claim 4 (depends on 3) The method according to claim 3, further including displaying information relating to at least one of the combinations of the at least two most probable failures and the associated probability values. ### Claim 5 (depends on 1) The method according to claim 1, wherein generating the failure data includes processing the symptom data using the approximate marginal inference algorithm, and wherein the marginal inference algorithm generates a ranked list of probable failures. ### Claim 6 (depends on 1) The method according to claim 1, wherein the at least two most probable failures not including a user-configured number of the at least two most probable failures are provided as absent failures data to be processed using the L-rank inference class algorithm. ### Claim 7 (depends on 6) The method according to claim 6, wherein each said combination of the at least two most probable failures in a said failure set has an associated confidence/probability value and the confidence/probability value is used to determine whether or not the processing using the L-rank inference class algorithm is to be used to generate at least one further failure set. ### Claim 8 (depends on 1) The method according to claim 1, further including halting the processing of the failure data and the symptom data using the L-rank inference class algorithm when a user-configurable number of said combinations of the at least two most probable failures has been generated. ### Claim 9 (depends on 1) The method according to claim 1, wherein the symptom data processed using the L-best inference class algorithm includes present symptoms that have at most a user-configurable number of parent failures. ### Claim 10 (depends on 1) The method according to claim 1, wherein a user-configurable number of the at least two most probable failures in the system, generated by generating the failure data, is included in the failure data to be processed using the L-rank inference class algorithm. ### Claim 11 An apparatus adapted to diagnose combinations of failures in a system, the apparatus including: a processor configured to receive symptom data including information relating to observed or detected symptoms in a system; a processor configured to generate failure data including information relating to at least two most probable failures in the system based on the symptom data; a processor configured to process the failure data and the symptom data using an L-rank inference class algorithm in order to generate failure set data, the failure set data including information relating to at least one most probable combination of the at least two most probable failures that explain the symptoms; and a processor configured to select an algorithm for use in generating the failure data, wherein a comparison based on a number of present symptoms in the symptom data and a user-configurable value is used to select the algorithm for use in generating the failure data, and if the number of present symptoms is greater than a user-configurable value then an approximate marginal inference algorithm is selected; otherwise, an exact marginal inference is selected, and wherein the processors may be the same processor or may include multiple processors. ### Claim 12 A non-transitory computer program product encoded with instructions that when executed by one or more processors cause a process of diagnosing combinations of failures in a system to be carried out, the process including: receiving symptom data including information relating to observed or detected symptoms in a system; generating failure data including information relating to at least two most probable failures in the system based on the symptom data; and processing the failure data and the symptom data using an L-rank inference class algorithm in order to generate failure set data, the failure set data including information relating to at least one most probable combination of the at least two most probable failures that explain the symptoms; selecting an algorithm for use in generating the failure data, wherein a comparison based on a number of present symptoms in the symptom data and a user-configurable value is used to select the algorithm for use in generating the failure data, and if the number of present symptoms is greater than a user-configurable value then an approximate marginal inference algorithm is selected; otherwise, an exact marginal inference is selected. ### Claim 13 (depends on 12) The non-transitory computer program product according to claim 12, wherein processing the failure and symptom data using the L-rank inference class algorithm includes using a Ranked Algorithm (RA) algorithm to produce a probability value associated with each said combination of the at least two most probable failures in the failure set data. ### Claim 14 (depends on 13) The non-transitory computer program product according to claim 13, further including at least one of: ordering the combinations of the at least two most probable failures in the failure set data in accordance with the associated probability values; and displaying information relating to at least one of the combinations of the at least two most probable failures and the associated probability values. ### Claim 15 (depends on 12) The non-transitory computer program product according to claim 12, wherein generating the failure data includes processing the symptom data using the approximate marginal inference algorithm, and wherein the marginal inference algorithm generates a ranked list of probable failures. ### Claim 16 (depends on 12) The non-transitory computer program product according to claim 12, wherein the failures not including the user-configured number of the most probable failures are provided as absent failures data to be processed using the L-rank inference class algorithm. ### Claim 17 (depends on 12) The non-transitory computer program product according to claim 12, the process further including halting the processing of the failure data and the symptom data using the L-rank inference class algorithm when a user-configurable number of said combinations of the failures, has been generated. ### Claim 18 (depends on 17) The non-transitory computer program product according to claim 17, wherein each said combination of the at least two most probable failures in a said failure set has an associated confidence/probability value and the confidence/probability value is used to determine whether or not the processing using the L-rank inference class algorithm is to be used to generate at least one further failure set.