JP2005248848A - Gas turbine diagnostic method and apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a gas turbine diagnosis method and apparatus capable of obtaining an accurate diagnosis based on appropriate model identification of a gas turbine to be applied.
Operation state information 9 detected by a sensor 3 from a gas turbine 2 to be diagnosed is divided into operation state information 10 for diagnosis and operation state information 11 for model identification. Based on the operation state information 11, the gas turbine The model identification unit 7 of the diagnosis unit 6 identifies the performance specifications of the gas turbine 2 to generate model constant information 16, and the operation state information 10 for diagnosis is compared with the model constant information 16 by the performance diagnosis unit 8 to obtain gas. Diagnosis of the soundness of the turbine 2, and maintenance plan information can be generated as needed.
Here, by generating the model constant information 16 over a certain period, the performance specifications of the gas turbine 2 can be identified, so that it is constantly compared with the operation state information 10 for diagnosis given in time series. By doing so, the soundness of the gas turbine 2 can be diagnosed.
[Selection] Figure 1

Description

  The present invention relates to a method and apparatus for function diagnosis of a gas turbine, and more particularly to a diagnosis method and apparatus suitable for a gas turbine for power generation.

  In recent years, for example, ESCO (Energy Service Company), which undertakes energy saving in general production facilities such as factories and building facilities, and energy such as electric power and steam are generated at facilities installed in customer factories and sold to customers Various energy service businesses, such as the O & M (Operation & Maintenance) business and the business that undertakes the operation and maintenance of customer facilities in a lump, are attracting attention.

  By the way, in such an energy service business, from the viewpoint of cogeneration, power generation equipment using a gas turbine as a prime mover has attracted attention in recent years. At this time, the gas turbine is a representative of mechanical equipment with deterioration. And when used for power generation, due to the high importance of securing power, particularly accurate maintenance is indispensable, and as a result, in the various businesses mentioned above, This has led to an increase in profits.

Therefore, in the prior art, by calculating a gas turbine control model and comparing it with the actual gas turbine operation information, the performance of the gas turbine is confirmed and abnormalities are diagnosed, so that accurate maintenance is always obtained. (For example, see Patent Document 1, Patent Document 2, and Patent Document 3).
JP 2003-83089 A JP 2001-90554 A JP 2001-174366 A

  The above prior art does not take into consideration that the performance of the gas turbine to be applied is known, and the application target is limited or accurate maintenance cannot be obtained. There was a problem.

  In the prior art, the control model of the gas turbine is calculated. The calculation at this time is based on the performance of the gas turbine to which the gas turbine is applied, and therefore the performance of the gas turbine must be known. ing.

  However, in the energy service business such as the ESCO business and the O & M business described above, there are cases where the specification details are unknown or the equipment whose performance specification state has changed from the initial state due to aging and is difficult to understand. Many.

  In particular, for gas turbines, manufacturers (manufacturers) usually rarely disclose information such as specifications and performance (manufacturer information), and the performance of gas turbines varies with operating conditions such as atmospheric conditions. Depending on the value, the value may fluctuate or may change to a value different from the original design value (specification value) due to aging, failure history, maintenance, parts replacement, and the like.

  In this way, if the manufacturer information about the gas turbine is not available and it is uncertain, in the prior art, it is necessary to use experience values or to adjust the model constant of the gas turbine by manual trial and error. For this reason, it is necessary to collect and search a large amount of maintenance records and correct specification information, which is troublesome and still does not always allow construction of a reliable model.

  Therefore, in the prior art, the target of application is limited to gas turbines whose performance is known, and otherwise, there is a problem that accurate maintenance cannot be obtained because a reliable model cannot be constructed. is there.

  An object of the present invention is to provide a gas turbine diagnosis method and apparatus capable of obtaining an accurate diagnosis based on an appropriate model identification of a gas turbine to be applied.

  In the gas turbine diagnosis method for comparing the operating state information detected from the gas turbine with the performance specification information of the gas turbine set in advance, and diagnosing the soundness of the gas turbine based on the comparison result, This is achieved by identifying model constant information of the gas turbine based on operating state information detected from the gas turbine, and diagnosing the soundness of the gas turbine using the model constant as the performance specification information.

  At this time, even if the calculation process for identifying the model constant information is shared with the calculation process for diagnosing the soundness of the gas turbine, the above object can be achieved.

  Further, the object is to compare the operating state information detected from the gas turbine with the performance specification information of the gas turbine set in advance, and to diagnose the soundness of the gas turbine based on the comparison result A model identification unit that generates model constant information based on the operating state information detected from the gas turbine, and the model constant information generated by the model identification unit is used as performance specification information of the gas turbine. This is achieved even if the health is diagnosed.

  At this time, the model identification means and the arithmetic processing means for diagnosing the soundness of the gas turbine have a common arithmetic unit, and the arithmetic unit is shared by the model identification means and the arithmetic processing means by mode switching. Even so, the above object can be achieved.

  According to the present invention, model constants are automatically identified, so that a model with excellent reliability can be identified with good reproducibility without being influenced by human experience and proficiency. Accurate diagnosis based on this can be obtained efficiently with labor saving.

  In addition, according to the present invention, even when the operating state changes over time, not only can effective diagnosis be performed by re-determining the model, but the performance has greatly changed from the initial performance state due to deterioration over time. Equipment can be diagnosed with high accuracy and reliability.

  Hereinafter, a gas turbine diagnosis method and apparatus according to the present invention will be described in detail with reference to embodiments shown in the drawings.

  Here, first, the present invention compares the predicted performance calculated based on the operating conditions of the gas turbine power generation facility (hereinafter referred to as gas turbine) with the performance obtained from the actual measurement data of the gas turbine. In the method and device for diagnosing performance degradation and abnormality, even if the specification information of the target gas turbine is unknown, the predicted performance such as output and efficiency according to the operating conditions can be predicted with high accuracy, Performance degradation and abnormality can be realized with high reliability. Therefore, in one embodiment of the present invention, when the diagnosis is performed, the system configuration shown in FIG. 1 is used.

  Here, the system of FIG. 1 is roughly divided into an actual equipment 1, an operation data management unit 4, and a gas turbine diagnosis unit 6. First, the actual equipment 1 includes a gas turbine 2 and various sensors 3 installed in the gas turbine. Next, the operation data management unit 4 stores operation state information 11 of the gas turbine 2. A stored operation record database 5 is provided.

  Further, the gas turbine diagnosis unit 6 includes a model identification unit 7 and a performance diagnosis unit 8. First, the model identification unit 7 is based on the model condition constant 11 based on the operation state information 11 supplied from the operation record database 5. It serves to identify information 16. Here, the model constant information 16 is information necessary for predictive calculation of performance according to the operating conditions of the gas turbine 2, which includes specification information of the gas turbine 2 and the like.

  Next, the performance diagnosing unit 8 functions to diagnose deterioration or abnormality of the performance of the gas turbine 2 based on the operation state information 10 representing the actual daily operation state of the gas turbine 2 and the model constant information 16.

  At this time, the operation state information representing the actual daily operation state of the gas turbine 2 is detected by a plurality of sensors 3 provided in the gas turbine 2 and output from the gas turbine 2 as operation state information 9 and 10. ing.

  Of these, first, the driving state information 9 is transmitted to the driving data management unit 4 via the information paths 12 and 13, where it is stored in the driving record database 5 as a driving state history, while the driving state information 10 is The information is transmitted to the model identification unit 7 via the information paths 12 and 14 and used for diagnosis.

  Here, in the case of this embodiment, the model identification unit 7 and the performance diagnosis unit 8 are not divided into independent parts as shown in the figure, but actually, as will be described later, the gas turbine diagnosis unit 6 share the same main arithmetic unit, and have a structure for realizing model identification and diagnosis functions by switching information processing systems in accordance with calculation modes.

  In the gas turbine diagnosis unit 6, there are different operation modes corresponding to the functions of model identification and diagnosis, and this is realized by mode switching. Therefore, hereinafter, one operation mode is referred to as a model identification mode, and the other operation mode is referred to as a diagnosis mode.

  First, in the model identification mode, the model identification unit 7 extracts a plurality of operation state information 11 from the operation record database 5 and identifies model constant information of the gas turbine 2 based on the extracted plurality of operation state information 11. And output as model constant information 16. Details of the processing procedure and contents of the model identification unit 7 at this time will be described later.

  Here, the operation state information 11 extracted from the operation record database 5 by the model identification unit 7 roughly includes information related to operation conditions and information related to performance states.

  Representative items of information related to one operating condition include atmospheric conditions such as the intake temperature, intake pressure, and humidity of the gas turbine (hereinafter referred to as atmospheric condition information), the fuel heat input amount of the gas turbine, and the refrigerant injection amount. There are heat input to the gas turbine and material input condition (hereinafter referred to as input condition information). Hereinafter, these are collectively referred to as operation condition information.

  Also, representative items of information relating to the other performance state include the amount of power generated by the gas turbine, the exhaust temperature, and the compressor discharge pressure. Hereinafter, these conditions are collectively referred to as performance state information.

 In addition, about the exhaust temperature, when the turbine has a plurality of stages, it may be an outlet temperature of an arbitrary stage. The same applies to the case where the exhaust temperature is described below. Further, the compressor discharge pressure may be a pressure ratio indicating a ratio between the outlet pressure and the inlet pressure of the compressor or the turbine. The same applies to the pressure ratio described below.

  Next, the model constant information 16 output from the model identification unit 7 is primarily information such as the compressor efficiency of the gas turbine, the turbine efficiency, the combustion temperature, and the amount of mechanical and thermal loss energy. These are the design specifications or performance factors of the gas turbine, specifically, for example, compressor specification information, turbine specification information, and combustor specification information.

  Here, the compressor specification information is the physical shape of the compressor in which the compressor efficiency can be calculated by calculating aerodynamics by combining the intake air flow rate of the compressor and the efficiency of the compressor, or the atmospheric condition information and the intake air flow rate information. Information.

  The information on the physical shape of the compressor specifically includes, for example, the number of compressor stages, the representative diameter, the size diameter and shape of the moving blades and stationary blades of each stage, and the size and shape of the air passage. , Including at least one design value of a velocity triangle formed by air and wings.

  Next, the turbine specification information is information on the physical shape of the turbine in which the turbine efficiency can be calculated by calculating the aerodynamics in combination with the turbine efficiency or the information on the temperature, pressure and specific heat of the turbine inflow gas. For example, the number of turbine stages, representative diameter, size and shape of the moving blades and stationary blades of each stage, the size and shape of the passage path of the combustion gas, the design value of the speed triangle formed by air and blades, the cooling air flow rate Including at least one of the design values.

  The combustor specification information is information such as the combustion efficiency and combustion temperature of the combustor and the refrigerant injection amount.

  However, as mentioned above, manufacturers (manufacturers) usually rarely disclose all these information (manufacturer information), and these performance factors have values due to fluctuations in operating conditions such as atmospheric conditions. May change, or may change to a value different from the original design value due to aging deterioration, maintenance, parts replacement, or the like.

  Therefore, when it is assumed that these performance factors are used as input information for calculation for gas turbine performance prediction, rather than treating it as a specification value fixed to a fixed value, equipment such as operating conditions and deterioration over time. It is possible to increase the accuracy of the performance prediction calculation by treating it as including the functional relationship of the value change according to the state change.

  Therefore, in the following description, the information on these performance factors will be referred to as model constant information 16 in the sense of including constant information such as the above-described function-related coefficients.

  Note that the information paths 12, 13, and 14 specifically refer to wired or wireless information communication lines, or information transfer in an information processing process in the same device. Therefore, in the following examples, too. Unless otherwise specified, the information path indicates the same general communication line means.

  Next, in the diagnosis mode, the performance diagnosis unit 8 performs performance degradation of the gas turbine 2 based on the model constant information 16 and the operation state information 10 representing the daily time-series operation state associated with the operation of the gas turbine 2. Or abnormalities are diagnosed.

  Here, since the generation of the model constant information 16 is performed over a certain period as will be described later with reference to FIG. 3, the performance specifications of the gas turbine 2 can be identified. The soundness of the gas turbine 2 can be diagnosed by analyzing the performance state every moment by the performance diagnosis unit 8 based on the operation state information 10 given in time series.

  At this time, as will be described in detail later, the performance diagnosis unit 8 first separates the operating state information 10 into information on operating conditions (assuming 10A) and information on performance state (also 10B), The performance of the gas turbine 2 is predicted and calculated based on the information 10A related to the operating conditions and the model constant information 16, and then the result of the prediction calculation is compared with the information 10B related to the performance state and compared. The degree of deterioration, the presence / absence of abnormality, and the presence / absence of signs of deterioration and abnormality are diagnosed.

  Here, the diagnostic method and apparatus according to the embodiment of the present invention are particularly effective in the energy service business such as the ESCO business and the O & M business for energy saving described above. This is because, in this kind of business, as mentioned above, there are cases where the specification details are particularly unknown, or equipment whose performance specifications have changed from the initial state as a result of aging and are difficult to understand. Because there are many.

  As described above, when the manufacturer information about the gas turbine is not available and it is uncertain, as described above, it is necessary to use experience values, adjust model constants by manual trial and error, and perform extensive maintenance. Since it was necessary to collect and search records and correct specification information, it was time consuming and could not build a reliable model.

  In contrast, according to the embodiment of the present invention, model constants are automatically identified, so that highly accurate and reliable diagnosis can be saved, and diagnosis can be performed efficiently in a short time. Since the identification of model constants is automated, model identification with excellent reproducibility and reliability is possible regardless of human experience and proficiency.

  Furthermore, according to this embodiment, even when the operating state changes over time, not only effective diagnosis can be performed by re-determining the model, but also aging, repair, parts replacement, periodic inspection, overhaul, Highly accurate and reliable diagnosis is possible even for equipment whose performance has changed greatly from the initial performance state due to failure history.

  Next, the configuration and processing procedure of the gas turbine diagnosis unit 6 in the above embodiment will be described with reference to FIG.

  Here, the gas turbine diagnosis unit 6 shown in FIG. 2 shares the calculation part, and can be used by switching to the above-described processing of the model identification mode and the diagnosis mode. The switching between the model identification mode and the diagnosis mode is given by mode instruction information from the operator to the mode input unit 21.

  Therefore, the gas turbine diagnosis unit 6 includes a mode input unit 21 to which mode instruction information is given, a main calculation unit (operation value vs. actual machine value comparison information generation unit) 31 shared by each mode, and the main calculation unit 31. Information processing branch for processing the output information as the calculation result of the main calculation unit 31 according to the mode instruction information, and the mode determination unit 22 for performing the information processing branch for generating the input information for the mode instruction information according to the mode instruction information A mode determination unit 33 for performing the above is provided.

  In the following, the processing procedure by the gas turbine diagnosis unit 6 will be sequentially described separately for the model identification mode and the diagnosis mode.

<In model identification mode>
Here, it is assumed that mode instruction information indicating the model identification mode is given to the mode input unit 21 by the operator. Then, the mode determination unit 22 selects an information processing route toward the driving condition information extraction unit 23, and the mode determination unit 33 selects an information processing route toward the model constant information search unit 34.

  Therefore, the driving condition information extraction unit 23 takes in representative driving state information 11 under a plurality of driving conditions from the driving record database 5 (FIG. 1) via the information path 15. The typical operating state information 11 under a plurality of operating conditions at this time is, for example, typical operating state information when the intake air temperature is 0 ° C., 15 ° C., and 30 ° C. with respect to the atmospheric conditions described above. There is no need to limit this. Here, the specific items of the driving state information are as described above.

  Next, the model constant information initial value generation unit 27 is closest to the power generation output at this time based on the representative information of the power generation output information of the gas turbine among the operation state information processed by the operation state information extraction unit 23. The model constant information 28 of the power generation output gas turbine is extracted from the model constant information database 37.

  Then, the model constant information 28 is validated by the subsequent information processing, and the process for correcting the model constant information 28 is corrected as necessary, so that the validity of the model constant information 28 is improved so as to increase the validity. It is given as an initial value for

  Here, the model constant information database 37 shows the rated power output information and model constant information for a different model different from the target gas turbine 2 or a part of which the model constant information is different, such as equipment of the same type but with a different product number. The database is stored in contrast to form a database.

  If the model constant information database 37 can output the model constant information from the power generation output information, for example, the model constant information database 37 holds the relationship between the power generation output information and the model constant information in a functional relationship. It may be in the form of a model constant information estimation unit that outputs model constant information from information input.

  Next, the calculation input unit 30 accepts the initial values of the driving state information 11 output from the driving state information extraction unit 23 and the model constant information 28 output from the model constant information initial value generation unit 27, and these are received next. Input information to the main calculation unit 31 of the first.

  Therefore, the main calculation unit 31 calculates the predicted performance of the gas turbine 2 under a plurality of operating conditions, and generates calculation result information as information for comparing the actual performance information of the actual machine included in the operation state information 11. To do.

  The details of the generation procedure of the calculation result information by the main calculation unit 31 at this time are as shown in FIG. 3. Here, first, the information separation step 52 is operation state information in a plurality of operation conditions in the calculation input unit 30. 11 is separated into information relating to the operating condition and information relating to the performance state described above, the former being used as the operation state information 53 for performance prediction computation in the subsequent performance computation step 56, and the latter being for verifying the performance prediction computation result. Information 54 is sorted.

  First, the operating condition information 53 for performance prediction calculation input is supplied to the performance calculation step 56 together with the model constant information initial value 28, and here, based on these input information 53, 28, The performance prediction calculation is executed, and information for comparing with the verification information 54 of the performance information of the actual machine, that is, calculation result information 57 is generated.

  The calculation here may be based on at least one of thermodynamic calculation and hydrodynamic calculation in a normal gas turbine cycle, and information obtained as a calculation result includes at least power generated by the gas turbine and gas turbine compression. It includes the discharge pressure of the compressor (hereinafter referred to as compressor discharge pressure) and the exhaust temperature of the gas turbine.

  The comparison between the gas turbine performance calculation result information 57 and the verification information 54 is the same as the operating condition information 53 between the loop start end 51 and the loop end 61 in FIG. It is repeated according to the condition number of each operation information of temperature and generated power, and the comparison result of the performance evaluation information in the whole condition number is stored in the storage area of the accumulated information database 60 as calculation / actual machine comparison information 59, After completion of the calculation over the conditions, the result output step 62 outputs the calculation / actual machine comparison information 63 to the calculation output unit 32.

  Returning to the description of FIG. 2, the calculation / actual machine comparison information 63 output to the calculation output unit 32 as described above indicates that the current processing mode is the model identification mode. The data is transferred to a constant information search unit (specification search unit) 34. The model constant information search unit 34 may be called a specification search unit.

  As shown in FIG. 4, the model constant information search unit 34 reduces the error of the comparison result in the order of the compressor discharge pressure, the exhaust temperature, and the generated power from the delivered calculation / actual machine comparison information 63. The information obtained by correcting the values of the model constants for adjusting each of them, that is, the corrected model constant information 74 for reducing the error between the calculation result of the compressor discharge pressure and the actual machine, the calculation result of the exhaust temperature and the error of the actual machine. Correction information 35 including correction model constant information 77 for reducing, and correction model constant information 80 for reducing the calculation result of the generated power and the error of the actual machine is generated.

  At this time, the specific correction procedure and correction items of each correction model constant information are as follows. First, by the loop of the compressor discharge pressure evaluation determination step 72 and the compressor discharge pressure model constant information correction step 73, the error between the predicted value of the calculation program and the actual machine value is constant for the pressure ratio or the compressor outlet pressure. The efficiency of the compressor, the ratio of the cooling air flow rate extracted from the compressor and used for cooling the turbine, the intake air flow rate of the compressor, or at least one numerical value of the compressor specification information so as to be below the range The modified model constant information 74 of the compressor discharge pressure is generated by changing the magnitude of.

  Next, by the loop of the exhaust temperature evaluation determination step 75 and the exhaust temperature model constant information correction step 76, the turbine exhaust temperature is set so that the error between the predicted value of the calculation program and the actual machine value is within a predetermined range. The exhaust temperature correction model constant information 77 is generated by changing the shape or size of the turbine, work efficiency, the effect of cooling by the cooling air, or the value of at least one of the turbine specification information. The

  Finally, by the loop of the generated power evaluation determination step 78 and the generated power model constant information correction step 79, the generated output is set so that the error between the predicted value of the calculation program and the actual machine value is within a predetermined range. The modified model constant information 80 of the generated power is generated by changing the magnitude of at least one of the mechanical loss of the generator and the turbine, the heat dissipation loss of the combustor or the turbine, or the combustor specification information. Will be.

  The correction information 35 is input again to the calculation input unit 30 (FIG. 2), and the calculation of the comparison information by the calculation of the main calculation unit 31 and the evaluation are repeated, and the convergence calculation is performed so that the error is minimized.

  Then, when the convergence calculation is completed, that is, when the result is OK in the generated power evaluation determination 78, the model constant information is stored in the model constant information database 37 as the determined model constant information 16, and here the model identification mode The process by is terminated. At this time, the model constant information is also output to the outside.

<Diagnosis mode>
Next, assuming that mode instruction information indicating a diagnosis mode is given from the operator to the mode input unit 21, the mode determination unit 22 selects an information processing route toward the time-series operation state information acquisition unit 25.

  Therefore, the time-series operation state information acquisition unit 25 executes the acquisition process of the actual machine time-series data 10 and inputs it to the calculation input unit 30 as the actual machine operation state information 26. This indicates a process of sequentially taking daily operation state information of the gas turbine 2 in the actual equipment 1 as time series data.

  Note that this processing does not necessarily have to be real-time, and may be time-series data of operation state information during a certain period of the gas turbine 2, for example, operation records provided in the operation data management unit 4 (FIG. 1). Operation state information during a certain past period may be acquired from the database 5.

  At this time, the model constant acquisition unit 29 takes in the definite model constant information 16 obtained in the model identification mode from the model constant information database 37 and similarly inputs it to the calculation input unit 30.

  The actual machine operating state information 26 and the confirmed model constant information 16 acquired in this way are input to the main calculation unit 31 via the calculation input unit 30 and processed. Therefore, the specific procedure will be described below with reference to FIG.

  First, the calculation input unit 30 receives the actual machine operating state information 26 and the confirmed model constant information 16 and uses them as input information for the next main calculation unit 31.

  Next, the main calculation unit 31 calculates the predicted performance of the gas turbine 2 corresponding to the operating conditions (the operating condition information) of the gas turbine included in the actual machine operating state information 26, and calculates the calculation result as the actual machine operating state information 26. Is generated as information for comparison with the actual performance information (the performance state information) included in the.

  The generation procedure of the calculation result information by the main calculation unit 31 at this time is as follows. First, the information separation step 52 takes the actual machine operation state information 26 in the calculation input unit 30 and separates it into the above-described operation condition information and performance state information. The former is classified as operation state information 53 for performance prediction calculation in the subsequent performance calculation step 56, and the latter is classified as verification information 54 for verifying the degree of performance degradation in comparison with the performance prediction calculation result. Since the verification information 54 represents the performance state of the actual machine, it may be called an actual machine value.

  First, the operation state information 53 for performance prediction calculation input is supplied to the performance calculation step 56 together with the definite model constant information 16, where the performance of the gas turbine 2 is based on these input information 53 and 16. A prediction calculation is executed, and information for comparison with the verification actual machine performance information 54, that is, calculation result information 57 is generated.

  The calculation here may be based on at least one of thermodynamic calculation and hydrodynamic calculation in a normal gas turbine cycle, and information obtained as a calculation result includes at least the power generated by the gas turbine and the gas turbine. It includes compressor discharge pressure and gas turbine exhaust temperature.

  Such a comparison between the performance calculation result information 57 of the gas turbine and the actual machine performance information 54 for verification is between the loop start end 51 and the loop end 61 in FIG. Alternatively, if the information 26 has been acquired, the information 26 is repeated according to the target period, and the comparison result of the performance information during this period is stored in the storage area of the accumulated information database 60 as calculation / actual machine comparison information 59 After completion of the calculation over the entire period, the result output step 62 outputs the calculation / actual machine comparison information 63 to the calculation output unit 32.

  That is, in the diagnosis mode, the main calculation unit 31 inputs the time-series operating condition information of the actual machine under a constant model constant condition to predict and calculate the performance state of the gas turbine, and the predicted value of the performance state by this calculation Is output as calculation / actual machine comparison information 63 in comparison with the actual performance state information of the real machine.

  Here, returning to the description of FIG. 2, the calculation / actual machine comparison information 63 output to the calculation output unit 32 as described above is the trend diagnosis from the mode determination unit 33 because the current processing mode is the diagnosis mode. Passed to part 36.

  Therefore, the trend diagnosing unit 36 diagnoses performance deterioration, abnormality presence / absence / prediction, and the like based on the predicted value of the performance state of the gas turbine and the comparison information of the actual machine value indicated in the calculation / actual machine comparison information 63.

  In this diagnosis, as shown in FIG. 5, the calculation / actual machine comparison information 63 is input to the trend analysis step 92, and the trend analysis step 92 compares the predicted value with the actual machine value and information on the change in the trend with the passage of time. As described below, the analysis information is output as degradation / abnormality evaluation information 93.

  Here, the relationship between the comparison information between the predicted value and the actual machine value by the performance calculation and the diagnosis of the presence / absence / prediction of the performance deterioration or abnormality is as follows, for example.

・ When the actual value of the generated power is lower than the predicted value or when there is a continuous downward trend → The output performance of the gas turbine as a whole is decreasing.

• When the actual compressor intake flow rate is lower than the expected value or when there is a continuous downward trend → The compressor intake capacity is particularly low among gas turbines.

・ When the actual value of the compressor outlet temperature is higher than the predicted value or when the deviation continuously increases → Among the gas turbines, the efficiency of the compressor is particularly low.

・ When the actual gas turbine exhaust temperature is higher than the predicted value or when the deviation continuously increases → The efficiency of the turbine is decreasing.

・ When the actual value of the gas turbine exhaust temperature is lower than the predicted value, or when the deviation continuously decreases → The cooling port of the turbine blade has deteriorated and the cooling air flow rate has increased, or the combustor has Deteriorated and combustion efficiency is reduced.

・ If the actual compressor discharge pressure value is lower than the predicted value or if the deviation continuously decreases → The efficiency of at least one of the compressor or turbine is decreasing.

  Furthermore, in the trend analysis process 92, the performance degradation of the gas turbine as a whole is summarized by combining the information on the performance degradation of the gas turbine as a whole and the degradation of the parts such as the compressor, turbine, and combustor. The rate, the performance degradation factor and the performance degradation rate for each part in it are estimated.

  As this estimation method, for example, the estimated content of the performance reduction rate of each part described above (eg, intake air flow rate reduction rate, compressor efficiency reduction rate, turbine efficiency reduction rate, cooling air flow rate increase rate, combustion efficiency) And the like, and the corrected model constant information 16 is corrected to the content assuming deterioration (hereinafter, corrected model constant information 16). In addition to the above-described actual machine operating state information 26, it is input to the main calculation unit 31 to calculate the gas turbine performance, and the calculation result and the actual machine comparison information 63 are obtained by the same procedure as described above with reference to FIG. There is a method of adjusting by changing the numerical value of the estimated content of the performance deterioration rate of each part (hereinafter referred to as the part performance deterioration rate) so that the calculated value of the gas turbine performance in the comparison information 63 is equal to the actual machine value. .

  The adjustment of the site performance reduction rate here is to adjust the site performance reduction rate so that the calculation results of the compressor discharge pressure, the turbine exhaust temperature, and the generated power are equivalent to the actual machine values.

  As the adjustment method, for example, in order of priority, compressor discharge pressure, turbine exhaust temperature, and generated power, so that the error between the calculation result and the actual measurement value is equal to or less than a preset target value. There is a method of repeating the search by mechanically increasing / decreasing the value of the performance degradation rate, but it is not necessary to be limited to this.

  For example, a plurality of combinations of contribution ratios of each part to the overall performance degradation, that is, a combination of part performance deterioration rates (hereinafter referred to as deterioration patterns) are prepared in advance, and a plurality of correction model constant information corresponding to the plurality of deterioration patterns 16 is input to the main calculation unit 31 together with the operating state information 26 in the same manner as described above to calculate the performance of the gas turbine, and an error obtained by comparing the calculation result of the calculation / actual machine comparison information 63 with the actual measurement value. It is possible to select a degradation pattern that minimizes.

  In this case, the part performance deterioration rate of the pattern with the smallest error selected is used as a reference, and the whole or a part of the part deterioration rate is further multiplied by a certain ratio value, or a constant value is not added. By subtracting, at least one other deterioration pattern is newly generated, and the corrected model constant information 16 is generated in the same manner as described above according to the deterioration pattern, and is input to the main calculation unit 31 together with the operation state information 26. Calculating the performance of the gas turbine, selecting a pattern in which the error between the calculation result of the calculation / actual machine comparison information 63 and the actual measurement value is further minimized, and performing this operation on a desired value in which the calculation result and the actual measurement error are preset. It is more effective to repeat until the following.

  Alternatively, not limited to these methods, a known multivariable optimization method (such as a genetic algorithm) that automatically searches for a combination of multiple variables that minimizes the value of the target evaluation function (in this case, an error) is used. Needless to say, it may be.

  In the trend analysis step 92, the progress of the future deterioration is predicted based on the information on the deterioration state of the performance of the whole gas turbine and each part estimated as described above.

  As a method for predicting the deterioration progress, for example, at least one type of model information in which the performance deterioration curve indicating the progress of the overall performance of the gas turbine and the performance deterioration for each part over time is prepared in advance in the step 92. By fitting the function model with the time degradation of the performance degradation rate of the entire gas turbine and the performance degradation rate of each part at each time point estimated as described above to this function model by a method such as the least squares method, By determining the calculation formula for deterioration of the entire turbine performance and the performance of each part over time, and inputting future date and time information into this calculation formula, the progress of the performance degradation of the whole gas turbine and each part in the future Can be predicted.

  Regarding the above trend analysis, there are various usage methods in the operation of this system. For example, if the performance characteristics of the equipment change due to periodic inspections, maintenance, or parts replacement, or if the performance characteristics of the equipment change after a certain point due to sudden abnormalities or failures, etc. There are cases where trends are evaluated, or there is a case where data for a certain period in the past is analyzed for a specific purpose.

  The need for trend analysis in this way is not limited to simply monitoring daily operating conditions, but also includes data analysis for any period in the past. In the case of the conventional method, it has been difficult to diagnose the above-described method since it requires labor from trial to error until data is identified and the model constant is identified.

  However, according to the present invention, the entire evaluation system has an identification unit and a diagnosis unit that can switch modes by sharing a main arithmetic unit, and the processing of the identification unit is as described with reference to FIG. Since it can be automatically and efficiently executed with high accuracy, it is possible to quickly and flexibly diagnose deterioration of the performance state with high accuracy according to the various analysis purposes as described above.

  Returning to FIG. 2, the subsequent maintenance plan drafting unit 39 formulates an optimum maintenance plan under this degradation progress condition, reflecting the current performance degradation state and degradation / abnormality evaluation information 93 indicating future degradation prediction. .

  The optimal maintenance plan here is, for example, when the purpose is to minimize costs, and what kind of maintenance measures should be taken at what time in consideration of the progress of performance deterioration of gas turbines in the future. Whether the total cost considering the cost and the effect of reducing the operational cost due to the performance recovery due to the maintenance measure can be minimized, the maintenance measure database 38 storing the information on the maintenance cost according to the content of the maintenance measure and the performance recovery effect The evaluation calculation is performed based on the information, and the maintenance plan information 40 including the maintenance execution time and the maintenance content is output.

In this example, cost minimization has been taken as an example, but the purpose of optimization need not be limited to cost. For example, the amount of fuel consumed is multiplied by CO ₂ emissions per unit of fuel (CO ₂ basic unit). For example, the purpose may be to minimize the cumulative value of CO ₂ emission during a certain period.

  In addition, cost minimization has been explained on the assumption that the total cost plus maintenance cost and operation cost is minimized in the above example, but depending on the evaluation purpose during actual operation, the maintenance cost may be minimized. Needless to say, it may be the minimization of the operating cost or the cumulative burden cost of the gas turbine operator from the viewpoint of the gas turbine operator, and is not limited to the above total cost minimization. .

  In the following description, the total cost minimization will be described as an example. However, even if the information processing procedure is basically the same as that for the total cost minimization, the evaluation purpose is not limited to this.

  The details of the processing of the maintenance plan drafting unit 39 will be described below with reference to FIG.

  First, the maintenance countermeasure scenario generation processing step 94 refers to the maintenance countermeasure database 38 under the current deterioration state and the future deterioration prediction state indicated in the deterioration / abnormality evaluation information 98, and the previous deterioration diagnosis result information. Based on (deterioration abnormality evaluation information 98), zero or more maintenance time candidates and zero or more maintenance countermeasure content candidates are combined to generate at least one maintenance countermeasure scenario 96 corresponding to the deterioration state. Here, the maintenance execution time of 0 represents a scenario in which maintenance is not performed and the current situation is maintained.

  In addition, the maintenance countermeasure database 38 stores information such as maintenance contents corresponding to the maintenance contents and performance recovery by maintenance. Specifically, the maintenance countermeasure expenses for each part and the restoration degree of the part performance at the time of maintenance are stored. Information, maintenance measure cost according to the type of periodic inspection, and overall performance and / or information on the degree of restoration of part performance are stored.

  These pieces of information may be case information, certain standard set values, or a function of these according to the type, capacity, operation years, etc. of the gas turbine.

  When the maintenance measure scenario 96 for each case is thus generated, the maintenance measure scenario 96 is first processed in the life cycle profit calculation processing step 98 for each maintenance measure scenario 96, and the result is calculated in the calculation result evaluation processing 99. The optimum scenario update process 100 is executed according to the determination result.

  Then, this process is repeated for each case of the maintenance countermeasure scenario 96 from the loop start end 97 to the loop end 101, and then the maintenance plan output process 102 is executed to obtain the optimum maintenance plan information 40 according to the purpose. Is output.

  That is, first, in the life cycle profit calculation 98, the total cost including the maintenance cost and the reduction of the operation cost due to the performance recovery (the sum of the maintenance cost and the operation cost) according to the maintenance measure scenario 96 up to an arbitrary period during the operation period. Calculated as the cumulative profit / loss of life (life cycle profit / loss calculation).

  Information on the calculation result of the life cycle profit / loss corresponding to each maintenance measure scenario 96 is accumulated in the information storage means 103 and processed in the subsequent calculation result evaluation step 99 and the optimum scenario update step 100.

  The calculation result evaluation step 99 compares whether the calculation result of the profit / loss (cost) is lower than the minimum value of the profit / loss (cost) calculated so far each time the maintenance scenario is calculated for the life cycle profit / loss. If the cost of the calculation result is smaller than the previous minimum value, the process proceeds to the optimum scenario update step 100 (in this case, since there is no comparison target in the case of the first maintenance scenario, the profit / loss of the scenario is Considered minimal).

  The optimal scenario update process 100 determines which scenario is the most optimal at that time during the repeated evaluation of life cycle profit / loss according to the maintenance countermeasure scenario (the scenario with the lowest cost when cost minimization is intended). ), And the optimum scenario information is updated according to the result of the previous calculation result evaluation step 99.

  By repeating such processing between the loop ends 97 and 101 for all scenarios, a maintenance scenario in which life cycle profit / loss is the minimum cost is determined as a result, and the maintenance timing and maintenance content are proposed as a maintenance plan. Information 40 is output.

  As described above, an optimal maintenance plan can be calculated based on the quantitative evaluation result of the deterioration factor and the effect at the time of maintenance.

  Regarding the above maintenance planning, there are various usage methods in the operation of this system. For example, when the gas turbine continues to operate normally, the possibility of deferring the maintenance contents at the next periodic inspection time is predicted for each part of the gas turbine, or conversely, When the operating condition of the turbine is rapidly deteriorated, the deterioration factor is estimated, and what kind of maintenance is carried out as soon as possible in consideration of the constraints of the schedule that can shut down the gas turbine plant. There are cases where measures are evaluated to determine whether it is good.

  When considering a maintenance plan according to such a purpose, the selection of the period of data to be diagnosed is naturally different. In this way, when planning maintenance plans not only for daily operation status monitoring but also for data in various past periods, identification of model constants, trend analysis, maintenance planning Is a series of inseparable steps.

  In this regard, the conventional method is not easy because it requires labor from trial and error from data acquisition to model constant identification, trend analysis analysis, and maintenance planning every time maintenance planning is required. However, in the present invention, the overall evaluation system has an identification unit and a diagnosis unit that can switch modes by sharing a main arithmetic unit, and the processing of the identification unit is automated as described with reference to FIG. In the diagnostic process, the process from trend analysis to maintenance plan planning is executed quickly, so that it can be executed quickly and flexibly according to the various maintenance plan consideration purposes. It is possible to devise appropriate maintenance measures according to the deterioration state based on the accurate deterioration diagnosis result of the performance state.

  Therefore, according to the above embodiment, even if the specification is unknown, the model can be easily constructed and adjusted from normal operation data, and high accuracy can be obtained without installing a special instrument for estimating the specification. Model identification can be obtained, a highly reliable diagnosis can be performed even for a gas turbine of unknown specifications, and an optimum maintenance plan can be drawn up.

  Further, as described above, according to this embodiment, for example, measurement of 10 or more points around the gas turbine to be diagnosed, such as compressor outlet air temperature, exhaust temperature, generated power, combustion heat input, and the like. Thus, since the internal state of the gas turbine can be estimated, high accuracy can be maintained even if the number of monitoring instruments required for diagnosis is reduced.

  Since high accuracy can be maintained in this way, for example, the monitoring data transmission frequency is suppressed when the operation state of the gas turbine to be diagnosed is constant and the frequency is increased when a change occurs. The frequency can be adjusted according to the diagnosis state, and as a result, the amount of information transmitted when remotely monitoring can be reduced.

1 is a block diagram illustrating an embodiment of a gas turbine diagnosis method and apparatus according to the present invention. It is a block diagram which shows the detail of the gas turbine diagnostic part in one Embodiment of this invention. It is a block diagram which shows the detail of the calculation input part and main calculation part in one Embodiment of this invention. It is a block diagram which shows the detail of the model constant information search part in one Embodiment of this invention. It is a block diagram which shows the detail of the tendency diagnosis part in one Embodiment of this invention, and a maintenance plan planning part.

Explanation of symbols

1: Actual equipment,
2: Gas turbine 3: Sensor 4: Operation data management unit 5: Operation record database 6: Gas turbine diagnosis unit 7: Model identification unit 8: Performance diagnosis unit 9: Operation state information 10: Operation state information 11: Operation state information 12 : Information line 13: Information line 14: Information line 15: Information line 16: Confirmed model constant information 21: Mode input unit 22: Mode determination unit 23: Operating condition data extraction unit 27: Model constant information initial value generation unit 28: Model Constant information 30: Calculation input unit 35: Correction information 37: Model constant information database 31: Main calculation unit (calculation value vs. actual machine value comparison information generation unit)
52: Information separation step 52: Operation condition information for performance prediction calculation input 53: Information for performance prediction calculation result verification

Claims (4)

In the gas turbine diagnosis method for comparing the operating state information detected from the gas turbine with the performance specification information of the gas turbine set in advance and diagnosing the soundness of the gas turbine based on the comparison result,
A gas turbine diagnosis method characterized by identifying model constant information of the gas turbine based on operating state information detected from the gas turbine, and diagnosing the soundness of the gas turbine using the model constant as the performance specification information . The gas turbine diagnosis method according to claim 1,
A gas turbine diagnosis method characterized in that an arithmetic processing unit for identifying the model constant information is shared with an arithmetic processing unit for diagnosing the soundness of the gas turbine. In the gas turbine diagnostic device for comparing the operating state information detected from the gas turbine with the performance specification information of the gas turbine set in advance and diagnosing the soundness of the gas turbine based on the comparison result,
Provided with model identification means for generating model constant information based on the operating state information detected from the gas turbine,
A gas turbine diagnostic apparatus for diagnosing the soundness of the gas turbine using the model constant information generated by the model identification means as performance specification information of the gas turbine. The gas turbine diagnostic device according to claim 1,
Computation processing means for diagnosing soundness of the model identification means and the gas turbine includes a common computation unit,
The gas turbine diagnosis apparatus according to claim 1, wherein the calculation unit is shared by the model identification unit and the calculation processing unit by mode switching.

Images