Starter Control Valve Failure Prediction Machine To Predict and Trend Starter Control Valve Failures In Gas Turbine Engines Using A Starter Control Valve Health Prognostic, Program Product and Related Methods

• US 20110040470A1
• Filed: 08/14/2009
• Published: 02/17/2011
• Est. Priority Date: 08/14/2009
• Status: Active Grant

• Current Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Original Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Inventors: Iyer, Naresh Sundaram, Qiu, Hai, Yan, Weizhong
• Granted Patent: US 8,370,045 B2
• Time in Patent Office: Days
• US Class Current: 701/100
• CPC Class Codes: F01D 19/00 Starting of machines or engines; Regulating, controlling, or safety means in connection therewith warming-up before starting F01D25/10; turning or inching gear F01D25/34
• CPC Class Codes: F02C 7/26 Starting; Ignition
• CPC Class Codes: F02C 7/262 Restarting after flame-out
• CPC Class Codes: F05D 2270/708 with comparison tables

Abstract

Starter control valve failure prediction machines, systems, program products, and computer implemented methods to predict and trend starter control valve failures in gas turbine engines using a starter control valve health prognostic and to make predictions of starter control valve failures, are provided. A computer implemented method according to an embodiment of the present invention can include the steps of generating a continuous starter control valve deterioration trend function responsive to a plurality of health indices derived from gas turbine engine startup data downloaded from gas turbine engine sensors for a plurality of startups and analyzing the continuous starter control valve deterioration trend function to identify potential starter control valve failure points where the points on the starter control valve deterioration trend function correlate to a starter control valve health prognostic responsive to historic gas turbine engine startup data downloaded from gas turbine engine sensors.

Claims

1. A computer implemented method to predict starter control valve failures in gas turbine engines using a starter control valve health prognostic comprising the steps of:

extracting one or more health index features from operational gas turbine engine startup data received for at least one gas turbine engine startup of one or more gas turbine engines of a certain gas turbine engine model;

calculating a starter control valve health index for each of the at least one gas turbine engine startup responsive to the one or more health index features extracted from the operational gas turbine engine startup data to thereby define one or more health indices;

comparing each of the one or more health indices to a starter control valve health prognostic for the certain gas turbine engine model; and

determining, responsive to the step of comparing, a potential predictive starter control valve failure point, the potential predictive starter control valve failure point being located where one of the one or more health indices correlates to the starter control valve health prognostic for the certain gas turbine engine model.

2. A computer implemented method as defined in claim 1,

wherein the at least one gas turbine engine startup is a plurality of gas turbine engine startups;

and

wherein the computer implemented method further comprises the steps of;

constructing a continuous starter control valve deterioration trend function responsive to a plurality of starter control valve health indices calculated for each of the plurality of gas turbine engine startups responsive to the one or more health index features extracted from the operational gas turbine engine startup data; and

determining, using the continuous starter control valve deterioration trend function, a point on the continuous starter control valve deterioration trend function that correlates to the starter control valve health prognostic for the certain gas turbine engine model to thereby identify the potential predictive starter control valve failure point.

3. A computer implemented method as defined in claim 2,

wherein the one or more gas turbine engines of the certain gas turbine engine model are associated with an aircraft;

and

wherein the computer implemented method further comprises the steps of;

analyzing the operational gas turbine engine startup data,

determining, responsive to analyzing the operational gas turbine engine startup data, the existence of one or more mechanical problems associated with one or more of an auxiliary power unit, a compressor, an air turbine, an ignition system, a fuel system, or a lube pressure system of the aircraft at the potential predictive starter control valve failure point,

determining, responsive to analyzing the operational gas turbine engine startup data, the existence of pilot initiated purging of a fuel injection system of each of the one or more gas turbine engines at the potential predictive starter control valve failure point,

and

reducing false positive indications of a starter control valve failure responsive to determining the existence of one or more mechanical problems and determining the existence of pilot initiated purging of the fuel injection system of each of the one or more gas turbine engines at the potential predictive starter control valve failure point.

4. A computer implemented method as defined in claim 3,

wherein the continuous starter control valve deterioration trend function is a function of two variables, a first variable representing one or more of the one or more extracted health index features and a second variable representing the gas turbine engine startup number associated with the one or more of the one or more extracted health index features;

and

wherein a mathematical limit of the continuous starter control valve deterioration trend function approaches zero as the gas turbine engine startup number associated with the one or more of the one or more extracted health index features increases.

5. A computer implemented method as defined in claim 4,

wherein the starter control valve health prognostic for the certain gas turbine engine model is constructed responsive to one or more health index features extracted from a starter control valve pre-failure signature determined responsive to historic gas turbine engine startup data received for the plurality of gas turbine engine startups for one or more gas turbine engines of the certain gas turbine engine model, the gas turbine engine startup data including an indication that at least one of the plurality of gas turbine engine startups failed due to a starter control valve failure;

wherein the one or more gas turbine engines is at least three gas turbine engines;

wherein the operational gas turbine engine startup data includes RPM sensor readings for each of the plurality of gas turbine engine startups of each of the at least three gas turbine engines; and

wherein the computer implemented method further comprises the steps of;

determining the lowest and highest RPM sensor readings of each of the at least three gas turbine engines produced during the plurality of gas turbine engine startups,

determining two median RPM sensor readings,

calculating an average difference between the two median RPM sensor readings and the lowest RPM sensor reading when the lowest RPM sensor reading does not equal the highest RPM sensor reading to thereby define a first difference;

calculating an average difference between the two median RPM sensor readings and the highest RPM sensor reading when the lowest RPM sensor reading does not equal the highest RPM sensor reading to thereby define a second difference, and

identifying one of the at least three gas turbine engines that is associated with the potential predictive starter control valve failure point, the identified one of the at least three gas turbine engines being the gas turbine engine associated with the highest RPM sensor reading when the first difference is greater than the second difference and the identified one of the at least three gas turbine engines being the gas turbine engine associated with the lowest RPM sensor reading when the first difference is less than the second difference.

6. A computer implemented method to predict starter control valve failures in gas turbine engines using a starter control valve health prognostic comprising the steps of:

receiving historic gas turbine engine startup data for a plurality of gas turbine engine startups of one or more historic gas turbine engines of a certain gas turbine engine model, the historic gas turbine engine startup data including an indication that at least one of the plurality of gas turbine engine startups failed due to a starter control valve failure;

determining, in a first computer process by a computer, a starter control valve pre-failure signature for the one or more historic gas turbine engines responsive to the historic gas turbine engine startup data;

extracting, in a second computer process by the computer, one or more health index features from the starter control valve pre-failure signature of the one or more historic gas turbine engines;

constructing, in a third computer process by the computer, a starter control valve health prognostic for the certain gas turbine engine model responsive to the one or more health index features extracted from the starter control valve pre-failure signature of the one or more historic gas turbine engines;

receiving operational gas turbine engine startup data for at least one gas turbine engine startup of one or more operational gas turbine engines of the certain gas turbine engine model;

extracting, in a fourth computer process by the computer, one or more health index features from the operational gas turbine engine startup data for each of the at least one gas turbine engine startups;

calculating, in a fifth computer process by the computer, a starter control valve health index for each of the at least one gas turbine engine startups responsive to the one or more health index features extracted from the operational gas turbine engine startup data to thereby define one or more health indices;

comparing, in a sixth computer process by the computer, each of the one or more health indices calculated by the fifth computer process to the starter control valve health prognostic for the certain gas turbine engine model; and

determining in a seventh computer process by the computer, responsive to the step of comparing, a potential predictive starter control valve failure point, the potential predictive starter control valve failure point being located where one of the one or more health indices correlates to the starter control valve health prognostic for the certain gas turbine engine model.

7. A computer implemented method as defined in claim 6,

wherein the operational gas turbine engine startup data received from the one or more operational gas turbine engines of the certain gas turbine engine model for the at least one gas turbine engine startup is for a plurality of gas turbine engine startups;

wherein the one or more operational gas turbine engines are a plurality of operational gas turbine engines; and

wherein the computer implemented method further comprises the steps of;

constructing a continuous starter control valve deterioration trend function responsive to a plurality of starter control valve health indices calculated for each of the plurality of gas turbine engine startups responsive to the one or more health index features extracted from the operational gas turbine engine startup data,

determining, using the continuous starter control valve deterioration trend function, a point on the continuous starter control valve deterioration trend function that correlates to the starter control valve health prognostic for the certain gas turbine engine model to thereby identify the potential predictive starter control valve failure point, and

identifying, in an eighth computer process by the computer, one of the one or more operational gas turbine engines that is associated with the potential predictive starter control valve failure point determined by the seventh computer process.

8. A computer implemented method as defined in claim 6, further comprising the steps of:

identifying one or more sensors of the certain gas turbine engine model usable to predict failure of a starter control valve of a gas turbine engine of the certain gas turbine engine model prior to an actual event of failure;

parsing the historic gas turbine engine data and the operational gas turbine engine data to select data from the identified one or more sensors usable to predict failure of a starter control valve of a gas turbine engine of the certain gas turbine engine model prior to an event of failure;

segmenting the historic gas turbine engine data and the operational gas turbine engine data to thereby select only gas turbine engine startup data; and

analyzing the historic gas turbine engine startup data and the operational gas turbine engine startup data to reduce irregular gas turbine engine operations and data outliers.

9. A computer implemented method as defined in claim 8,

wherein the historic gas turbine engine startup data is associated with a first vehicle;

wherein the operational gas turbine engine startup data is associated with a second vehicle;

wherein the one or more sensors are legacy sensors;

wherein the identified one or more legacy sensors includes one or more of;

an auxiliary power unit sensor, a plurality of engine start switch sensors, each start switch sensor recording the start switch state for the engine associated with the start switch, and a plurality of gas turbine engine RPM sensors;

wherein the one or more health index features include one or more of;

the total gas turbine engine RPM rise time and one-half of the total gas turbine engine RPM rise time; and

wherein the computer implemented further includes the step of decoding the historic gas turbine engine data and the operational gas turbine engine data from a binary trend file to an ASCII trend file.

10. A computer implemented method as defined in claim 9, wherein the computer implemented method further comprises the steps of:

analyzing operational gas turbine engine RPM sensor data received for the at least one gas turbine engine startup of the one or more operational gas turbine engines of the certain gas turbine engine model;

determining, responsive to analyzing operational gas turbine engine RPM sensor data, the existence of one or more mechanical problems associated with one or more of an auxiliary power unit, a compressor, an air turbine, an ignition system, a fuel system, or a lube pressure system of the aircraft at the potential predictive starter control valve failure point;

determining, responsive to analyzing operational gas turbine engine RPM sensor data, the existence of pilot initiated purging of a fuel injection system of each of the one or more operational gas turbine engines at the potential predictive starter control valve failure point; and

reducing false positive indications of a starter control valve failure responsive determining the existence of one or more mechanical problems and determining the existence of pilot initiated purging of the fuel injection system of each of the one or more operational gas turbine engines at the potential predictive starter control valve failure point.

11. A computer implemented method as defined in claim 10,

wherein the continuous starter control valve deterioration trend function is a function of two variables, a first variable representing one or more of the one or more extracted health index features and a second variable representing the gas turbine engine startup number associated with the one or more of the one or more extracted health index features;

and

wherein a mathematical limit of the continuous starter control valve deterioration trend function approaches zero as the gas turbine engine startup number associated with the one or more of the one or more extracted health index features increases

12. A computer implemented method as defined in claim 10,

wherein the first computer process, the second computer process, and the third computer process are performed by a first computer;

wherein the fourth computer process, the fifth computer process, the sixth computer process, and the seventh computer process are performed by a second computer;

wherein the first vehicle is an aircraft;

wherein the second vehicle is an aircraft; and

wherein the computer implemented method further comprises the steps of;

storing the historic gas turbine engine startup data in a first database associated with the first computer,

storing the one or more health index features in a second database associated with the first computer,

storing the starter control valve health prognostic for the certain gas turbine engine model in a third database associated with the first computer,

storing the operational gas turbine engine startup data in a fourth database associated with the second computer, and

storing the one or more starter control valve health indices in a fourth database associated with the second computer.

13. A starter control valve failure prediction machine to predict and trend starter control valve failures in gas turbine engines using a starter control valve health prognostic, the failure prediction machine comprising:

a first computer readable storage medium storing historic gas turbine engine startup data for a plurality of gas turbine engine startups of one or more historic gas turbine engines of a certain gas turbine engine model, the historic gas turbine engine startup data including an indication that at least one of the plurality of gas turbine engine startups failed due to a starter control valve failure;

a second computer readable storage medium storing operational gas turbine engine startup data for at least one gas turbine engine startup of one or more operational gas turbine engines of the certain gas turbine engine model;

a computer adapted to be in communication with the first computer readable storage medium and the second computer readable storage medium and having a processor, a memory, and a computer readable program product stored in the memory of the computer, the computer readable program product including;

a starter control valve pre-failure signature determiner configured to determine at least one starter control valve pre-failure signature for the one or more historic gas turbine engines responsive to the historic gas turbine engine startup data received from the first computer readable storage medium,

a health prognostic constructor configured to construct a starter control valve health prognostic for the certain gas turbine engine model responsive to one or more health index features extracted from the at least one starter control valve pre-failure signature of the one or more historic gas turbine engines,

a health index determiner configured to determine a starter control valve health index for each of the at least one gas turbine engine startups of the one or more operational gas turbine engines responsive to one or more health index features extracted from the operational gas turbine engine startup data received from the second computer readable storage medium to thereby define one or more health indices, and

a signature comparator configured to compare each of the one or more health indices with the starter control valve health prognostic for the certain gas turbine engine model to thereby identify a potential predictive starter control valve failure point, the potential predictive starter control valve failure point being located where one of the one or more health indices correlates to the starter control valve health prognostic for the certain gas turbine engine model.

14. A starter control valve failure prediction machine as defined in claim 13,

wherein the operational gas turbine engine startup data for the at least one gas turbine engine startup is for a plurality of gas turbine engine startups;

wherein the one or more operational gas turbine engines are a plurality of operational gas turbine engines;

wherein the starter control valve failure prediction machine further comprises a gas turbine engine identifier configured to identify one of the one or more operational gas turbine engines associated with the potential predictive starter control valve failure point; and

wherein the signature comparator is further configured to;

construct a continuous starter control valve deterioration trend function responsive to a plurality of starter control valve health indices calculated for each of the plurality of gas turbine engine startups and responsive to the one or more health index features extracted from the operational gas turbine engine startup data, and

determine, using the continuous starter control valve deterioration trend function, a point on the continuous starter control valve deterioration trend function that correlates to the starter control valve health prognostic for the certain gas turbine engine model to thereby identify the potential predictive starter control valve failure point.

15. A starter control valve failure prediction machine as defined in claim 13,

wherein the historic gas turbine engine startup data is associated with a first vehicle;

wherein the operational gas turbine engine startup data is associated with a second vehicle; and

wherein the starter control valve failure prediction machine further includes an error minimizer configured to minimize false positive indications of a starter control valve failure to thereby verify the potential predictive starter control valve failure point by determining the existence of one or more mechanical problems associated with one or more of an auxiliary power unit, a compressor, an air turbine, an ignition system, a fuel system, or a lube pressure system of the second vehicle at the potential predictive starter control valve failure point, and determining the existence of pilot initiated purging of a fuel injection system of the identified gas turbine engine of the second vehicle at the potential predictive starter control valve failure point.

16. A starter control valve failure prediction machine as defined in claim 15,

wherein the first vehicle and the second vehicle are the same vehicle;

wherein the one or more health index features include one or more of;

the total gas turbine engine RPM rise time and one-half of the total gas turbine engine RPM rise time; and

wherein the first computer readable storage medium and the second computer readable storage medium are both stored together on a third computer readable storage medium and the third computer readable storage medium is associated with the computer.

17. A starter control valve failure prediction machine as defined in claim 15,

wherein the computer further comprises a graphical user interface device and a user interface input device;

wherein the signature comparator is further configured to display via the graphical user interface of the computer the continuous starter control valve deterioration trend function and to receive from a user input via the user interface input device of the computer identifying a potential predictive starter control valve failure point of the second vehicle where a point of the starter control valve deterioration trend function correlates to the starter control valve health prognostic for the certain gas turbine engine model; and

wherein the error minimizer is further configured to display via the graphical user interface device of the computer the potential predictive starter control valve failure point and to receive input via the user interface input device of the computer from the user indicating whether the potential predictive starter control valve failure point is a false positive indication of a starter control valve failure.

18. A starter control valve failure prediction machine as defined in claim 17,

wherein the continuous starter control valve deterioration trend function is a function of two variables, a first variable representing one or more of the one or more extracted health index features and the second variable representing the gas turbine engine startup number associated with the one or more of the one or more extracted health index features;

and

wherein a mathematical limit of the continuous starter control valve deterioration trend function approaches zero as the gas turbine engine startup number associated with the one or more of the one or more extracted health index features increases.

19. A starter control valve failure prediction machine as defined in claim 17,

wherein the starter control valve failure prediction machine further comprises an historic data downloader configured to:

download the historic gas turbine engine startup data from the first computer readable storage medium

parse the historic gas turbine engine data to thereby select data usable to predict failure of a starter control valve of a gas turbine engine of the certain gas turbine engine model prior to an event of failure,

segment the historic gas turbine engine data to certain only gas turbine engine startup data, and

analyze the historic gas turbine engine startup data to reduce irregular gas turbine engine operations and data outliers; and

wherein the starter control valve failure prediction machine further comprises an operational data downloader configured to;

download the operational gas turbine engine startup data from the second computer readable storage medium;

parse the operational gas turbine engine data to thereby select data usable to predict failure of a starter control valve of a gas turbine engine of the certain gas turbine engine model prior to an event of failure,

segment the operational gas turbine engine data to certain only gas turbine engine startup data, and

analyze the operational gas turbine engine startup data to reduce irregular gas turbine engine operations and data outliers.

20. A computer readable program product stored in a tangible computer medium and including instructions that when executed by a computer cause the computer to perform the operations of

determining a starter control valve pre-failure signature for one or more gas turbine engines of a certain gas turbine engine model responsive to historic gas turbine engine startup data received for a plurality of gas turbine engine startups of one or more gas turbine engines of the certain gas turbine engine model, the historic gas turbine engine startup data including an indication that at least one of the plurality of gas turbine engine startups failed due to a starter control valve failure;

constructing a starter control valve health prognostic for the certain gas turbine engine model responsive to one or more health index features extracted from the starter control valve pre-failure signature of the one or more gas turbine engines;

calculating a starter control valve health index for each of a plurality of operational gas turbine engine startups of a plurality of gas turbine engines operationally mounted on a vehicle and of the certain gas turbine engine model responsive to health index features extracted from operational gas turbine engine startup data to thereby define one or more health indices;

comparing each of the one or more health indices to the starter control valve health prognostic for the certain gas turbine engine model;

determining, responsive to the step of comparing, a potential predictive starter control valve failure point, the potential predictive starter control valve failure point being located where one of the one or more health indices correlates to the starter control valve health prognostic for the certain gas turbine engine model; and

identifying the gas turbine engine of the vehicle associated with the potential predictive starter control valve failure point.

21. A computer readable program product as defined in claim 20:

wherein the operational gas turbine engine startup data includes gas turbine engine RPM sensor data; and

wherein the computer readable program product further comprising instructions that when executed by the computer cause the computer to perform the operations of;

analyzing the operational gas turbine engine RPM sensor data;

determining, responsive to analyzing operational gas turbine engine RPM sensor data, the existence of one or more mechanical problems associated with one or more of an auxiliary power unit, a compressor, an air turbine, a ignition system, a fuel system, or a lube pressure system of the vehicle at the potential predictive starter control valve failure point;

determining, responsive to analyzing operational gas turbine engine RPM sensor data, the existence of pilot initiated purging of a fuel injection system of the identified gas turbine engine of the vehicle at the potential predictive starter control valve failure point;

displaying via a graphical user interface device associated with the computer the potential predictive starter control valve failure point, and responsive to the step of determining, the existence of one or more mechanical problems and the existence of pilot initiated purging;

receiving input from a user via a user interface input device associated with the computer indicating whether the potential predictive starter control valve failure point is a false positive indication of a starter control valve failure; and

minimizing false positive indications of a starter control valve failure to thereby verify the potential predictive starter control valve failure point responsive to receiving input from the user via the user interface input device associated with the computer indicating whether the potential predictive starter control valve failure point is a false positive indication of a starter control valve failure.

22. A computer readable program product as defined in claim 20, wherein the computer readable program product further comprises instructions that when executed by the computer cause the computer to perform the operations of:

constructing a continuous starter control valve deterioration trend function responsive to a plurality of starter control valve health indices calculated for each of the plurality of operational gas turbine engine startups and responsive to the one or more health index features extracted from the operational gas turbine engine startup data;

displaying via a graphical user interface of the computer the continuous starter control valve deterioration trend function;

receiving input from a user via a user interface input device of the computer identifying a potential predictive starter control valve failure point where a point on the starter control valve deterioration trend function correlates to the starter control valve health prognostic for the certain gas turbine engine model; and

identifying, responsive to receiving input from the user, the potential predictive starter control valve failure point.

23. A computer readable program product as defined in claim 20,

wherein the continuous starter control valve deterioration trend function is a function of two variables, a first variable representing one or more of the one or more extracted health index features and a second variable representing the gas turbine engine startup number associated with the one or more of the one or more extracted health index features;

wherein a mathematical limit of the continuous starter control valve deterioration trend function approaches zero as the gas turbine engine startup number associated with the one or more of the one or more extracted health index features increases; and

wherein the starter control valve health prognostic is a single value.

24. A computer readable program product as defined in claim 20,

wherein the starter control valve health prognostic includes a plurality of values to thereby create a starter control valve health prognostic threshold;

and

wherein the computer readable program product further comprises instructions that when executed by the computer cause the computer to perform the operation of determining whether each of the one or more health indices correlates to at least one of the plurality of values of the health prognostic threshold to thereby identify the potential predictive starter control valve failure point of the vehicle where one of the one or more health indices correlates to at least one of the plurality of values of the starter control valve health prognostic for the certain gas turbine engine model.

25. A computer readable program product as defined in claim 20,

wherein the starter control valve health prognostic is a rate of declination;

and

wherein the computer readable program product further comprises instructions that when executed by the computer cause the computer to perform the operations of;

determining a health index rate of declination for the one or more starter control valve health indices, and

determining whether the health index rate of declination is greater or equal to the starter control valve health prognostic to thereby identify a potential predictive starter control valve failure point of the vehicle where the health index rate of declination is greater or equal to the starter control valve health prognostic.