Machine operation management

• US 9,076,271 B2
• Filed: 06/08/2010
• Issued: 07/07/2015
• Est. Priority Date: 07/08/2009
• Status: Active Grant

• Current Assignee: Rolls-Royce PLC (Rolls-Royce Holdings Plc)
• Original Assignee: Optimized Systems and Solutions Ltd (Rolls-Royce Holdings Plc)
• Inventors: Shi, Dongfeng, Clarke, Toby, Dibsdale, Charles, Douglas, Richard C. T., Shone, Peter, Puglia, William
• Primary Examiner(s): Breene, John
• Assistant Examiner(s): Hagos, Eyob
• Application Number:
  US13/379,196
• Publication Number:
  US 20120150491A1
• Time in Patent Office: 1,855 Days
• Field of Search: 702/184
• US Class Current: 1/1
• CPC Class Codes: F05B 2260/80 Diagnostics
• 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
• CPC Class Codes: G07C 3/00 Registering or indicating the condition or the working of machines or other apparatus, other than vehicles

Abstract

A method for managing machine operation comprising sensing a plurality of operational variables for a machine during use thereof so as to generate operational data for said variables. The operational data is processed so as to determine features within the operational data which are indicative of a divergence from a desired operational state of the machine. Confidence values associated with said features are determined and used to assess whether the plurality of features and associated confidence values are indicative of a predetermined diagnosis for said machine. A confidence value for said diagnosis is determined based upon the associated feature confidence values and used to generate a signal indicative of an operational state of the machine. The invention may be used for engine health monitoring applications and may be used for determining necessary servicing or repair work for the engine.

Claims

1. A method for managing machine operation comprising:

defining a network relationship structure having three bands, the first band comprising possible features, the second band comprising possible symptoms and the third band comprising possible diagnoses, wherein the features are linked to symptoms and symptoms are linked to diagnoses, the network relationship structure being defined such that each single symptom is linked to one or a plurality of features and each single symptom is linked to one or a plurality of diagnoses;

sensing a plurality of operational variables for a machine during use thereof so as to generate operational data for said variables and store said operational data in a memory;

processing, using a hardware processor, said operational data so as to determine features within the operational data indicative of an abnormal event representing a divergence from a desired operational state;

determining a confidence value associated with said features;

assessing whether a plurality of features and associated confidence value are indicative of a predetermined diagnosis for said machine;

determining a confidence value for said diagnosis based upon the associated feature confidence value, the confidence value for said diagnosis being different from the associated feature confidence value; and

outputting a signal indicative of an operational state of the machine if the diagnosis confidence value exceeds a threshold diagnosis confidence.

2. A method according to claim 1, wherein

the features and diagnoses are represented as nodes within said network structure and links are formed there-between to identify which features impact on which diagnoses.

3. A method according to claim 1, wherein

assessing whether a plurality of features and associated confidence values are indicative of a predetermined diagnosis for said machine, comprises determining one or more symptom confidence values from the plurality of feature confidence values and assessing whether the one or more symptom confidence values are indicative of a predetermined diagnosis.

4. A method according to claim 3, wherein

the one or more symptom confidence values are representative of a summation or averaging of the feature confidence values related thereto.

5. A method according to claim 3, wherein

the diagnosis confidence value is representative of a summation or averaging of the symptom confidence values related thereto.

6. A method according to claim 1, wherein

feature confidence values are determined based upon a divergence of said operational data from a desired operational state beyond a predetermined threshold value or rate of change.

7. A method according to claim 1, wherein

feature confidence values are determined based upon divergence of operational data from a desired operational state for a plurality of operational variables.

8. A method according to claim 1, wherein

the determination of any or any combination of a feature, symptom and/or diagnosis confidence value comprises definition of a time period value associated with said feature, symptom or diagnosis.

9. A method according to claim 8, wherein

features occurring within the time period prescribed for another feature are determined to be linked and the determining of the confidence value for said associated diagnosis is based upon said linked feature confidence values.

10. A method according to claim 8, wherein

the processing said operational data comprises scanning for linked features which occur within said time period.

11. A method according to claim 1, comprising

scheduling testing, maintenance or repair work for said machine based upon the diagnosis confidence value.

12. A tool for diagnosis or prognosis of machine operation characteristics comprising:

one or more hardware processors arranged for data communication with a plurality of sensors, said sensors arranged to record a plurality of operational variables during use of a machine so as to generate machine operational data for said variables and store said operational data in a memory,

the one or more processors arranged to receive said machine operational data and to;

define a network relationship structure having three bands, the first band comprising possible features, the second band comprising possible symptoms and the third band comprising possible diagnoses, wherein the features are linked to symptoms and symptoms are linked to diagnoses, the network relationship structure being defined such that each single symptom is linked to one or a plurality of features and each single symptom is linked to one or a plurality of diagnoses;

control the processing of said operational data so as to determine features within the operational data indicative of an abnormal event representing a divergence from a desired operational state;

determine a confidence value associated with said features;

assess whether a plurality of features and associated confidence value are indicative of a predetermined diagnosis for said machine;

determine a confidence value for said diagnosis based upon the associated feature confidence value, the confidence value for said diagnosis being different from the associated feature confidence value; and

control the output of a signal indicative of an operational state of the machine dependent on the diagnosis confidence value.

13. A data carrier comprising non-transitory machine readable instruction for controlling operation of one or more processors to perform the processing steps of:

defining a network relationship structure having three band, the first band comprising possible features, the second band comprising possible symptoms and the third band comprising possible diagnoses, wherein the features are linked to symptoms and symptoms are linked to diagnosis, the network relationship structure being defined such that each single symptom is linked to one or a plurality of features and each single symptom is linked to one or a plurality of diagnoses;

receiving operational data for a plurality of operational variables for a machine;

determine features within the operational data indicative of an abnormal event representing a divergence from a desired operational state of said machine;

determining a confidence value associated with said features;

assessing whether a plurality of features and associated confidence value are indicative of a predetermined diagnosis for said machine; and

determining a confidence value for said diagnosis based upon the associated feature confidence value, the confidence value for said diagnosis being different from the associated feature confidence value.