System and method for monitoring a gas turbine

Abstract

System (20) for monitoring the behavior of a gas turbine in use (10), comprising: a storage element having a database (26) with a fleet model containing collected historical parameter information about the operation, repairs and / or maintenance of comparable gas turbines, the fleet model being configured to enable projections of parameter information during future loads using the collected historical parameter information; an input device (34), the input device (34) being configured to generate a unit data signal (30) containing parameter information from the gas turbine in use (10) and a risk signal (32) containing a risk value for the gas turbine in use (10); and a processor (22) in communication with the memory element and the input device (34), the processor (22) being configured to input the input data signal (30) into the database (26) with the collected parameter information from comparable gas turbines, to update the fleet model using the parameter information from the in-service gas turbine (10) to generate a unit model, to generate future-projected parameter information for the in-service gas turbine (10) using the unit model, and to combine the future-projected parameter information from the unit model with the risk signal (32) to calculate a conditional risk that the future-projected parameter information for the in-service gas turbine (10) will reach a predetermined parameter limit;and, based on the calculated conditional risk, generate an output signal (42) containing at least one of repair and maintenance planning information;

Description

FIELD OF THE INVENTION

[0001] The present invention generally encompasses a system and method for monitoring the health of a gas turbine. More specifically, the present invention describes a system and method that adapts a generic gas turbine model using actual information about an individual gas turbine to project or predict future repair and / or maintenance intervals for the individual gas turbine. BACKGROUND TO THE INVENTION

[0002] Gas turbines are widely used in industrial and commercial applications. As in Fig. 1, a typical gas turbine in use 10 includes a compressor 12, particularly an axial-flow compressor at the front, one or more combustors 14 approximately in the middle, and a turbine 16 at the rear. The compressor 12 includes multiple stages of rotating blades and stationary vanes. Ambient air enters the compressor 12, and the rotating blades and stationary vanes impart increasing kinetic energy to the working fluid (the air) to bring it to an energetic state. The working fluid exits the compressor 12 and flows to the combustors 14, where it mixes with a fuel 18 and ignites to produce combustion gases having a high temperature, high pressure, and high velocity. The combustion gases exit the combustors 14 and flow to the turbine 16, where they expand to perform work.

[0003] Gas turbines, like any other mechanical device, require periodic repairs and maintenance to ensure proper functioning. As a general approach, past experience with the "fleet" of gas turbines, particularly comparable gas turbines of a similar class or design, can be statistically analyzed to develop a fleet model that can project the expected wear and damage experienced by other gas turbines into the future. Based on the fleet model, projections, or forecasts, can be made, and repairs and maintenance activities can be scheduled at optimal intervals, minimizing the risk of both unplanned shutdowns for repairs and unnecessary shutdowns for unnecessary preventive maintenance.

[0004] However, the actual behavior of individual gas turbines may differ from the fleet model. For example, individual gas turbines may have slight differences in configuration, manufacturing tolerances, and design, which may result in different levels of wear and damage compared to the fleet model. Furthermore, the actual operating, repair, and maintenance histories of individual gas turbines may differ from the fleet average. For example, gas turbines operating in humid and corrosive environments may require more frequent repair and maintenance actions compared to the fleet model to address problems related to corrosion, pitting, and emissions. Conversely, other gas turbines that experience fewer start-up and shutdown cycles may require less frequent shutdowns to perform preventative maintenance related to cyclic loading compared to the fleet model.In each example, adjustments to the fleet model based on actual data associated with individual gas turbines would improve the ability to optimally plan repair and maintenance activities.

[0005] US 2006 / 0 265 183 A1 describes a system and method for monitoring the behavior and functioning of a gas turbine engine of an aircraft in use. During a flight, real-time flight parameters are recorded using sensors, and snapshot data representing the parameters during specific flight phases is extracted from the recorded flight parameters. The snapshot data is then used to evaluate the difficulty of the flight. In particular, for certain components with a limited lifetime, a K-factor is determined from the snapshot data. This K-factor represents a measure of the difficulty of the flight, i.e., it characterizes the lifetime usage of the respective component consumed during the flight.The K-factor data for each individual limited-life component is summed and compared to specified limits for the component according to a U.S. Federal Aviation Administration guideline to determine when the specific limited-life component should be repaired or replaced.

[0006] Furthermore, an improved system and method for monitoring the behavior and functioning of a gas turbine in use would be desirable. BRIEF DESCRIPTION OF THE INVENTION

[0007] Aspects and advantages of the invention are set forth below in the following description, or may be obvious from the description, or may be learned by practice of the invention.

[0008] According to one aspect of the present invention, a system for monitoring the behavior of a gas turbine in use is provided. The system includes a storage element having a database with a fleet model containing collected historical parameter information about the operation, repairs, and / or maintenance activities of comparable gas turbines, the fleet model being configured to enable projections of parameter information during future stresses using the collected historical parameter information, and an input device, the input device being configured to generate a unit data signal containing parameter information from the gas turbine in use and a risk signal containing a risk value for the gas turbine in use.A processor, communicatively coupled to the memory element and the input device, is configured to: incorporate the unit data signal into the database of collected parameter information from comparable gas turbines, update the fleet model using the parameter information from the in-service gas turbine to generate a unit model, generate future projected parameter information for the in-service gas turbine using the unit model, and combine the future projected parameter information from the unit model with the risk signal to calculate a conditional risk that the future projected parameter information for the in-service gas turbine will reach a predetermined parameter limit.The processor is further configured to generate an output signal containing at least one of repair and maintenance planning information based on the calculated conditional risk.

[0009] According to another aspect of the present invention, a method for monitoring the behavior of a gas turbine in use is provided.The method includes: receiving parameter information from comparable gas turbines from a fleet model that includes collected historical parameter information about the operation, repairs and / or maintenance of comparable gas turbines, the fleet model being configured to enable projections of parameter information during future stresses using the collected historical parameter information, receiving parameter information from the in-service gas turbine and a risk signal that includes a risk value for the in-service gas turbine, adding the parameter information from the in-service gas turbine to the parameter information from comparable gas turbines to update the fleet model using the parameter information from the in-service gas turbine to produce a unit model, and forecasting orProjecting the parameter information for the in-service gas turbine into the future using the unit model. The method further includes calculating a conditional risk that the projected parameter information for the in-service gas turbine will reach a predetermined parameter limit by combining the projected parameter information from the unit model with the risk signal, and generating an output signal containing at least one of a repair and / or a maintenance plan for the in-service turbine based on the calculated conditional risk.

[0010] Those skilled in the art will better understand the features and aspects of such and other embodiments after reviewing the specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] A complete and practical disclosure of the present invention, including the best mode thereof, to those skilled in the art is set forth in greater detail in the remainder of the specification, which includes reference to the accompanying figures, in which: Fig. 1 illustrates a simplified block diagram of a typical gas turbine system; Fig. 2 illustrates a functional block diagram of a system for monitoring a gas turbine in use according to an embodiment of the present invention; Fig. 3 illustrates an algorithm for updating and validating a fleet model; Fig. 4 illustrates an algorithm for updating and validating a unit model; Fig. 5 illustrates an algorithm for performing a unit risk analysis; Fig. 6 illustrates an algorithm for calculating the remaining useful life of a part or component; Fig. 7 graphically illustrates hypothetical loss progression curves that may be generated by a unit risk analysis according to an embodiment of the present invention; and Fig. 8 graphically illustrates hypothetical service life curves according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0012] Reference will now be made in detail to the present embodiments of the invention, one or more examples of which are illustrated in the accompanying drawings. The detailed description uses numerical and letter designations to refer to features in the drawings. Like or similar reference numerals in the drawings and the description are used to designate like or similar parts of the invention.

[0013] Each example is intended to illustrate the invention, not to limit the invention. Indeed, it will be apparent to those skilled in the art that modifications and changes can be made to the present invention without departing from its scope or spirit. For example, features illustrated or described as part of one embodiment may be utilized in another embodiment to yield a still further embodiment. Thus, the present invention is intended to include such modifications and changes as come within the scope of the appended claims and their equivalents.

[0014] The systems and methods discussed herein refer to processors, servers, memory, databases, software applications, and / or other computer-based systems, as well as actions performed on or taken by such systems, and information sent to and from such systems. One skilled in the art will recognize that the inherent flexibility of computer-based systems allows for a wide variety of possible configurations, combinations, and divisions of tasks and functionality among the components. For example, computer-implemented processes described herein may be implemented using a single server or processor, or multiple such elements operating in combination with one another. Databases and other storage / media elements and applications may be implemented on a single system or distributed among multiple systems.Distributed components may operate sequentially or in parallel. All such variations are intended to be within the scope of the present subject matter, as would be understood by those skilled in the art.

[0015] When data is received or accessed between a first and a second computer system, a first and a second processing device, or components thereof, the actual data may be exchanged directly or indirectly between the systems. For example, if a first computer accesses a file or data from a second computer, the access may involve one or more intermediary computers, proxies, or the like. The actual file or data may be passed between the computers, or a computer may provide a pointer or metafile that the second computer uses to access the actual data from a computer other than the first computer.

[0016] The various computer systems described herein are not limited to any particular hardware architecture or configuration. Embodiments of the methods and systems provided herein may be implemented by one or more general-purpose or custom computing devices configured in any suitable manner to provide the desired functionality. The device(s) may be configured to provide additional functionality that is either complementary to or unrelated to the present subject matter. For example, one or more computing devices may be configured to provide the described functionality by accessing software instructions embodied in computer-readable form.When software is used, any suitable programming language, scripting language, or other suitable language, or combinations of languages, may be used to implement the teachings contained herein. However, software need not be used exclusively or not at all. As will be understood by those skilled in the art, without requiring further detailed explanation, some embodiments of the methods and systems provided and disclosed herein may also be implemented using hardwired logic or other circuitry, including, but not limited to, application-specific circuitry. Of course, various combinations of computer-executed software and hardwired logic or other circuitry may also be suitable.

[0017] It will be understood by those skilled in the art that embodiments of the methods disclosed herein may be performed by one or more suitable computing devices that enable the device(s) to perform such methods. As noted above, such devices may access one or more computer-readable media containing computer-readable instructions that, when executed by at least one computer, cause the at least one computer to perform one or more embodiments of the methods according to the present subject matter.Any suitable computer-readable medium or media may be used to implement or carry out the subject matter disclosed herein, including, but not limited to, floppy disks, drives and other magnetic-based storage media, optical storage media including discs (including CD-ROMs, DVD-ROMs and variants thereof), flash memory, RAM, ROM memory and other semiconductor storage devices, and the like.

[0018] Condition-based maintenance systems apply stochastic analyses of fleet models, unit-specific data, and operator-selected risk parameters to create a cost-effective system and process for optimizing repair and / or maintenance intervals for high-value systems, such as gas turbines. A fleet model for each specific failure or damage mechanism for a gas turbine can be developed by applying multi-stage stochastic modeling techniques, such as Bayesian perturbation and Markov Chain Monte Carlo (MCMC) simulation, to historical fleet data.The accuracy of each fleet model can be periodically verified and / or validated, and unit-specific data obtained from a specific gas turbine can be added to each fleet model to adjust or update the fleet model and generate a unit model that more accurately models the specific gas turbine for each specific failure or damage mechanism. Applying operator-selected risk parameters to the updated fleet model improves the ability to schedule repair and / or maintenance units at optimal intervals, increasing operational availability, reducing unplanned and unnecessary outages, and / or extending the useful life of the specific gas turbine.

[0019] For example, if the unit-specific data for the particular gas turbine indicates less wear or damage compared to the future projections (forecasts) provided by the fleet model, the availability of the particular gas turbine can be improved by extending the intervals between repair and / or maintenance events. Conversely, if the unit-specific data for the particular gas turbine indicates greater wear or damage compared to the projections provided by the fleet model, the intervals between repair and / or maintenance events can be reduced, resulting in a planned outage instead of the more costly unplanned outage.In any case, the adapted repair and / or maintenance plan improves the reliability, operational safety and functioning of the specific gas turbine, resulting in more accurate utilization and possibly a longer service life for the specific gas turbine.

[0020] Fig. 2 shows a system 20 for monitoring a field gas turbine 10 according to an embodiment of the present invention. The term "field gas turbine" refers to a particular or specific gas turbine as distinct from the fleet of gas turbines. The system 20 generally includes a processor 22 containing programming or coding to access one or more storage / media elements. The processor 22 receives a fleet model signal 24 from a database 26 and a fleet data signal 28, a unit data signal 30, and / or a risk signal 32 from an input device 34. The term "signal" refers to any electrical transmission of information or data. The fleet model signal 24 includes parameter information for comparable gas turbines projected into the future by a fleet model contained in the database 26.The system 20 applies multi-stage stochastic modeling techniques, Bayesian perturbation and MCMC simulation as described by block 36 and the method shown in . Fig. 3, to verify and validate the projected parameter information contained in the fleet model signal 24 and generate an updated fleet model signal 33. The system 20 adds parameter information from the gas turbine in use 10 contained in the unit data signal 30 to the updated fleet model signal 33 to generate an updated fleet model, preferably referred to as a unit model, which is executed by block 38 and the algorithm illustrated in Fig. 4. The unit model generates future projected parameter information 41 for the gas turbine in use 10. A unit risk analysis performed by the unit 40 in Fig. 2 and the Fig. 5 and Fig. 6, combines the projected parameter information 41 from the unit model with the risk signal 32 to produce an output signal 42 representing plans for repair 44 and / or maintenance 46 and / or a useful life forecast 48 for the gas turbine in service 10.

[0021] The processor 22 discussed herein is not limited to any particular hardware architecture or configuration. Rather, the processor 22 may comprise a general-purpose or custom computing device configured to access storage media (e.g., blocks 36, 38, and / or 40 in Fig. 2), databases, and other hardware to provide the described functionality in a manner guided by software instructions represented in computer-readable form or by programmed circuitry. For example, processor 22 may comprise a single server, a single microprocessor, hard-wired logic, including, but not limited to, application-specific circuitry, or multiple such elements operating in combination.

[0022] Database 26 contains historical parameter information about the "fleet" of gas turbines, particularly comparable gas turbines of a similar class or design, collected from available sources. Database 26 may contain storage / media elements and applications implemented on a single system or distributed across multiple systems. If distributed components are used, they may operate sequentially or in parallel.

[0023] The historical parameter information contained in database 26 includes data reflecting the operation, repairs, and / or maintenance actions of the comparable gas turbines. The historical parameter information may specifically include data referred to as stress data and damage data. Stress data includes any information describing the operational history of a comparable gas turbine that can be statistically associated with the prediction of a failure mode or mechanism. For example, stress data may include operating hours, number of start-up and shutdown cycles, firing temperatures, and the number of unscheduled trips. Damage data includes any hardware failure mechanisms that have occurred with statistical significance.A failure mechanism encompasses any deterioration in physical or functional characteristics from nominal values ​​that results in a reduction in output power, a loss of efficiency, or the inability to operate the comparable gas turbine. Examples of known failure mechanisms include corrosion, pitting, deformation, fatigue, foreign object damage, oxidation, thermal barrier coating (TBC) spalling, plugging / contamination, fracture, cracking, and wear. These failure mechanisms may be detected or recorded as a result of enhanced borescope inspections, on-site monitoring, operational logs, repair logs, maintenance logs, and the like.

[0024] Available sources of historical information include, for example, operational experience databases, operational records, part inspection records, and field inspection reports. Examples of historical information contained in these sources include, but are not limited to, enhanced borescope inspection reports, electronic records, monitoring and diagnostic data, downtime event reports, operating durations, start-ups, and tripping reports, as well as maintenance shop or repair data.

[0025] The collection of historical information, such as stress and damage data, is statistically analyzed and normalized to develop the fleet model, also referred to as a data accumulation model. The fleet model projects or predicts parameter information, such as damage growth, during future stresses using the collected historical information, and the fleet model and / or the projected parameter information is transmitted to the processor 22 via the fleet model signal 24.

[0026] Input device 34 enables a user to communicate with system 20 and may include any structure to provide an interface between the user and system 20. For example, input device 34 may include a keyboard, a computer, a terminal, a tape drive, and / or any other device for receiving input from a user and generating fleet data signal 28, unit data signal 30, and / or risk signal 32 for system 20.

[0027] Fig. 3 shows an algorithm for updating and validating the fleet model and / or the fleet model signal 24, referred to above in the form of block 36 according to Fig. 2. In block 50, the algorithm imports the fleet data signal 28, which includes, for example, newly acquired parameter information from comparable gas turbines in the fleet, such as stress data 52 and damage data 54. For illustrative purposes only, assume that the fleet data signal 28 indicates that, with 10,000 hours of operation, 20 start-up and shutdown cycles, and two unscheduled trips, borescope inspections in a particular component detected cracks of sizes 0.1, 0.2, 0.1, 0.2, 0.3, and 0.2. In block 56, the algorithm sorts the imported stress data 52 and damage data 54 and organizes them, for example, by assigning a variable L nto each inspection result in ascending order according to the amount of damage detected to produce the following result: L1=0.1, L2=0.1, L3=0.2, L4=0.2, L5=0.2 and L6=0.3. In block 58, the algorithm groups the sorted stress data 52 and damage data 54, e.g., by assigning a variable R n to each inspection result by the same amount to produce the following result: R1=2 / 6, R2=2 / 6, R3=3 / 6, R4=3 / 6, R5=3 / 6, and R6=1 / 6. In block 60, the algorithm compares the sorted and grouped data 52, 54 with the fleet model signal 24, which includes distribution parameter information, such as the projected damage results based on the fleet model, to determine whether the fleet model is statistically accurate. Statistical accuracy can be measured using many individual or combined statistical criteria, including, for example, the value of the coefficient of determination (R 2) or the standard deviation (σ). If the comparison indicates that the fleet model provides a statistically accurate projection of the actual damage, block 62, the algorithm then updates the database 26 of historical parameter information with the newly acquired parameter information from comparable gas turbines in the fleet and provides the updated fleet model signal 33 for further analysis. The updated fleet model becomes the unit model if indicated by the Fig. 4 is accessed. If the comparison indicates that the fleet model does not provide a statistically accurate projection of the actual damage, the algorithm then generates a flag or other signal in block 66 indicating the need to investigate the error between the fleet model projections and the actual damage data.

[0028] Fig. Figure 4 shows an algorithm for updating and validating the unit model referred to above as block 38 in Fig. 2. In block 68, the algorithm imports the unit data signal 30, which includes, for example, newly acquired parameter information from the gas turbine in service 10, such as stress data 70 and damage data 72. For illustrative purposes only, assume again that the unit data signal 30 indicates that, with 10,000 hours of operation, 20 start-up and shutdown cycles, and two unscheduled trips, borescope inspections in a particular component detected cracks of sizes 0.1, 0.3, 0.1, 0.3, 0.3, and 0.2. In block 74, the algorithm sorts the imported unit data 70, 72 and organizes them, for example, by assigning a variable L nto each inspection result in ascending order according to the amount of damage detected to produce the following result: L1=0.1, L2=0.1, L3=0.2, L4=0.3, L5=0.3, and L6=0.3. In block 76, the algorithm groups the sorted unit data 70, 72, for example, by assigning a variable R n to each inspection result by the same amount to produce the following result: R1=2 / 6, R2=2 / 6, R3=1 / 6, R4=3 / 6, R5=3 / 6, and R6=3 / 6. In block 78, the algorithm compares the sorted and grouped unit data 70, 72 with the unit model, which contains distribution parameter information, such as the projected damage results based on the unit model, to determine whether the unit model is statistically accurate. Statistical accuracy can be measured by a number of individual or combined statistical criteria, including, for example, the value of the coefficient of determination (R 2) or the standard deviation (σ). If the comparison indicates that the unit model provides a statistically accurate projection of the actual damage, block 80, the algorithm then updates the unit model with the newly acquired parameter information from the in-service gas turbine 10 and generates updated parameter information 41 from the unit model for further analysis. If the comparison indicates that the unit model does not provide a statistically accurate projection of the actual damage, the algorithm then generates a flag 84 or other signal indicating the need to investigate the error between the projections and the actual damage.

[0029] Fig. Figure 5 shows an algorithm for performing the unit risk analysis referred to above in the form of block 40 in Fig. 2. The unit risk analysis combines the updated parameter information 41 from the unit model with the risk signal 32 to generate the output signal 42 representing the repair plans 44 and / or maintenance 46 and / or the useful life projection or forecast 48 for the in-service gas turbine 10. In block 86, the algorithm imports the risk signal 32, which includes, for example, unit stress data, allowable risk levels for each failure mechanism, and / or the next maintenance interval for the in-service gas turbine 10. In block 88, the algorithm imports the updated parameter information 41 from the unit model, which includes, for example, the distribution unit parameter information, e.g., the projected damage results based on the unit model. In block 90, the algorithm loads or accesses risk analysis equations associated with each failure mechanism.The risk analysis equations may use any of several techniques known in the art for modeling distribution curves of future states based on known data. For example, the risk analysis equations may use a Weibull loglinear model, a Weibull proportional damage model, or a lognormal loglinear model.

[0030] In block 92, the algorithm calculates a conditional risk associated with each particular failure mechanism using the risk analysis equations. The conditional risk is the probability that a unit parameter will reach or exceed a predetermined limit at some time in the future. The predetermined parameter limit may be any condition, metric, measure, or other criteria specified by the user. For example, the predetermined parameter limit may be an operational limit, such as a crack size, of a part or component, which, if exceeded, may require user action such as performing an additional inspection, removing the part or component from service, repairing the part or component, or limiting the performance of the gas turbine in service 10.The future time may be the next inspection interval for the gas turbine in service 10, measured chronologically by operating hours, starts, shutdowns, unscheduled trips, or any other stress data provided by the user and related to the failure mechanism.

[0031] In block 94, the algorithm calculates the operational reliability at the current state of the gas turbine in service 10. The calculated reliability is the probability that a part or component will be able to successfully perform its intended function(s) at the rated limits at least until some future time. In other words, the calculated reliability is the probability that a part or component will not fail due to an identified failure mechanism before some future time.As with the conditional risk calculation, the future time may be the next inspection interval for the gas turbine in service 10, as measured chronologically, by operating hours, starts, shutdowns, unscheduled trips, or any other stress data provided by the user and related to the failure mechanism.

[0032] In block 96, the algorithm calculates the remaining useful life for the part or component, and Fig. Figure 6 shows an algorithm for performing this calculation. In blocks 98 and 100, the algorithm imports the risk signal 32 and the updated parameter information 41, respectively, as described above in connection with blocks 86 and 88 in Fig. 5. In block 102, the algorithm calculates the mean damage value for each particular failure mechanism for the gas turbine in service 10. In block 104, the algorithm calculates the probability that the part or component will reach or exceed a predetermined operational limit at various future stress times (e.g., operating hours, start-ups, shutdowns, unscheduled trips, etc.). In block 106, the algorithm calculates the most limiting stress time based on the allowable risk level provided by the user for each failure mechanism.If, using the data presented in the previous illustrative examples, the user provides an acceptable risk level of 5% for the crack size and the predetermined operational limit for the crack size is 0.5, block 106 of the algorithm calculates the stress time at which the conditional risk that a crack size of 0.5 will be present is 5%. In block 108, the algorithm calculates the remaining service life of the part or component based on the difference between the current stress time and the most limiting stress time calculated in block 106.

[0033] Returning to Fig. 5, the unit risk analysis algorithm generates the output signal 42, which reflects the results of the unit risk analysis. For example, the output signal 42 may include repair plans 44 and / or maintenance plans 46 and / or a useful life projection for the gas turbine in service 10 or a component therein.

[0034] Fig. Figure 7 graphically illustrates hypothetical damage history curves that may be generated by the unit risk analysis algorithm according to an embodiment of the present invention. The horizontal axis represents the stress interval (e.g., the operating hours, starts, shutdowns, unscheduled trips, or any other stress data associated with a failure mechanism) between outages for repair and / or maintenance, and the vertical axis represents the degree of damage to a part or component in the in-service gas turbine 10. A horizontal line above the graph represents the predetermined parameter limit 110 or operational limit of a part or component, as set by the user.

[0035] Each curve on the graph in Fig. 7 represents a hypothetical damage progression curve. For example, the curve labeled 112 reflects a 5% risk according to the fleet model that a part or component without detected damage will exceed the predetermined parameter limit 110 before the stress interval labeled 114. The curve labeled 116 reflects a 95% risk according to the fleet model that a part or component without detected damage will exceed the predetermined parameter limit 110 before the stress interval labeled 118. The curve labeled 120 reflects a 5% risk according to the updated fleet model or unit model that a part or component without detected damage will exceed the predetermined parameter limit 110 before the stress interval labeled 122.The curve labeled 124 reflects a 95% risk, according to the updated fleet model or unit model, that a part or component will exceed the predetermined parameter limit 110 without detected damage before the stress interval labeled 126. The various data points labeled 128 represent actual inspection results, variously referred to above as unit parameter information or damage data 72, which are communicated to the processor 22 via the unit data signal 30. Referring again to FIG. Fig. 2, this damage data 72 is added to the unit model in block 38 to generate the updated parameter information 41. The unit risk analysis combines the updated parameter information 41 with information in the risk signal 32 to determine the actual risk curve for the gas turbine in service 10.

[0036] Fig. Figure 8 graphically illustrates hypothetical service life curves generated by the algorithm described above in connection with Fig. 6. In this illustration, the horizontal axis represents the stress limit in operating hours, and the vertical axis represents the stress limit for start-ups. Other stress limits may be applicable depending on various factors, such as the failure mechanism, the specific part or component, the stress data for the gas turbine in service 10, etc. The curve labeled 130 represents a hypothetical service life curve for a part or component for a particular failure mechanism. Point 132 represents a design service life for a part or component for a given combination of start-ups and operating hours. The curve labeled 134 represents a new service life curve for the part or component, as determined by blocks 106 and 108 in Fig.6. As illustrated, the new service life curve 134 shows the increased number of starts and hours of operation that the part or component may have before the failure mechanism occurs.

[0037] This description uses examples to disclose the invention, including the best mode, and also to enable any person skilled in the art to practice the invention, including making and using any devices or systems and performing any incorporated methods. The patentable scope of the invention is defined by the claims and may include other examples that occur to those skilled in the art. Such other examples are intended to be within the scope of the claims if they include structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insubstantial differences from the literal language of the claims.

[0038] A system 20 for monitoring a field gas turbine 10 includes a database 26 containing information from comparable gas turbines and an input device 34 that generates a unit data signal 30 and a risk signal 32. A processor 22, communicatively coupled to the memory and input device 34, inputs the unit data signal 30 into the database 26, projects information for the field gas turbine 10 into the future, and calculates a conditional risk that the field gas turbine 10 will reach a limit. An output signal 42 contains repair or maintenance schedules. A method for monitoring a field gas turbine 10 includes receiving information from comparable gas turbines, adding information from the field gas turbine 10 to the information from comparable gas turbines, and projecting information for the field gas turbine 10 into the future.The method further includes calculating a conditional risk that the gas turbine in use 10 will reach a limit and generating an output signal 42 containing repair or maintenance schedules. Parts list: 10 gas turbines in use 12 compressors 14 Combustion chamber 16 turbines 20 systems 22 processor 24 Fleet model signal 26 Database 28 Fleet data signal 30 standard data signal 32 Risk signal 33 updated fleet model signal 34 Input device 36 Fleet Model Verification Block 38 Unit model generation block 40 Unit Risk Analysis Block 41 Parameter information from the unit model 42 Output signal 44 Repair plan 46 Maintenance plan 48 Useful life projection / forecast 50 Fleet data import block 52 Fleet usage data 54 Fleet damage data 56 Fleet data sorting block 58 Fleet data grouping block 60 Fleet comparison block 62 Fleet Update Block 66 Fleet Flag Block 68 Unit data import block 70 unit stress data 72 unit damage data 74 Unit data sorting block 76 Unit data grouping block 78 Unit Comparison Block 80 Unit Update Block 84 Unit Flag Block 86 Input block of the unit risk analysis 88 Import block of the uniform risk analysis 90 Risk analysis equation loading block of the unit risk analysis 92 Conditional risk calculation block of the unit risk analysis 94 Reliability calculation block of the unit risk analysis 96 Useful life calculation block of the unit risk analysis 98 Input block for the useful life calculation 100 import block for the useful life calculation 102 Damage mean value calculation block for the service life calculation 104 Failure probability calculation block for the useful life calculation 106 Calculation of downtime at risk for the useful life calculation 108 Calculation of the remaining useful life for the useful life calculation 110 operational limit 112 5% fleet curve 114 5% fleet limit 116 95% fleet curve 118 95% fleet limit 120 5% unit curve 122 5% unit limit 124 95% unit curve 126 95% unit limit 128 inspection results 130 hypothetical service life curve 132 points 134 new service life curve

Claims

[1] A system (20) for monitoring the behavior of a gas turbine in use (10), comprising: a storage element having a database (26) with a fleet model containing collected historical parameter information about the operation, repairs and / or maintenance of comparable gas turbines, the fleet model being configured to enable projections of parameter information during future loads using the collected historical parameter information; an input device (34), the input device (34) being configured to generate a unit data signal (30) containing parameter information from the gas turbine in use (10) and a risk signal (32) containing a risk value for the gas turbine in use (10); and a processor (22) in communication with the memory element and the input device (34), the processor (22) being configured to input the input data signal (30) into the database (26) with the collected parameter information from comparable gas turbines, to update the fleet model using the parameter information from the in-service gas turbine (10) to generate a unit model, to generate future-projected parameter information for the in-service gas turbine (10) using the unit model, and to combine the future-projected parameter information from the unit model with the risk signal (32) to calculate a conditional risk that the future-projected parameter information for the in-service gas turbine (10) will reach a predetermined parameter limit;and, based on the calculated conditional risk, generate an output signal (42) containing at least one of repair and maintenance planning information; [2] The system (20) of claim 1, further comprising a fleet model signal (24) between the database (26) and the processor (22), the fleet model signal (24) containing the parameter information of comparable gas turbines. [3] The system (20) of any of claims 1-2, wherein the unit data signal (30) includes data representative of at least one of operation and / or repairs and / or maintenance of the gas turbine in service (10). [4] The system (20) of any of claims 1-3, wherein the processor (22) is arranged to generate the output signal (42) based on a comparison of the calculated conditional risk with the risk value. [5] The system (20) of any of claims 1-4, wherein the output signal (42) includes a projected useful life of a component in the gas turbine in service (10). [6] A method for monitoring the behavior of a gas turbine in use (10), comprising: (a) receiving parameter information from comparable gas turbines from a fleet model containing collected historical parameter information on the operation, repairs and / or maintenance of comparable gas turbines, the fleet model being configured to enable projections of parameter information during future loads using the collected historical parameter information; b) receiving parameter information from the gas turbine in use (10) and a risk signal (32) containing a risk value for the gas turbine in use (10); c) adding the parameter information from the gas turbine in use (10) to the parameter information from comparable gas turbines to update the fleet model using the parameter information from the gas turbine in use (10) to create a unit model; d) projecting the parameter information for the gas turbine in use (10) into the future using the unit model; e) calculating a conditional risk that the projected parameter information for the gas turbine in use (10) will reach a predetermined parameter limit by combining the future projected parameter information from the unit model with the risk signal (32); and f) generating an output signal (42) containing at least one of a repair plan and a maintenance plan for the gas turbine in use (10) based on the calculated conditional risk. [7] The method of claim 6, further comprising comparing the conditional risk with a predetermined risk value. [8] The method of claim 7, further comprising delaying a repair or maintenance if the conditional risk is less than the predetermined risk value. [9] The method of claim 7, further comprising bringing forward a repair or maintenance if the conditional risk is not less than the predetermined risk value.

Images