A gas turbine unit monitoring method and system based on a proxy model

Abstract

The application relates to a gas turbine unit monitoring method and system based on an agent model, which comprises the following steps: establishing a simulation model of a gas turbine unit; simulation model calibration: determining parameters for monitoring the performance of the gas turbine unit and performance indexes thereof; agent model generation: for each monitoring performance parameter, an agent model is generated in the following manner: selecting multiple simulation operation condition parameters, combining them into multiple different simulation operation condition conditions, simulating and calculating by using the simulation model of the gas turbine unit to obtain the monitoring performance parameters corresponding to each simulation operation condition condition; correcting multiple actual monitoring performance parameters corresponding to each exhaust temperature to eliminate the influence of boundary conditions and obtain corrected monitoring performance parameters; obtaining a function relationship formula between the corrected monitoring performance parameter reference values and the exhaust temperature by fitting; monitoring model establishment: obtaining the performance indexes corresponding to each monitoring performance parameter and monitoring the performance indexes of each monitoring performance parameter.

Description

Technical Field

[0001] This invention relates to the field of gas turbine technology, and more specifically to a gas turbine unit monitoring method based on a surrogate model. Background Technology

[0002] Gas turbines are widely used in natural gas transmission, power generation, and marine applications, playing a crucial role as core components. However, due to their complex structure and harsh operating conditions, they are prone to failure within the entire system. A malfunction can lead to a decline in unit performance, and in severe cases, require shutdown for maintenance, impacting the power plant's economic viability. Therefore, to improve operational stability and reduce accidents, online monitoring of gas turbines and early warning systems for potential faults are essential.

[0003] There are many methods for monitoring the condition of gas turbines, such as monitoring the temperature of turbine blades, monitoring the condition of gas turbines through sound, monitoring the condition of gas turbine lubrication systems based on oil spectral analysis, monitoring gas turbine performance, and monitoring gas turbine vibration. Conventional online monitoring of units generally requires the installation of an additional online monitoring system, involving repeated iterative calculations, which is time-consuming and prone to non-convergence. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the technical problem to be solved by the present invention is to provide a gas turbine unit monitoring method based on a proxy model, which can realize the online monitoring function of gas turbine without adding additional measuring points, the model calculation is convenient and fast, no need for repeated iterative calculations, and the accuracy is high.

[0005] To achieve the above objectives, this invention provides a gas turbine unit monitoring method based on a surrogate model, comprising the following steps:

[0006] S1. Establish a simulation model of the gas turbine unit;

[0007] S2. Simulation Model Calibration: The simulation model is calibrated using actual gas turbine unit performance test data to ensure that the accuracy of the simulation model meets the requirements.

[0008] S3. Determine the parameters used to monitor the performance of the gas turbine unit, denoted as monitoring performance parameters, including gas turbine power, gas turbine efficiency, compressor pressure ratio, and compressor isentropic efficiency index, and determine the performance index corresponding to the monitoring performance parameters:

[0009] S4. Proxy Model Generation: Let X represent each monitoring performance parameter. For each monitoring performance parameter, a proxy model is generated in the following manner:

[0010] S41. Set the operating condition range, including the range of exhaust temperature EGT, the range of ambient temperature T, the range of ambient pressure P, the range of compressor inlet guide vane opening IGV, and the range of ambient relative humidity Phi.

[0011] S42. Select N1 EGTs within the operating condition range. s N2 T s N3 P s , N4 IGV s and N5 Phi s These parameters are used as simulation operating condition parameters and combined to form N. 总 N different simulation operating conditions 总 =N1*N2*N3*N4*N5, using the simulation model of the gas turbine unit to perform simulation calculations, the monitoring performance parameters corresponding to each simulated operating condition are obtained, and denoted as the actual monitoring performance parameter X. s Thus, N is obtained. 总 A set of simulation calculation data {EGT s T s P s IGV s Phi s X s};

[0012] S43, for each EGT s The corresponding multiple actual monitoring performance parameters X s After making corrections to eliminate the influence of boundary conditions, the corresponding corrected monitoring performance parameters are obtained, denoted as X. c EGT s The corresponding X c Constituting the data group {EGT s X c}; The exhaust gas temperature EGT for each simulation calculation is obtained. s The corresponding corrected monitoring performance parameter X c This results in multiple data sets {EGT} s X c};

[0013] S44. Using a polynomial fitting method, for all data sets {EGT} s X c By fitting the data, the corrected baseline value X of the monitoring performance parameter is obtained. cstd The functional relationship between X and exhaust temperature cstd =f{EGT}, thus obtaining the baseline value X of the corrected monitoring performance parameter at any exhaust temperature. cstd ;

[0014] S5. Monitoring Model Establishment: Using the proxy model from step S4, obtain the performance index X corresponding to each monitoring performance parameter. index =[(X c -X cstd ) / X cstd *100%; Real-time monitoring of the performance index corresponding to various monitoring performance parameters, reflecting the deviation of the gas turbine unit from the ideal state, and conducting online monitoring of the unit.

[0015] Furthermore, in step S2, after a certain period of operation, the simulation model needs to be recalibrated using the actual operating data of the gas turbine unit.

[0016] Further, in step S41, the exhaust temperature EGT ranges from 520℃ to 620℃, the ambient temperature T ranges from -15℃ to 45℃, the ambient pressure P ranges from 0.95 bar to 1.05 bar, the compressor inlet guide vane opening IGV ranges from 0% to 100%, and the ambient relative humidity Phi ranges from 0% to 100%.

[0017] Furthermore, in step S42, N1, N2, N3, N4, and N5 are all greater than or equal to 5.

[0018] Furthermore, in step S42, the simulation calculation results are organized to obtain each EGT. s The corresponding multiple actual monitoring performance parameters X s And it is displayed in the form of a chart.

[0019] Further, in step S43, for each EGT s The corresponding multiple actual monitoring performance parameters X s The methods for making corrections include:

[0020] S431, From all simulation calculation data sets {EGTs, T s P s IGV s Phi s X s Select to include the EGT. s The simulation calculation data set is denoted as EGT. s The runtime parameter data group {EGTs, T s P s IGV s Phi s X s};

[0021] S432. The EGT is corrected through four single-factor corrections: environmental pressure correction, environmental temperature correction, IGV correction, and environmental relative humidity correction. s The corresponding multiple actual monitoring performance parameters X s Make corrections, and each single-factor correction is the same, including the following steps:

[0022] S4321. Determine the data group before this correction, denoted as the data group to be corrected. If this correction is the first correction, then the data group to be corrected is the running parameter data group {EGTs, T} of the EGTs selected in step S431. s P s IGV s Phi s X s If this is not the first revision, then the data set to be revised is the data set obtained after the last single-factor revision.

[0023] S4322, T s P s IGV s Phi s Of these four factors, the factor related to this correction is represented by Y, and the other three factors together are represented by M. All data sets to be corrected are organized, and data with the same M but different Y are grouped into a set of fitting calculation data sets. Each set of fitting calculation data sets includes multiple data sets.

[0024] S4323. Use the fitting calculation data set to perform fitting calculation, and compare the actual monitoring performance parameters corresponding to different Y values ​​in the same fitting calculation data set with the benchmark monitoring performance parameters X under the benchmark operating conditions. 基 The ratio is processed, and then Y is used as the independent variable and the ratio is used as the dependent variable. Mathematical methods are used to fit the data to obtain the functional relationship between the ratio and Y. Then, based on the functional relationship between the ratio and Y, the actual monitoring performance parameters in each data group to be corrected are corrected, and the corrected actual monitoring performance parameters are calculated. Then, the actual monitoring performance parameters in the data group to be corrected are replaced with the corrected actual monitoring performance parameters to obtain the corrected data group.

[0025] Further, in step S4323, the reference operating conditions include: ambient temperature 15°C, ambient pressure 1 bar, compressor inlet guide vane opening 100%, and ambient relative humidity 60%.

[0026] The present invention also provides a gas turbine unit monitoring system based on a proxy model, including an operation data acquisition device and a monitoring and processing module. The operation data acquisition device acquires the operation parameter data of the gas turbine unit and transmits the relevant operation data to the monitoring and processing module. The monitoring and processing module stores a gas turbine monitoring computer program, which, when executed, can realize the above-mentioned gas turbine unit monitoring method.

[0027] As described above, the gas turbine unit monitoring method and system of the present invention have the following beneficial effects:

[0028] A novel monitoring method is proposed, featuring convenient and rapid model calculations that eliminate the need for repeated iterations, exhibit easy convergence, and high accuracy. On one hand, it can monitor deviations from benchmark values ​​in real time by calculating four key performance indices—gas turbine power, gas turbine efficiency, compressor pressure ratio, and compressor isentropic efficiency—reflecting the unit's operational status. On the other hand, different gas turbine faults have varying impacts on these performance indices; different threshold combinations can be set to provide early warnings of unit faults, offering reference suggestions for power plant operators. This gas turbine unit monitoring method can be based on the power plant's existing control system, eliminating the need for additional monitoring software and saving costs. Attached Figure Description

[0029] Figure 1 This is a flowchart illustrating the gas turbine unit monitoring method based on a proxy model according to the present invention.

[0030] Figure 2 This is a graph showing the relationship between the actual gas turbine power and exhaust temperature obtained from the simulation calculations in this invention.

[0031] Figure 3 This is a graph showing the relationship between the actual gas turbine efficiency and exhaust temperature obtained from the simulation calculations in this invention.

[0032] Figure 4 This is a graph showing the relationship between the actual compressor pressure ratio and exhaust temperature obtained from the simulation calculations in this invention.

[0033] Figure 5 This is a graph showing the relationship between the actual compressor isentropic efficiency and exhaust temperature obtained from simulation calculations in this invention.

[0034] Figure 6 This is a graph showing the relationship between the corrected gas turbine power reference value and the exhaust temperature obtained by fitting in this invention.

[0035] Figure 7 This is a graph showing the relationship between the corrected gas turbine efficiency benchmark value and the exhaust temperature obtained by fitting in this invention.

[0036] Figure 8This is a graph showing the relationship between the corrected compressor pressure ratio reference value and the exhaust temperature obtained by fitting in this invention.

[0037] Figure 9 This is a graph showing the relationship between the corrected compressor isentropic efficiency benchmark value and the exhaust temperature obtained by fitting in this invention. Detailed Implementation

[0038] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification.

[0039] It should be understood that the structures, proportions, sizes, etc., depicted in the accompanying drawings of this specification are merely for illustrative purposes to aid those skilled in the art in understanding and reading the content disclosed herein, and are not intended to limit the conditions under which the invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in proportions, or adjustments to the size, without affecting the effects and objectives achieved by the invention, should still fall within the scope of the technical content disclosed herein. Furthermore, the terms such as "upper," "lower," "left," "right," and "middle" used in this specification are merely for clarity and are not intended to limit the scope of the invention. Changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention's implementation.

[0040] See Figures 1 to 9 This invention provides a gas turbine unit monitoring method based on a surrogate model, comprising the following steps:

[0041] S1. Establish a simulation model of the gas turbine unit, whose main components include the compressor, combustion chamber, turbine, and intake and exhaust modules. The simulation model can be established using modular and mechanistic modeling methods, which are conventional approaches and will not be elaborated upon further.

[0042] S2. Simulation Model Calibration: The simulation model is calibrated using performance test data from actual gas turbine units to ensure that the accuracy of the simulation model meets the requirements. During calibration, preferably, operating conditions with relatively extreme boundary conditions during actual unit operation are selected: high relative humidity, low relative humidity, high ambient pressure, low ambient pressure, high ambient temperature, and low ambient temperature. When the deviations of the simulation model's calculation results at non-design points are all less than the required values, for example, all deviations are less than 1%, the established simulation model of the gas turbine unit is considered to have high accuracy, meets the needs of engineering applications, and can be used to generate subsequent proxy models.

[0043] In this embodiment, as a preferred design, considering the aging effect of the gas turbine unit, the simulation model is recalibrated using the actual operating data of the gas turbine unit after a certain period of operation (e.g., 10,000 operating hours) to ensure that the simulation model can truly reflect the performance of the unit.

[0044] S3. Determine the parameters used to monitor the performance of the gas turbine unit, denoted as monitoring performance parameters, including turbine power PW and turbine efficiency η. e Compressor pressure ratio PR and compressor isentropic efficiency η c And determine the performance index corresponding to the monitored performance parameters.

[0045] Generally, parameters reflecting the performance status of a gas turbine unit include: gas turbine power, gas turbine efficiency, compressor pressure ratio, compressor isentropic efficiency, turbine inlet temperature, turbine efficiency, and exhaust temperature. However, some of these parameters lack operational measurement points and cannot be directly measured. In actual operation, the power level of the gas turbine unit can be adjusted by regulating the exhaust temperature, which indirectly affects other performance parameters. Therefore, it can be considered that there is a certain coupling relationship between exhaust temperature and other parameters, and it can be used as an independent variable for performance parameters. Therefore, the following performance indices can be selected for monitoring performance parameters:

[0046] Gas turbine power index: It represents the deviation between the actual corrected output power and the corresponding reference value at a given exhaust temperature, and is used to reflect the condition of the entire gas turbine unit.

[0047] Gas turbine efficiency index: It represents the deviation between the actual corrected gas turbine efficiency and the corresponding reference value at a given exhaust temperature, and is used to reflect the condition of the entire gas turbine unit.

[0048] Compressor pressure ratio index: It represents the deviation between the actual corrected compressor pressure ratio and the corresponding reference value at a given exhaust temperature, and is used to reflect the condition of the compressor and the possible decrease in flow capacity.

[0049] Compressor isentropic efficiency index: It represents the deviation between the actual corrected compressor isentropic efficiency and the corresponding reference value at a given exhaust temperature, and is used to reflect the condition of the compressor.

[0050] S4. Proxy Model Generation:

[0051] In this invention, surrogate model generation refers to the process of calculating the actual corrected values ​​and corresponding reference values ​​of monitored performance parameters. Gas turbine power PW, gas turbine efficiency η e Compressor pressure ratio PR and compressor isentropic efficiency η c The proxy models for these four monitoring performance parameters are generated in the same way. For ease of explanation, let X represent each monitoring performance parameter, that is, let X refer to PW, η e PR or ηc For each monitored performance parameter, a proxy model is generated in the following manner:

[0052] S41. Set the operating condition range, including the range of exhaust temperature EGT, ambient temperature T, ambient pressure P, compressor inlet guide vane opening IGV, and ambient relative humidity Phi. Preferably, the exhaust temperature EGT range is 520℃~620℃, the ambient temperature T range is -15℃~45℃, the ambient pressure P range is 0.95bar~1.05bar, the compressor inlet guide vane opening IGV range is 0~100%, and the ambient relative humidity Phi range is 0~100%. This operating condition range can cover the actual environmental conditions of a conventional power plant. Furthermore, fuel composition is generally considered relatively stable; the actual fuel composition can be used as input during simulation.

[0053] S42. Select N1 EGTs within the operating condition range. s N2 T s N3 P s , N4 IGV s and N5 Phi s These parameters are used as simulation operating condition parameters and combined to form N. 总 N different simulation operating conditions 总 =N1*N2*N3*N4*N5, using the simulation model of the gas turbine unit to perform simulation calculations, the monitoring performance parameters corresponding to each simulated operating condition are obtained, and denoted as the actual monitoring performance parameter X. s Thus, N is obtained. 总 A set of simulation calculation data {EGT s T s P s IGV s Phi s X s Simultaneously, the simulation results are organized to obtain each EGT. s The corresponding multiple actual monitoring performance parameters X s And it can be displayed in the form of charts.

[0054] In this embodiment, preferably, N1, N2, N3, N4, and N5 are all greater than or equal to 5, resulting in a suitable amount of simulation data. When selecting simulation runtime parameters, N1 EGTs are used. s N2 T s N3 P s , N4 IGV s and N5 Phi s When selecting, it is preferable to elect all at equal intervals, for example, 5 exhaust temperature EGTs.s Divided into 520℃, 545℃, 570℃, 595℃ and 620℃, see [link / reference]. Figures 2 to 5 .

[0055] Through simulation calculations, the gas turbine power PW and gas turbine efficiency η were obtained. e Compressor pressure ratio PR and compressor isentropic efficiency η c The simulation results for these four performance parameters are as follows. Among them, the gas turbine power PW and gas turbine efficiency η are... e The compressor pressure ratio PR can be obtained directly or by acquiring relevant parameters and calculating it, which is a conventional method, so it will not be described in detail here.

[0056] In this invention, the compressor isentropic efficiency is not a directly measured value. It needs to be calculated based on conventional measurement data such as compressor inlet temperature, compressor inlet pressure, compressor outlet temperature, and compressor outlet pressure to obtain the compressor isentropic efficiency under the current operating conditions. The compressor isentropic efficiency module mainly involves the following parameters: compressor inlet temperature T... cin Compressor inlet pressure P cin Compressor outlet temperature T cout Compressor outlet pressure P cout The isentropic efficiency η of the compressor c =(h couts -h1) / (h cout -h1), where: η c The compressor isentropic efficiency, expressed as a percentage (%); h couts h1 is the isentropic enthalpy at the compressor outlet, in kJ / kg; h2 is the enthalpy at the compressor inlet, in kJ / kg; h3 is the isentropic enthalpy at the compressor outlet, in kJ / kg; h4 is the isentropic enthalpy at the compressor outlet, in kJ / kg; h5 is the isentropic enthalpy at the compressor outlet, in kJ / kg; h6 is the isentropic enthalpy at the compressor outlet, in kJ cout The value represents the compressor outlet enthalpy, expressed in kJ / kg. The compressor inlet pressure P can be obtained by operating the measuring points. cin and compressor outlet pressure P cout Using the formula The compressor pressure ratio was calculated. Additionally, the compressor inlet temperature T was obtained through operational measuring points. cin and compressor outlet temperature T cout By substituting the known temperature into the solution for the enthalpy, the enthalpy at the compressor inlet (h1) and the enthalpy at the compressor outlet (h) can be obtained. couts The method for determining the enthalpy value from a known temperature can employ existing, well-established methods, such as: "Calculation and Analysis of Thermal Properties of Working Fluids in Gas Turbines" (Huang Di, Zhou Dengji, et al., Journal of Naval University of Engineering, 2017, 12). Separately, determine... In the formula: h1 is the compressor inlet enthalpy, in kJ / kg; γ is the adiabatic index of air, which is generally taken as 1.4; R is the gas constant of air, in J / (kg·K); T cin π represents the compressor inlet temperature in Kelvin (K). CLet be the compressor pressure ratio. Substituting the parameters obtained above into the formula, the current compressor isentropic efficiency η can be calculated. c The calculation of the compressor's isentropic efficiency can be performed automatically by setting up a calculation program.

[0057] The simulation results of gas turbine power PW are compiled and presented in graphical form to obtain different EGT. s The corresponding actual gas turbine power PW s See Figure 2 As shown. Regarding the gas turbine efficiency η e The simulation results were compiled and presented in chart form to obtain different EGT values. s The corresponding actual gas turbine efficiency η es See Figure 3 As shown, the simulation results of the compressor pressure ratio PR are compiled and presented in graphical form to obtain different EGT. s The corresponding actual compressor pressure ratio PR s See Figure 4 As shown. The isentropic efficiency η of the compressor. c The simulation results were compiled and presented in chart form to obtain different EGT values. s The corresponding actual compressor isentropic efficiency η cs See Figure 5 As shown.

[0058] The calculation results show that at each exhaust temperature, due to differences in boundary conditions (including ambient temperature T, ambient pressure P, compressor inlet guide vane opening IGV, and ambient relative humidity Phi), the actual monitored performance parameter X... s The relationship between these parameters is not one-to-one; instead, multiple actual monitoring performance parameters are obtained. Therefore, the generation process of the surrogate model first requires correcting the actual monitoring performance parameters. By using a fitting method, the measured monitoring performance parameters are corrected to the parameters corresponding to the baseline operating conditions, eliminating the influence of different boundary conditions, in order to obtain a one-to-one correspondence between exhaust temperature and performance parameters.

[0059] S43. For each EGTs, there are multiple actual monitoring performance parameters X. s After making corrections to eliminate the influence of boundary conditions, the final corrected monitoring performance parameter is denoted as X. c Multiple actual monitoring performance parameters X s The corresponding corrected monitoring performance parameters may be the same or different, therefore an EGT s Corresponding to one or more different X c, However, they all fluctuate within a small range around a basic value, EGT s Its corresponding X cConstituting one or more data sets {EGT s X c}; Get each EGT s The corresponding corrected monitoring performance parameter X c This results in multiple data sets {EGT} s X c}

[0060] In this embodiment, preferably, for each EGT s The corresponding multiple actual monitoring performance parameters X s The methods for making corrections include:

[0061] S431, From all simulation calculation data sets {EGT s T s P s IGV s Phi s X s Select to include the EGT. s The simulation calculation data set is denoted as EGT. s The running parameter data group {EGT s T s P s IGV s Phi s X s}, used for this EGT s The corresponding actual monitoring performance parameter X s Make corrections for each EGT s Both have multiple running parameter data groups {EGT s T s P s IGV s Phi s X s}

[0062] S432. The EGT is corrected through four single-factor corrections: environmental pressure correction, environmental temperature correction, IGV correction, and environmental relative humidity correction. s The corresponding multiple actual monitoring performance parameters X s Corrections are made, and each single-factor correction aims to eliminate the influence of one of the following factors—environmental pressure, ambient temperature, IGV, or relative humidity—on the actual monitoring performance parameter X. sThe order of the environmental pressure factor correction, environmental temperature factor correction, IGV factor correction, and environmental relative humidity factor correction is not required. These four single-factor corrections can be performed sequentially. The first single-factor correction is based on the simulation calculation results of step S42 above, and subsequent single-factor corrections are based on the previous single-factor corrections.

[0063] The method for each single-factor adjustment is the same, including the following steps:

[0064] S4321. Determine the data group before this correction, and denot it as the data group to be corrected. If this correction is the first correction, then the data group to be corrected is the EGT selected in step S431. s The running parameter data group {EGT s T s P s IGV s Phi s X s If this is not the first revision, then the data set to be revised is the data set obtained after the last single-factor revision.

[0065] S4322, T s P s IGV s and Phi s Of these four parameters, the factor relevant to this correction is represented by Y, and the other three factors together are represented by M. For example, when this correction is for environmental stress factors, P... s Let Y represent P s IGV s and Phi s Let M represent all the data sets to be corrected. All data sets with the same M but different Y values ​​are grouped into a single set for fitting calculations. Each set of fitting calculation data sets contains multiple data sets.

[0066] S4323. Use the fitting calculation data set to perform fitting calculation, and compare the actual monitoring performance parameters corresponding to different Y values ​​in the same fitting calculation data set with the benchmark monitoring performance parameters X under the benchmark operating conditions. 基 The ratio is processed, and then Y is used as the independent variable, with the ratio as the dependent variable. A mathematical fitting method is used to obtain the functional relationship between the ratio and Y. Then, based on this functional relationship, the actual monitoring performance parameters are corrected using Y and the actual monitoring performance parameters in each data set to be corrected. The corrected actual monitoring performance parameters are calculated, and then all the actual monitoring performance parameters in the data sets to be corrected are replaced with the corrected monitoring performance parameters to obtain the corrected data sets. In this invention, the ambient temperature T in the baseline operating conditions... 基Environmental pressure P 基 IGV (Inlet Guide Vane Opening) of Compressor 基 and ambient relative humidity Phi 基 It can be set according to the actual situation. Preferably, it is set to P. 基 =1 bar, Phi 基 =60, IGV 基 =100%, T 基 =15℃.

[0067] The EGT s Each running parameter data group {EGT s T s P s IGV s Phi s X s After four single-factor corrections—environmental pressure correction, environmental temperature correction, IGV correction, and environmental relative humidity correction—the corresponding corrected monitoring performance parameter X is finally obtained. c Thus, the corrected data set {EGTs, T} is obtained. s P s IGV s Phi s X c}, different corrected data sets {EGT s T s P s IGV s Phi s X c X in} c They may be the same or different, but they all fluctuate within a small range around a base value, that is, for a certain exhaust temperature EGT s It is possible to obtain multiple different X values. c .

[0068] In this embodiment, the monitoring performance parameter X is corrected. c Compared with actual monitoring performance parameter X s The fitting function relationship between them can be denoted as X. c =f{P s Phi s IGV s T s X s At the same exhaust temperature, EGT s Below, X c With X s The ratio between them and P s Phi s IGV s and T sFour conditions are related; therefore, in this embodiment, X is set. c =X s / (f1×f2×f3×f4), where f1 is the environmental pressure correction factor, f2 is the environmental relative humidity correction factor, f3 is the IGV correction factor, and f4 is the environmental temperature correction factor. In this embodiment, the order of environmental pressure correction, environmental temperature correction, IGV correction, and environmental relative humidity correction is taken as an example, and the specific explanation is as follows:

[0069] A. Correction of environmental stress factors, including the following steps:

[0070] A1. Determine the data group before this correction, and denot it as the data group to be corrected. Since this is the first correction, the data group to be corrected is the EGT selected in step S431. s The running parameter data group {EGT s T s P s IGV s Phi s X s If this is not the first revision, then the data set to be revised is the data set obtained after the last single-factor revision.

[0071] A2. Organize all data sets to be corrected, and those with the same T s Phi s and IGV s And different P s The data is categorized into a set of fitting calculation data groups. For simplicity, we will use M instead of {T}. s Phi s IGV s}, then each fitted calculation data set obtained includes N3 data sets {EGTs, M, P}. s X s}, and can obtain multiple sets of fitting calculation data.

[0072] A3. Use the data set for fitting calculations to perform fitting calculations, and combine different P values ​​from the same data set. s Corresponding actual monitoring performance parameter X s Benchmark monitoring performance parameter X under benchmark operating conditions 基 The ratio is processed, and then environmental pressure is used as the independent variable, with the ratio as the dependent variable. Conventional mathematical methods are used to fit the relationship between the ratio and environmental pressure, yielding a functional relationship. Specifically, in this embodiment, X is set... c =X sWhen / (f1×f2×f3×f4), since the fitting calculation is performed on all data sets in the dataset, T is used to calculate T. s Phi s and IGV s When f2 × f3 × f4 remain constant, the functional relationship f1 = f(P) can be obtained by fitting the data using existing mathematical fitting methods. For example, linear fitting can be used, resulting in f1 = k1 × P + k2, where k1 and k2 are constants. The ambient temperature T in the reference operating condition... 基 Environmental pressure P 基 IGV (Inlet Guide Vane Opening) of Compressor 基 and ambient relative humidity Phi 基 It can be set according to the actual situation, for example, P 基 =1 bar, Phi 基 =60%, IGV 基 =100%, T 基 =15℃.

[0073] Then, based on the functional relationship between the ratio and environmental pressure, the data sets {EGT} to be corrected are processed sequentially. s T s P s IGV s Phi s X s X in} s All are corrected: using the data set to be corrected {EGT} s T s P s IGV s Phi s X s The environmental pressure P in} s and actual monitoring performance parameters X s The corrected monitoring performance parameters were calculated and, for ease of explanation, are denoted as X. s修 just- P In this embodiment, that is, X s Correction -P =f1×X s =(k1×P s +k2)×X s Then, the data group to be corrected {EGT} s T s P s IGV s Phi s X s X in} s Replace all with X s修正-P Thus, the corrected data set {EGT} is obtained. sT s P s IGV s Phi s X s修正-P After correcting the actual monitoring performance parameters in each data set to be corrected using the above method, the resulting X values ​​are... s修正-P They may be the same or they may be different.

[0074] After correction using the above method, the environmental pressure effect on the monitoring performance parameters was eliminated in the obtained data set.

[0075] B. Correction for relative humidity is basically the same as the correction for environmental pressure above, and includes the following steps:

[0076] B1. Determine the data set before this correction, denoted as the data set to be corrected. Since this is the second correction, the data set to be corrected is the data set obtained after the previous environmental stress factor correction, i.e., {EGT}. s T s P s IGV s Phi s X s修正-P}

[0077] B2. Organize all data sets to be corrected, and those with the same T s P s and IGV s And different Phi s The data is categorized into a set of fitting calculation data groups. Specifically, for simplicity, we will use M instead of {T}. s P s IGV s}, then each fitted calculation data set obtained includes N5 data sets {EGTs, M, Phi}. s X s}, and can obtain multiple sets of fitting calculation data.

[0078] The fitting calculation is performed using a set of data for fitting calculation, and different Phi values ​​in the same set of data for fitting calculation are... s The corresponding monitoring performance parameters and the benchmark monitoring performance parameters X under the benchmark operating conditions. 基 The ratio is processed, and then the relative humidity is used as the independent variable, with the ratio as the dependent variable. Conventional mathematical methods are used to fit the relationship between the ratio and the relative humidity, yielding a functional relationship. Specifically, in this embodiment, X is set... c =X sWhen / (f1×f2×f3×f4), since the fitting calculation is performed on all data sets in the dataset, T is used to calculate T. s P s IGV s When the values ​​are the same, that is, f1×f3×f4 remain unchanged, the functional relationship f2=f(Phi) can be obtained by fitting the data using existing mathematical fitting methods. For example, linear fitting can be used to obtain f2=k3×Phi+k4, where k3 and k4 are constants.

[0079] Then, based on the functional relationship between the ratio and the ambient relative humidity, the data sets {EGT} to be corrected are... s T s P s IGV s Phi s X s修正-P All are corrected: using Phi from the data group to be corrected. s and X s修正-P The corrected monitoring performance parameters are calculated and, for ease of explanation, can be denoted as X. s修正-Phi In this embodiment, that is, X s修正-Phi =f2×X s修正-P =(k3×Phi) s +k4)×X s修正-P Then, replace all the original monitoring performance parameters in the data group to be corrected with X. s修正-Phi That is, the {EGT} obtained above s T s P s IGV s Phi s X s修正-P} updated to {EGT s T s P s IGV s Phi s X s修正-Phi}, thus obtaining the corrected data set. Through the above method, after correcting the actual monitoring performance parameters in each data set to be corrected, the resulting X values ​​are... s修正-Phi They may be the same or they may be different.

[0080] After correction using the above method, the influence of ambient relative humidity on monitoring performance parameters is eliminated in the obtained data set.

[0081] C. IGV factor correction includes the following steps:

[0082] C1. Determine the data set before this correction, denoted as the data set to be corrected. Since this is the third correction, the data set to be corrected is the data set obtained after the previous correction for the environmental relative humidity factor, i.e., {EGT}. s T s P s IGV s Phi s X s修正-Phi}

[0083] C2. Organize all data sets to be corrected, and those with the same T s P s and Phi s And different IGV s The data is categorized into a set of fitting calculation data groups. For simplicity, we will use M instead of {T}. s P s Phi s}, then each fitted calculation data set obtained includes N4 data sets {EGTs, M, Phi}. s X s}, and can obtain multiple sets of fitting calculation data.

[0084] The fitting calculation is performed using a set of data for fitting calculation, and different IGV values ​​are calculated from the same set of data for fitting calculation. s The corresponding monitoring performance parameters and the benchmark monitoring performance parameters X under the benchmark operating conditions. 基 The ratio is processed, and then the compressor inlet guide vane opening is used as the independent variable, with the ratio as the dependent variable. Conventional mathematical methods are used for fitting to obtain the functional relationship between the ratio and the compressor inlet guide vane opening. Specifically, in this embodiment, X is set... c =X s When / (f1×f2×f3×f4), since the fitting calculation is performed on all data sets in the dataset, T is used to calculate T. s P s and Phi s When the values ​​are the same, that is, f1×f2×f4 remain unchanged, the functional relationship f3=f(IGV) can be obtained by fitting with an existing suitable mathematical fitting method. For example, linear fitting can be used to obtain f3=k5×IGV+k6, where k5 and k6 are constants.

[0085] Then, based on the functional relationship between the ratio and the compressor inlet guide vane opening, the data sets {EGT} to be corrected are... s T s P s IGV s Phi s X s修正-PhiAll are corrected: using the IGV data in the data set to be corrected. s and X s修正-Phi The corrected monitoring performance parameters are calculated and, for ease of explanation, can be denoted as X. s修正-IGV In this embodiment, that is, X s修正-IGV =f3×X s修正-Phi =(k5×IGV) s +k6)×X s修正-Phi Then, replace all the actual monitoring performance parameters in the data set to be corrected with X. 修正-IGV Specifically, in this embodiment, the {EGT} obtained above is... s T s P s IGV s Phi s X s修正-Phi} updated to {EGT s T s P s IGV s Phi s X s修正-IGV}, thus obtaining the corrected data set. After correcting the actual monitoring performance parameters in each data set to be corrected, the resulting X values ​​are... s修正-IGV They may be the same or they may be different.

[0086] After the above corrections, the influence of the compressor inlet guide vane opening on the monitoring performance parameters was eliminated in the resulting data set.

[0087] D. Correction for ambient temperature factors, including the following steps:

[0088] D1. Determine the data set before this correction, denoted as the data set to be corrected. Since this is the fourth correction, the data set to be corrected is the data set obtained after the previous IGV factor correction, i.e., data set {EGT}. s T s P s IGV s Phi s X s修正-IGV}

[0089] D2. Organize all data sets to be corrected, and those with the same P s IGV s and Phi s And different T s The data is categorized into a set of fitting calculation data groups. Specifically, for simplicity, we will use M instead of {P}. s IGV s Phi s}, then each fitted calculation data set obtained includes N2 data sets {EGTs, M, Phi}. s X s}, and can obtain multiple sets of fitting calculation data.

[0090] The fitting calculation is performed using a set of data for fitting calculation, and different T values ​​are calculated from the same set of data for fitting calculation. s The corresponding monitoring performance parameters and the benchmark monitoring performance parameters X under the benchmark operating conditions. 基 The ratio is processed, and then the ambient temperature is used as the independent variable, and the ratio is used as the dependent variable. Conventional mathematical methods are used to fit the relationship between the ratio and the ambient temperature to obtain a functional relationship. Specifically, in this embodiment, X is set... c =X s When / (f1×f2×f3×f4), due to the fitting calculation of P in all data sets in the data set. s IGV s and Phi s When the values ​​are the same, that is, f1×f2×f3 remain unchanged, the functional relationship f4=f(T) can be obtained by fitting with an existing suitable mathematical fitting method. For example, linear fitting can be used to obtain f4=k7×T+k8, where k7 and k8 are constants.

[0091] Then, based on the functional relationship between the ratio and the ambient relative humidity, the data sets {EGT} to be corrected are... s T s P s IGV s Phi s X s修正-Phi All are corrected: using T from the data set to be corrected. s and X s修正-Phi The corrected monitoring performance parameters are calculated and, for ease of explanation, can be denoted as X. s修正-T In this embodiment, that is, X s修正-T =f4×X s修正-Phi =(k7×T) s +k8)*X s修正-Phi Then replace and modify all monitoring performance parameters in the data group to be corrected with X. s修正-T Specifically, in this embodiment, the {EGT} obtained above is... s T s P s IGV s Phi s X s修正-IGV} updated to {EGT s T s P s IGV sPhi s X s修正-T}, thus obtaining the revised set of operating parameter data.

[0092] After this correction, the influence of ambient relative humidity on the actual monitoring performance parameters has been eliminated in the resulting set of operating parameter data.

[0093] After undergoing the above four single-factor corrections, the EGT s Each corresponding runtime parameter data group {EGT s T s P s IGV s Phi s X s X in} s After all corrections, X was obtained. c Specifically, in this embodiment, X c =X s修正-T Furthermore, the EGT s The corresponding X c There may be multiple different values, but they are all within a certain range and are used in subsequent mathematical fitting calculations.

[0094] Using the same method, the exhaust temperature EGT was calculated for N1 simulations. s The corresponding actual monitoring performance parameter X s All are corrected to obtain each EGT. s The corresponding corrected monitoring performance parameter X c This forms multiple data groups {EGT} s X c} can be represented as points in a chart, see [link / reference]. Figures 6 to 9 As shown.

[0095] S44. Through mathematical fitting, for all data sets {EGT} s X c By fitting the data, the corrected baseline value X of the monitoring performance parameter is obtained. cstd The functional relationship between X and exhaust temperature cstd = f(EGT). This allows us to obtain the baseline value X of the corrected monitoring performance parameter at any exhaust temperature. cstd In this embodiment, X is preferably obtained through polynomial fitting. cstd =f(EGT)=K1+K2×EGT-K3×EGT 2 +K4×EGT 2.5 -K5×EGT 3 Where K1, K2, K3, K4, and K5 are polynomial constants. For a given exhaust temperature EGT... sIt is possible to obtain multiple different X values. c However, they all fluctuate within a small range around a basic value, which will not affect the accuracy of the fitting results.

[0096] In this embodiment, the gas turbine power PW and gas turbine efficiency η are obtained through the above-described proxy model generation method. e Compressor pressure ratio PR and compressor isentropic efficiency η c For the proxy models of these four performance monitoring parameters, see [link to relevant documentation]. Figures 6 to 9 As shown.

[0097] S5. Monitoring Model Establishment: Using the proxy model from step S4, obtain the performance index X corresponding to each monitoring performance parameter. index =[(X c -X cstd ) / X cstd *100%. In this embodiment, it specifically includes: gas turbine power index PW index =[(PW c -PW cstd ) / PW cstd *100%; Gas turbine efficiency index: η eindex =[(η ec -η ecstd ) / η ecstd *100%, Compressor pressure ratio index PR index =[(PR c -PR cstd ) / PR cstd *100%, compressor isentropic efficiency index η cindex =[(η cc -η ccstd ) / η ccstd ]*100%

[0098] The monitoring model mainly achieves online monitoring of the gas turbine unit in two aspects: (a) By monitoring the performance index corresponding to various monitoring performance parameters in real time, it reflects the deviation of the gas turbine unit from the ideal state and the operating status of the unit. If the performance index remains stable for a period of time, it indicates that the unit is operating stably. If a sudden change occurs, it indicates that a fault has occurred in the unit, and an alarm is issued; (b) When different faults occur in the gas turbine, such as compressor fouling, excessive pressure loss in the intake duct, increased blade tip clearance, and reduced turbine efficiency, the degree of impact on each performance index is also different. Therefore, different faults of the unit can be characterized by setting different combinations of performance index thresholds, thereby realizing early warning of unit faults.

[0099] This invention also provides a gas turbine unit monitoring system based on a surrogate model, including an operation data acquisition device and a monitoring and processing module. The operation data acquisition device collects operation parameter data of the gas turbine unit, which mainly includes the actual power of the gas turbine PW, ambient temperature T, ambient pressure P, compressor inlet guide vane opening IGV, ambient relative humidity Phi, natural gas LHV, and natural gas flow rate m. f Compressor inlet temperature T cin Compressor inlet pressure P cin Compressor outlet temperature T cout Compressor outlet pressure P cout The collected operational data is transmitted to the monitoring and processing module. This module has storage and execution functions. The gas turbine monitoring computer program includes the monitoring model established using the aforementioned gas turbine unit monitoring method. When executed, the program calculates the performance index X corresponding to each monitoring performance parameter based on the operational parameter data collected by the data acquisition device. index .

[0100] The gas turbine unit monitoring method and system of the present invention have the following beneficial effects:

[0101] A novel monitoring method is proposed, featuring convenient and rapid model calculations that eliminate the need for repeated iterations, exhibit easy convergence, and high accuracy. On one hand, it can monitor deviations from benchmark values ​​in real time by calculating four key performance indices—gas turbine power, gas turbine efficiency, compressor pressure ratio, and compressor isentropic efficiency—reflecting the unit's operational status. On the other hand, different gas turbine faults have varying impacts on these performance indices; different threshold combinations can be set to provide early warnings of unit faults, offering reference suggestions for power plant operators. This gas turbine unit monitoring method can be based on the power plant's existing control system, eliminating the need for additional monitoring software and saving costs.

[0102] In summary, this invention effectively overcomes the various shortcomings of the prior art and has high industrial application value.

[0103] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A gas turbine unit monitoring method based on a surrogate model, characterized in that: Includes the following steps: S1. Establish a simulation model of the gas turbine unit; S2. Simulation Model Calibration: The simulation model is calibrated using actual gas turbine unit performance test data to ensure that the accuracy of the simulation model meets the requirements. S3. Determine the parameters used to monitor the performance of the gas turbine unit, denoted as monitoring performance parameters, including gas turbine power, gas turbine efficiency, compressor pressure ratio, and compressor isentropic efficiency index, and determine the performance index corresponding to the monitoring performance parameters: S4. Proxy Model Generation: Let X represent each monitoring performance parameter. For each monitoring performance parameter, a proxy model is generated in the following manner: S41. Set the operating condition range, including the range of exhaust temperature EGT, the range of ambient temperature T, the range of ambient pressure P, the range of compressor inlet guide vane opening IGV, and the range of ambient relative humidity Phi. S42. Select N1 EGTs within the operating condition range. s N2 T s N3 P s , N4 IGV s and N5 Phi s These parameters serve as simulation operating condition parameters and are combined to form multiple N values. 总 N different simulation operating conditions 总 =N1*N2*N3*N4*N5, using the simulation model of the gas turbine unit to perform simulation calculations, the monitoring performance parameters corresponding to each simulated operating condition are obtained, and denoted as the actual monitoring performance parameter X. s Thus, N is obtained. 总 A set of simulation calculation data {EGT s T s P s IGV s Phi s X s }; S43, for each EGT s The corresponding multiple actual monitoring performance parameters X s After making corrections to eliminate the influence of boundary conditions, the corresponding corrected monitoring performance parameters are obtained, denoted as X. c EGT s The corresponding X c Constituting the data group {EGT s X c }; S44. Using a polynomial fitting method, for all data sets {EGT} s X c By fitting the data, the corrected baseline value X of the monitoring performance parameter is obtained. cstd The functional relationship between X and exhaust temperature cstd =f(EGT), thus obtaining the baseline value X of the corrected monitoring performance parameter at any exhaust temperature. cstd ; S5. Monitoring Model Establishment: Using the proxy model from step S4, obtain the performance index X corresponding to each monitoring performance parameter. index =[(X c -X cstd ) / X cstd ]*100%; Real-time monitoring of the performance indices corresponding to various performance parameters enables online monitoring of the unit.

2. The gas turbine unit monitoring method according to claim 1, characterized in that: In step S2, after a certain period of operation, the simulation model is recalibrated using the actual operating data of the gas turbine unit.

3. The gas turbine unit monitoring method according to claim 1, characterized in that: In step S41, the exhaust temperature EGT ranges from 520℃ to 620℃, the ambient temperature T ranges from -15℃ to 45℃, the ambient pressure P ranges from 0.95 bar to 1.05 bar, the compressor inlet guide vane opening IGV ranges from 0% to 100%, and the ambient relative humidity Phi ranges from 0% to 100%.

4. The gas turbine unit monitoring method according to claim 1, characterized in that: In step S42, N1, N2, N3, N4 and N5 are all greater than or equal to 5.

5. The gas turbine unit monitoring method according to claim 1, characterized in that: In step S42, the simulation calculation results are organized to obtain each EGT. s The corresponding multiple actual monitoring performance parameters X s And it is displayed in the form of a chart.

6. The gas turbine unit monitoring method according to claim 1, characterized in that: In step S43, for each EGT s The corresponding multiple actual monitoring performance parameters X s The methods for making corrections include: S431, From all simulation calculation data sets {EGTs, T s P s IGV s Phi s X s Select to include the EGT. s The simulation calculation data set is denoted as EGT. s The runtime parameter data group {EGTs, T s P s IGV s Phi s X s }; S432. The EGT is corrected through four single-factor corrections: environmental pressure correction, environmental temperature correction, IGV correction, and environmental relative humidity correction. s The corresponding multiple actual monitoring performance parameters X s Make corrections, and each single-factor correction is the same, including the following steps: S4321. Determine the data group before this correction, denoted as the data group to be corrected. If this correction is the first correction, then the database to be corrected is the running parameter data group {EGTs, T} of the EGTs selected in step S431. s P s IGV s Phi s X s If this is not the first revision, then the data set to be revised is the data set obtained after the last single-factor revision. S4322, T s P s IGV s Phi s Of these four parameters, the factor related to this correction is represented by Y, and the other three factors are represented by M. All data sets to be corrected are organized, and data with the same M but different Y are grouped into a set of fitting calculation data sets. Each set of fitting calculation data sets includes multiple data sets. S4323. Use the fitting calculation data set to perform fitting calculation, and compare the actual monitoring performance parameters corresponding to different Y values ​​in the same fitting calculation data set with the benchmark monitoring performance parameters X under the benchmark operating conditions. 基 The ratio is processed, and then Y is used as the independent variable and the ratio is used as the dependent variable. Mathematical methods are used to fit the data to obtain the functional relationship between the ratio and Y. Then, based on the functional relationship between the ratio and Y, the actual monitoring performance parameters in each data group to be corrected are corrected, and the corrected actual monitoring performance parameters are calculated. Then, the actual monitoring performance parameters in the data group to be corrected are replaced with the corrected actual monitoring performance parameters to obtain the corrected data group.

7. The gas turbine unit monitoring method according to claim 6, characterized in that: In step S4323, the reference operating conditions include: ambient temperature 15℃, ambient pressure 1 bar, compressor inlet guide vane opening 100%, and ambient relative humidity 60%.

8. A gas turbine unit monitoring system based on a surrogate model, characterized in that: The system includes an operational data acquisition device and a monitoring and processing module. The operational data acquisition device collects operational parameter data of the gas turbine unit and transmits the relevant operational data to the monitoring and processing module. The monitoring and processing module stores a gas turbine monitoring computer program. The gas turbine monitoring computer program includes a monitoring model established by the gas turbine unit monitoring method as described in any one of claims 1 to 7. When the gas turbine monitoring computer program is executed, it can calculate the performance index X corresponding to each monitoring performance parameter based on the operational parameter data collected by the operational data acquisition device. index .

Images