JP2002054460A - Combustion vibration predicting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To facilitate regulation of a combustion control system and early detect the occurrence of combustion vibration during operation by predicting combustion vibration generated by the combustor of a gas turbine by a numeral formula model. SOLUTION: A combustion vibration predicting device comprises an input means 4 to input plant data, weather data, and the limit value of an internal pressure fluctuation; an internal pressure fluctuation characteristics grasping means 1 to effect numeral-formula-model of the internal pressure fluctuation of the combustor from inputted plant data and weather data; a combustion vibration region estimating means 2 to apply a limit value of an internal pressure fluctuation to a numeral formula model determined by the internal pressure fluctuation characteristics grasping means 1; and an output means 5 to output a combustion vibration region estimated result by a combustion vibration region estimating means 2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a combustion vibration predicting apparatus, and more particularly to a combustion vibration predicting apparatus for predicting combustion vibration generated in a combustor of a commercial or aircraft gas turbine.

[0002]

2. Description of the Related Art Various approaches have been made to adjust a control system by using a Kalman filter or a neural network. However, the phenomena related to the combustion oscillation of the gas turbine are complicated, and the field trial operation adjustment is still adjusted based on the experience and know-how of the operator.

[0003] Regarding the monitoring of combustion vibration during operation, data is collected by a pressure sensor installed in the combustor, and the frequency of combustion vibration is analyzed based on the data, so that it is possible to grasp early that combustion vibration is abnormal. There is known a technique for monitoring the operating state of a combustor with an emphasis on soundness.

FIG. 19 shows a conventional example as disclosed in JP-A-11-32.
1 shows a combustion abnormality monitoring device (gas turbine abnormality monitoring device) disclosed in Japanese Patent No. 4725. This monitoring device includes a pressure sensor 100 installed in a gas turbine combustor, an A / D converter 101 that converts a detection signal from the pressure sensor 100 into digital data and receives it, and decomposes the digital data into frequency components. Analysis device 102 for analyzing
And a determination condition setting unit 103 for variably setting reference data on a frequency component to be monitored based on a parameter defined by a gas turbine combustor output and a fuel supply amount thereof.
And, based on the reference data, extract a frequency component caused by the combustion oscillation phenomenon from the analysis data of the frequency component, and compare the amplitude value of the frequency component with the normal amplitude value of the data on the monitored frequency component. It has a judgment processing unit 104 for judging the combustion oscillation state of the gas turbine, and a result display unit 105 for displaying data relating to this judgment result.

In the above-described monitoring device, the frequency of the combustion vibration is analyzed based on the data from the pressure sensor 100, and whether the combustion vibration is abnormal is determined by comparing the amplitude value of the normal vibration for each frequency with the actual value. Can be determined.

[0006]

However, in the conventional monitoring device as described above, it is necessary to input the amplitude value of the combustion oscillation at the normal time for each frequency in advance. In effect, it becomes a management value such as an alarm, and it is not possible to detect combustion vibration at an early stage.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and facilitates adjustment of a combustion control system by predicting a combustion oscillation generated in a combustor of a gas turbine by a mathematical model. It is an object of the present invention to provide a combustion vibration prediction device that can detect the occurrence of combustion vibration at an early stage during operation, avoid damage to equipment, improve equipment utilization, and improve safety.

[0008]

In order to achieve the above object, a combustion vibration predicting apparatus according to the first aspect of the present invention constructs a mathematical model for explaining internal pressure fluctuation from plant data and weather data, Based on the model, a region where combustion vibration is likely to occur and a region where combustion vibration is unlikely to occur are obtained and output.

According to this configuration, a region where combustion oscillation is likely to occur and a region where combustion vibration is unlikely to occur are obtained based on a mathematical model for explaining internal pressure fluctuations constructed from plant data and weather data. Is output.

A combustion vibration prediction apparatus according to a second aspect of the present invention includes an input means for inputting plant data, weather data, a limit value of internal pressure fluctuation, etc., and an internal pressure fluctuation of a combustor based on the input plant data and weather data. Internal pressure fluctuation characteristic grasping means for modeling a mathematical model, and combustion vibration region estimating means for applying a limit value of the internal pressure fluctuation to the mathematical model obtained by the internal pressure fluctuation characteristic grasping means to obtain a region where combustion vibration is likely to occur, Output means for outputting a result of the combustion vibration region estimation by the combustion vibration region estimation means.

According to this configuration, the internal pressure fluctuation characteristic grasping means forms a mathematical model of the internal pressure fluctuation of the combustor from the plant data and weather data inputted by the input means, and the combustion vibration region estimating means restricts the internal pressure fluctuation to the mathematical model. A region where combustion oscillation is likely to occur is obtained by applying the value, and the combustion oscillation region estimation result is output from the output means.

Further, the combustion vibration predicting apparatus according to the third aspect of the present invention further includes a database for storing plant data, weather data, and the like input by the input means in a time series. Data is obtained from the database to formulate the internal pressure fluctuation of the combustor as a mathematical model.

According to this configuration, the plant data, weather data, and the like input by the input means are stored in the database in chronological order, and the internal pressure fluctuation characteristic grasping means obtains the data from the database and detects the internal pressure fluctuation of the combustor. Make a mathematical model.

According to a fourth aspect of the present invention, there is provided a combustion vibration predicting apparatus for inputting plant data, weather data and the like, and estimating an internal pressure fluctuation of a combustor from the input plant data and weather data. Means, and output means for outputting the internal pressure fluctuation estimation result estimated by the internal pressure fluctuation estimating means.

According to this configuration, the internal pressure fluctuation estimating means estimates the internal pressure fluctuation of the combustor from the plant data and the weather data input by the input means, and the estimated internal pressure fluctuation estimation result is output from the output means.

The apparatus for predicting combustion vibration according to the invention of claim 5 further includes a database for storing plant data, weather data, and the like input by the input means in a time series, wherein the internal pressure fluctuation estimating means includes a database. Is used to estimate the expected value of the internal pressure fluctuation based on the latest time data stored in.

According to this configuration, the plant data, weather data, and the like input by the input means are stored in the database in chronological order, and the internal pressure fluctuation estimating means uses the latest time data stored in the database to determine the internal pressure fluctuation. Estimate the expected value of.

Further, the combustion vibration predicting device according to the invention of claim 6 is a device for predicting internal pressure fluctuation, N based on plant data and weather data.
A mathematical model describing the amount of Ox emission is constructed, and based on the constructed mathematical model, a region where combustion oscillation is unlikely to occur and a region where combustion vibration is likely to occur are obtained and output.

According to this configuration, based on a mathematical model describing the internal pressure fluctuation and the NOx emission amount constructed from the plant data and the weather data, an area where combustion oscillation is unlikely to occur,
A region that is likely to occur is determined, and the result is output.

Further, the combustion vibration predicting apparatus according to the present invention comprises an input means for inputting plant data, weather data, a limit value of internal pressure fluctuation, a NOx control value, and the like, and an input means for inputting plant data and weather data. The internal pressure fluctuation characteristic grasping means for mathematically modeling the internal pressure fluctuation of the combustor, the NOx emission characteristic grasping means for mathematically modeling the NOx emission from the input plant data and weather data, and the internal pressure fluctuation characteristic grasping means are obtained. Applying the limit value of the internal pressure fluctuation to the obtained mathematical model, and applying the regulation value of NOx to the mathematical model obtained by the NOx emission characteristic grasping means,
The safety region estimating means for obtaining a region where the NOx emission amount is equal to or less than the regulation value and in which combustion oscillation is unlikely to occur, and an output means for outputting a safe region estimation result by the safety region estimating device.

According to this configuration, the internal pressure fluctuation characteristic grasping means forms a mathematical model of the internal pressure fluctuation of the combustor from the plant data and the weather data inputted by the input means, and the NOx emission characteristic grasping means is inputted by the input means. Formulate NOx emissions from plant data and weather data into mathematical models,
The safety region estimating means adds the internal pressure fluctuation limit value and N to the mathematical model.
A region where combustion oscillation is unlikely to occur is obtained by applying the regulation value of Ox, and the safety region estimation result is output from the output means.

Further, the combustion vibration predicting apparatus according to the invention of claim 8 is a device for predicting internal pressure fluctuation, N
A mathematical model for describing the amounts of Ox and CO emissions is constructed, and based on the constructed mathematical model, a region in which combustion oscillation is unlikely to occur and a region in which combustion vibration is likely to occur are obtained and output.

According to this configuration, a region in which combustion oscillation is unlikely to occur and a region in which combustion vibration is likely to occur are obtained based on a mathematical model for explaining internal pressure fluctuations, NOx emissions, and CO emissions constructed from plant data and weather data. Is performed and the result is output.

Further, the combustion vibration predicting apparatus according to the ninth aspect of the present invention is characterized in that: input means for inputting plant data, weather data, a limit value of internal pressure fluctuation, a NOx, CO regulation value, and the like;
Internal pressure fluctuation characteristic grasping means for mathematically modeling the internal pressure fluctuation of the combustor from the input plant data and weather data, and NOx emission characteristic grasping means for mathematically modeling the NOx emission from the input plant data and weather data A CO emission characteristic grasping means for mathematically modeling the CO emissions from the input plant data and weather data; and a limit value of the internal pressure fluctuation applied to the mathematical model obtained by the internal pressure fluctuation characteristic grasping means. The NOx regulation value is applied to the mathematical expression model obtained by the emission characteristic grasping means, and the CO regulation value is applied to the mathematical expression model obtained by the CO emission characteristic grasping means, whereby NOx and CO emissions are regulated. A safety area estimating means for obtaining an area which is less than or equal to the value and in which combustion oscillation is unlikely to occur, and a safety area estimating result obtained by the safety area estimating means. And output means for, those having a.

According to this structure, the internal pressure fluctuation characteristic grasping means forms a mathematical model of the internal pressure fluctuation of the combustor from the plant data and weather data inputted by the input means, and the NOx emission characteristic grasping means is inputted by the input means. Formulate NOx emissions from plant data and weather data into mathematical models,
The CO emission characteristic grasping means forms a mathematical model of the CO emission from the plant data and the weather data input by the input means, and the safety region estimating means applies the limit value of the internal pressure fluctuation and the NOx and CO regulation values to the mathematical model. Then, an area where the NOx and CO emissions are less than the regulation value and in which combustion oscillation is less likely to be generated is obtained, and the safety area estimation result is output from the output means.

The combustion vibration predicting apparatus according to the tenth aspect of the present invention includes an input means for inputting plant data, weather data, a limit value of internal pressure fluctuation, a NOx, CO control value, and the like; Focus setting means for selecting data to be used for mathematical modeling from data, means for grasping internal pressure fluctuation characteristics for mathematically modeling the internal pressure fluctuation of a combustor from selected plant data and weather data, and selected plant data and weather data From NOx
And an emission characteristic grasping means for mathematically modeling the CO emission, and applying a limit value of the internal pressure fluctuation to the mathematical model obtained by the internal pressure fluctuation characteristic grasping means, and a mathematical expression model obtained by the emission characteristic grasping means. NOx and CO
A safety area estimating means for obtaining an area in which NOx and CO emissions are less than or equal to the regulatory value and in which combustion oscillation is unlikely to occur, and an output means for outputting a safety area estimation result by the safety area estimating means. .

According to this configuration, the internal pressure fluctuation characteristic grasping means forms a mathematical model of the internal pressure fluctuation of the combustor from the plant data and the weather data input by the input means and selected by the focus setting means, and grasps the emission characteristic. Means for mathematically modeling NOx and CO emissions from the plant data and weather data selected by the focus setting means, and the safety region estimating means applying the internal pressure fluctuation limit value and the NOx and CO regulation values to the mathematical model. Thus, an area in which the NOx and CO emissions are less than the regulation value and in which fuel combustion oscillation is unlikely to occur is obtained, and the safety area estimation result is output from the output means.

[0028] In the combustion vibration predicting apparatus according to the eleventh aspect, the focus setting means may further include a plant data input from the input means based on an area specified by the input means or a setting mode. , To select weather data.

According to this configuration, the focus setting means selects the plant data and the weather data input from the input means based on the area specified by the input means or the setting mode.

A combustion vibration predicting apparatus according to a twelfth aspect of the present invention provides an input means for inputting plant data, weather data, internal pressure fluctuation limit values, NOx and CO control values, and the like. Focus determination means for selecting data to be used for mathematical modeling from data; internal pressure fluctuation characteristic grasping means for mathematically modeling internal pressure fluctuations of a combustor from selected plant data and weather data; selected plant data and weather data From NOx
And an emission characteristic grasping means for mathematically modeling the CO emission, and applying a limit value of the internal pressure fluctuation to the mathematical model obtained by the internal pressure fluctuation characteristic grasping means, and a mathematical expression model obtained by the emission characteristic grasping means. NOx and CO
The safety region estimating means for obtaining a region in which NOx and CO emissions are equal to or less than the regulation values and in which combustion oscillation is unlikely to occur, and the safety region estimation result obtained by the safety region estimating device, It has adjustment plan generating means for obtaining points to be measured, and output means for outputting the safety area estimation result by the safety area estimating means and the points to be measured by the adjustment plan generating means.

According to this configuration, the internal pressure fluctuation characteristic grasping means forms a mathematical model of the internal pressure fluctuation of the combustor from the plant data and the weather data input by the input means and selected by the focus determining means, and grasps the emission characteristic. Means for mathematically modeling NOx and CO emissions from the plant data and weather data selected by the focus determining means, and the safety region estimating means applies the internal pressure fluctuation limit value and the NOx and CO regulated values to the mathematical model. Then, an area in which the NOx and CO emissions are less than or equal to the regulation value and in which combustion oscillation is unlikely is obtained, and the adjustment plan generation means obtains a point to be measured next using the safety area estimation result by the safety area estimation means. Also, the safety area estimation result by the safety area estimation means and the next point to be measured by the adjustment plan generation means are output from the output means. It is.

In the combustion vibration predicting apparatus according to the thirteenth aspect of the present invention, the focus determination means may further include a mathematical model obtained based on plant data and weather data selected by the previous focus determination. The next focus is determined.

According to this configuration, the focus determining means determines the next focus based on the mathematical model obtained based on the plant data and weather data selected by the previous focus determination.

[0034]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a perspective view of a combustion vibration predicting apparatus according to an embodiment of the present invention.

Embodiment 1 FIG. 1 shows the configuration of Embodiment 1 of a combustion vibration prediction device according to the present invention. In FIG. 1, reference numeral 10 indicates the entire combustion vibration prediction device. Combustion vibration prediction device 10
Are internal pressure fluctuation characteristic grasping means 1 for modeling the internal pressure fluctuation of the combustor, combustion vibration area estimating means 2 for obtaining an area where combustion vibration is likely to occur, and a database 3 for storing plant data and weather data in time series. Input means 4 for inputting plant data, weather data, internal pressure fluctuation limit values, and the like.
And an output unit 5 for outputting a combustion vibration region estimation result. The input vibration unit 4 is connected to a plant 30 whose combustion vibration is to be predicted.

The internal pressure fluctuation characteristic grasping means 1 uses the data stored in the database 3 to construct a mathematical model for explaining the internal pressure fluctuation. For example, assuming that the number of combustors is n ₁ and the number of frequency bands to be modeled is n ₂ , the internal pressure fluctuation is modeled by a multiple regression model such as the following equation (1). Y _ij = a _{ij, 0} + a _{ij, 1} × X ₁₁ + a _{ij, 2} × X ₁₂ + a _{ij, 3} × X ₂₁ + a _{ij, 4} × X ₂₂ (1)

Here, Y _ij : the internal pressure fluctuation value of the j-th frequency band of the i-th combustor (i = 1
.. N ₁ , j = 1... N ₂ ) X ₁₁ : Value of manipulated variable 1 X ₁₂ : Value of manipulated variable 2 X ₂₁ : Value of state variable 1 that cannot be operated X ₂₂ : Value of state variable 2 that cannot be operated a _{ij, 0} , a _{ij, 1} , a _{ij, 2} , a _{ij, 3} , a _{ij, 4 are} coefficient parameters.

The internal pressure fluctuation characteristic grasping means 1 includes a database
3. Internal pressure fluctuation values organized and stored by time of day, operation
Using the quantity and the state quantity that cannot be operated
Data a _{ij, 0}, A_{ij, 1}, A_{ij, 2}, A_{ij, 3}, A_{ij, 4},
This is transmitted to the combustion oscillation region estimation means 2. Coefficient parameter
For example, a least squares method is used for solving the data.

The internal pressure fluctuation value referred to here is the plant 30
A / D conversion is performed on the data obtained from the pressure sensor (internal pressure sensor) 31 installed in the apparatus, and the result of the frequency analysis is divided into n ₂ frequency bands, and the maximum obtained within a certain time in each frequency band is obtained. It is an amplitude value. In the above description, for convenience of explanation, the operation amount is two variables and the state quantity that cannot be operated is two variables. However, the model expression is not limited to two variables.

The combustion vibration region estimating means 2 uses the mathematical model obtained by the internal pressure fluctuation characteristic grasping means 1 to find a region where combustion vibration is likely to occur.

[0041] For example, the operation amount 1, the operation amount 2, the state quantity 1 that can not be operated, the state quantity 2 that can not be operated, each _{X '11, X' 12,} X '21, X' i-th combustor when the ₂₂ The internal pressure fluctuation predicted value Y ′ _ij of the j-th frequency band is obtained by the following equation (2). _{Y 'ij = a ij, 0} + a ij, 1 × X' 11 + a ij, 2 × X '12 + a ij, 3 × X' 21 + a ij, 4 × X '22 ... (2)

Where a _{ij, 0} , a _{ij, 1} , a _{ij, 2} ,
a _{ij, 3} and a _{ij, 4} are coefficient parameters sent from the internal pressure fluctuation characteristic grasping means 1.

The internal pressure fluctuation in the j-th frequency band of the i-th combustor
Limits are set due to the structural aspects of the combustor and surrounding equipment.
ing. The j-th round of the i-th combustor transmitted from the input means 4
The limit value of the internal pressure fluctuation in the wavenumber band is Z_ijThen, Z_ij= A_{ij, 0}+ A_{ij, 1}× X '₁₁+ A_{ij, 2}× X '₁₂+ A_{ij, 3}× X '_{twenty one}+ A_{ij, 4}× X ' _{twenty two} … (3) X '₁₁, X '₁₂, X '_{twenty one}, X '_{twenty two}Will exist
You.

Now, suppose that the value of the state quantity 1 that cannot be operated and the value of the state quantity 2 that cannot be operated are input by the input means 4.
When these input values and sent to the combustion vibration region estimating unit 2, (3) out of the equation, X _'11, X' ₁₂ than is a constant, satisfying the expression (3) _{(X '1 1, X'} 12 ) Can be easily obtained. Αk (k = 1) transmitted from the input means 4
... The n ₃₎ becomes the _{gain, αk × Z ij = a ij} , 0 + a ij, 1 × X '11 + a ij, 2 × X' 12 + a ij, 3 × X '21 + a ij, 4 × X' 22 ... If (X ′ ₁₁ , X ′ ₁₂ ) is obtained as (4), n ₃ lines can be obtained for each frequency band of each combustor. FIG. 2 illustrates this. Here, the coefficient parameter a _{ij, 2}
Is positive, the upper side of the straight line is a region where combustion vibration is likely to occur, and the lower side is a region where combustion vibration is unlikely to occur. Conversely, if the coefficient parameter a _{ij, 2} is negative, the lower side of the straight line is a region where combustion vibration is likely to occur, and the upper side is a region where combustion vibration is less likely to occur.

The combustion oscillation region estimating means 2 transmits the limit value Z _ij (i = 1... N ₁ , j = 1... N ₂ ) of the internal pressure fluctuation in the j-th frequency band of the i-th combustor transmitted from the input means 4. , Gain αk
(K = 1... N ₃ ) and the values of the variables except for the specific two, and the coefficient parameter a transmitted from the internal pressure fluctuation characteristic grasping means 1.
_{ij, 0} , a _{ij, 1} , a _{ij, 2} , a _{ij, 3} , a _{ij, 4} (i = 1...
n ₁ , j = 1... n ₂ ), the above-mentioned straight lines are obtained for all the frequency bands of all the combustors. The area is obtained and transmitted to the output means 5.

The internal pressure fluctuation value, the operation amount, and the inoperable state amount are stored in the database 3 in a time-series manner arranged for each time, and when these data are transmitted from the input means 4, the database 3 It is additionally stored internally.

The input means 4 is provided outside the combustion vibration prediction device 10.
Plant data and weather data transmitted from the plant 30
The data is received and transmitted to the database 3. Plant de
Data and weather data are the above-mentioned internal pressure fluctuation values, manipulated variables,
Includes state quantities that cannot be produced. This input means
Keyboard and touch screen provided for 4
Control of the internal pressure fluctuation in the j-th frequency band of the i-th combustor
Limit value Z_ij(I = 1 ... n ₁, J = 1 ... n_Two), Gain αk
(K = 1 ... n_Three) And the values of the variables except for two specific ones
Then, it is transmitted to the combustion oscillation region estimation means 2.

The output means 5 outputs the estimation result transmitted from the combustion oscillation area estimation means 2. FIG. 3 shows an output example in which the combustion oscillation region is output with the horizontal axis being X ₁₁ and the vertical axis being X ₁₂ . In this example, the combustion oscillation region is expressed as a contour line for each gain αk, and the central portion is a region where combustion oscillation is less likely to occur, and the peripheral portion is a region where the combustion oscillation is more likely to occur. The output destination is a display device such as a CRT or a printing device provided in the output means.

As described above, according to the combustion vibration predicting apparatus 10, the combustion vibration generated in the combustor of the gas turbine is predicted by the mathematical model, and the combustion control system can be easily adjusted based on the prediction. It becomes possible,
It is possible to avoid equipment damage, improve the equipment utilization rate, and improve safety.

In the above embodiment, the plant data and the weather data are inputted from the plant 30. However, these data may be directly inputted manually from a keyboard provided in the input means 4. Also,
Although the model structure is described as a linear linear expression, it may be a second-order or higher-order model. In addition, although described as a model formula using the operation amount or the state amount that cannot be operated input from the plant 30, a value converted based on physical characteristics or the like may be used.

Embodiment 2 FIG. 4 shows the configuration of Embodiment 2 of a combustion vibration prediction device according to the present invention. In FIG. 4, reference numeral 20 indicates the entire combustion vibration prediction device. Combustion vibration prediction device 20
Are internal pressure fluctuation estimating means 12 for estimating the internal pressure fluctuation of the combustor.
And a database 13 for storing plant data and weather data in time series, an input means 14 for inputting plant data and weather data, and the like, and an output means 15 for outputting an internal pressure fluctuation estimation result. Is connected to a plant 30 whose combustion vibration is to be predicted.

The internal pressure fluctuation estimating means 12 stores the database 1
Estimating the expected value of the internal pressure fluctuation using the internal pressure fluctuation value, the operation amount, and the inoperable state amount at the latest time stored in 3;
The internal pressure fluctuation predicted value is transmitted to the database 13. For example, assuming that the number of combustors is n _l and the number of frequency bands is n ₂ , the internal pressure fluctuation predicted value is estimated using a multiple regression model as shown in the following equation (5). Y ′ _ij = a _{ij, 0} + a _{ij, 1} × X ₁₁ + a _{ij, 2} × X ₁₂ + a _{ij, 3} × X ₂₁ + a _{ij, 4} × X ₂₂ (5)

Here, Y ′ _ij is a predicted value of internal pressure fluctuation in the j-th frequency band of the i-th combustor (i = 1... N ₁ , j = 1... N ₂ ). X ₁₁ : Value of manipulated variable 1 X ₁₂ : Value of manipulated variable 2 X ₂₁ : Value of state variable 1 that cannot be operated X ₂₂ : Value of state variable 2 that cannot be operated a _{ij, 0} , a _{ij, 1} , a _{ij, 2} , a _{ij, 3} , a _{ij, 4} : This is the coefficient parameter.

The coefficient parameter is obtained by a preliminary analysis and is stored in the internal pressure fluctuation estimating means 12. The internal pressure fluctuation value referred to here is also obtained by A / D conversion of data obtained from a pressure sensor (internal pressure sensor) 31 installed in the plant 30.
The result of the conversion and frequency analysis is divided into n ₂ frequency bands, and is the maximum amplitude value obtained within a certain time in each frequency band.

In the above description, for convenience of explanation, the operation amount is set to 2
Although the model formula is described with variables and unoperable state quantities as two variables, this is not particularly limited to two variables. Further, the model structure is described as a linear linear expression, but may be a second-order or higher-order model or a nonlinear model such as a neural network. Further, the description is given as a model formula using the operation amount or the state amount that cannot be operated input from the plant 30, but a value converted based on a law such as a mass balance may be used.

In the database 13, the internal pressure fluctuation value, the operation amount, the inoperable state amount and the estimated internal pressure fluctuation value are arranged in time sequence and stored in a time series, and are stored in the input means 14 and the internal pressure fluctuation estimating means 12. When the data is transmitted, it is additionally stored in the database 13.

The input means 14 receives plant data and weather data transmitted from the plant 30 outside the combustion vibration prediction device 20 and transmits them to the database 13. The plant data and the weather data include the above-mentioned internal pressure fluctuation value, operation amount, and state amount that cannot be operated.

The output means 15 outputs data stored in the database 13. FIG. 5 shows an example in which the horizontal axis represents time and the vertical axis represents the measured internal pressure fluctuation value Y _ij and its predicted value Y ′ _ij in the j-th frequency band of the i-th combustor. The output destination is a display device such as a CRT or a printing device provided in the output means.

As described above, according to this embodiment, the estimated value of the internal pressure fluctuation and the actually measured value can be output simultaneously, and it can be determined whether the internal pressure fluctuation of the gas turbine combustor is at the expected level. Thus, it is possible to detect the occurrence of combustion vibration at an early stage, thereby avoiding equipment damage, improving the equipment utilization rate, and improving safety.

Third Embodiment FIG. 6 shows the configuration of a third embodiment of the combustion vibration prediction device according to the present invention. In FIG. 6, reference numeral 40 indicates the entire combustion vibration prediction device. Combustion vibration prediction device 40
Are internal pressure fluctuation characteristic grasping means 41 for modeling the internal pressure fluctuation of the combustor as a mathematical model, and a safe area estimating means 42 for finding a region where the NOx emission amount is equal to or less than a regulation value and in which combustion oscillation is hardly generated.
And a database 43 for storing plant data and weather data in time series, input means 44 for inputting plant data and weather data, a limit value of internal pressure fluctuation, a NOx limit value, and the like, and an output for outputting a safety region estimation result. Means 45;
A NOx emission characteristic grasping means 46 for mathematically modeling the NOx emission amount is provided. The input means 44 is connected to a plant 50 whose combustion vibration is to be predicted. In FIG. 6, reference numeral 49 denotes a connecting portion for connecting each means to each other (the same applies to other drawings).

The internal pressure fluctuation characteristic grasping means 41 uses the data stored in the database 43 to construct a mathematical model for explaining the internal pressure fluctuation. For example, assuming that the number of combustors is n _l and the number of frequency bands to be modeled is n ₂ , the internal pressure fluctuation is modeled by a multiple regression model as in the above equation (1).

The internal pressure fluctuation characteristic grasping means 41 uses the internal pressure fluctuation value, the operation amount, and the inoperable state amount arranged and stored for each time in the database 43, and uses the coefficient parameters a _{ij, 0 of} the above equation (1), a _{ij, 1} , a _{ij, 2} , a _{ij, 3} , and a _{ij, 4} are obtained and transmitted to the safe region estimation means 42. For solving the coefficient parameter, for example, the least squares method is used.

The internal pressure fluctuation value referred to here is the plant 50
A / D conversion is performed on data obtained from a pressure sensor (internal pressure sensor), not shown, which is installed in the system, and the result of frequency analysis is divided into n ₂ frequency bands, and each frequency band is obtained within a certain time. This is the maximum amplitude value. In the above description, for convenience of explanation, the operation amount is two variables and the state quantity that cannot be operated is two variables. However, the model expression is not limited to two variables.

The NOx emission characteristic grasping means 46 uses the data stored in the database 43 to construct a mathematical model for explaining the NOx emission. For example, the NOx emission amount is modeled by a multiple regression model represented by the following equation (6). E = b ₀ + b ₁ × X ₁₁ + b ₂ × X ₁₂ + b ₃ × X ₂₁ + b ₄ × X ₂₂ (6)

Here, E: NOx emission amount X ₁₁ : value of operation amount 1 X ₁₂ : value of operation amount 2 X ₂₁ : value of state amount 1 that cannot be operated X ₂₂ : value of state amount 2 that cannot be operated b ₀ , b ₁ , b ₂ , b ₃ , b ₄ : coefficient parameters

The NOx emission characteristic grasping means 46 uses the NOx emission amount, the operation amount, and the unoperable state amount arranged and stored for each time in the database 43 to obtain the coefficient parameters b ₀ , b of the above equation (6). ₁ , b ₂ , b ₃ , b ₄ are obtained and transmitted to the safe area estimation means 42. For solving the coefficient parameter, for example, the least squares method is used.

In the above description, for convenience of explanation, the operation amount is two variables, and the state quantity that cannot be operated is two variables. However, the model expression is not limited to the two variables.

The safety region estimating means 42 uses the mathematical model obtained by the internal pressure fluctuation characteristic grasping means 41 and the NOx emission characteristic grasping means 46 to determine an area where the NOx emission is equal to or less than the regulation value and in which combustion oscillation is unlikely to occur. Ask.

[0069] For example, the operation amount 1, the operation amount 2, the state quantity 1 that can not be operated, the state quantity 2 that can not be operated, each _{X '11, X' 12,} X '21, X' i-th combustor when the ₂₂ The internal pressure fluctuation predicted value Y'ij of the j-th frequency band is obtained by the above (2).
At this time, a _{ij, 0} , a _{ij, 1} , a _{ij, 2} , a _{ij, 0
ij, 3} and a _{ij, 4} are coefficient parameters sent from the internal pressure fluctuation characteristic grasping means 41.

The internal pressure fluctuation in the j-th frequency band of the i-th combustor has a limit value in terms of the structure of the combustor and surrounding equipment. The j-th combustion chamber transmitted from the input means 44
Assuming that the limit value of the internal pressure fluctuation in the frequency band is Z _ij , there are X ′ ₁₁ , X ′ ₁₂ , X ′ ₂₁ , and X ′ ₂₂ that satisfy the above expression (3).

Now, suppose that the values of the inoperable state quantity 1 and the inoperable state quantity 2 are input to the input means 44, and that these input values are transmitted to the safe area estimating means 42. , X ′ ₁₁ , and X ′ ₁₂ are constants, and (X ′ ₁₁ , X ′ ₁₂ ) satisfying the expression (3) can be easily obtained. Using the gain αk (k = 1... N ₃ ) transmitted from the input means 44,
If (X ′ ₁₁ , X ′ ₁₂ ) is obtained from the equation, n ₃ lines can be obtained for each frequency band of each combustor. FIG. 7 illustrates this. Here, the coefficient parameter a _{ij, 2}
Is positive, the upper side of the straight line is a region where combustion vibration is likely to occur, and the lower side is a region where combustion vibration is unlikely to occur. Conversely, if the coefficient parameter a _{ij, 2} is negative, the lower side of the straight line is a region where combustion vibration is likely to occur, and the upper side is a region where combustion vibration is less likely to occur.

The safety region estimating means 42 receives the limit value Z _ij (i = 1... N ₁ , j = 1... N ₂ ) of the internal pressure fluctuation in the j-th frequency band of the i-th combustor transmitted from the input means 44. Gain αk
(K = 1... N ₃ ) and the values of the variables except for two specific ones,
The coefficient parameters a _{ij, 0} , a _{ij, 1} , a _{ij, 2} , a _{ij, 3} , a _{ij, 4} (i = 1...) Transmitted from the internal pressure fluctuation characteristic grasping means 41.
n ₁ , j = 1... n ₂ ), the above-mentioned straight lines are obtained for all the frequency bands of all the combustors. Find the area. FIG. 8 illustrates this.

The safe area estimating means 42 performs the same processing as the above-described processing to control the NOx emission amount F transmitted from the input means 44 and the gain βk (k = 1... N ₃ ).
And the values of the variables except for the specific two, and the coefficient parameters b ₀ , b ₁ ,
From b ₂ , b ₃ , and b ₄ , an area where NOx is hardly generated and an area where NOx is easily generated are obtained. FIG. 9 illustrates this. Here, if positive coefficient parameter b _2, upper straight lines is likely to occur region, hardly lower is generated region of NOx. Conversely, if the coefficient parameter b ₂ is negative, the lower straight line is likely to occur region, the upper occurs hardly region of NOx.

Then, the safety region estimating means 42 determines NOx emission for the region where combustion oscillation is unlikely to occur and the region where combustion oscillation is likely to occur, and the region where NOx is unlikely to occur and the region where NOx is likely to occur based on the linear programming procedure. An area in which the amount is equal to or less than the regulation value and in which combustion oscillation is unlikely to occur is obtained and transmitted to the output means 45. FIG. 10 illustrates this.

In the database 43, the internal pressure fluctuation value, NO
The x-discharge amount, the operation amount, and the state amount that cannot be operated are arranged in time series and stored in chronological order. When these data are transmitted from the input means 44, they are additionally stored in the database 43.

The input means 44 receives plant data and weather data transmitted from the plant 50 outside the combustion vibration prediction device 40 and transmits them to the database 43. The plant data and weather data are the above-mentioned internal pressure fluctuation value, NOx
This includes the amount of discharge, the amount of operation, and the amount of state that cannot be operated. In addition, a device such as a keyboard or a touch screen provided to the input means 44 may be used to determine whether the j-th combustor
Limit value Z _ij (i = 1... N ₁ , j =
1 ... n ₂ ), NOx emission regulation value F, gain αk, β
k (k = 1... n ₃ ) and the values of the variables except for two specific ones are input and transmitted to the safe area estimation means 42.

For example, in the gas turbine 51 having the configuration shown in FIG.
In addition to Ox emissions, intake temperature, intake pressure, intake flow rate,
There are compressor outlet temperature, compressor outlet pressure, fuel flow, fuel temperature, fuel pressure, exhaust gas temperature, inlet guide vane angle, combustor bypass valve opening, fuel flow control valve opening, and the like. The fuel flow rate, fuel pressure, and fuel flow rate control valve include a main flame for main combustion and a pilot flame for main flame holding. The weather data includes atmospheric temperature, atmospheric pressure, humidity, and the like. The manipulated variable and the inoperable state variable used in the multiple regression model are selected from these. In FIG. 11, reference numeral 53 denotes a compressor, and reference numeral 5 denotes a compressor.
4 is a turbine, 55 is a combustor, 56 is a main fuel flow control valve, 57 is a pilot fuel flow control valve,
Reference numeral 58 denotes a combustor bypass valve, and reference numeral 59 denotes an inlet guide vane.

The output means 45 outputs the estimation result transmitted from the safe area estimation means 42. In FIG. 10, the horizontal axis is X ₁₁ ,
The vertical axis X ₁₂ shows an output example of outputting the safe area. In this example, the safety region is expressed as a contour line for each of the gains αk and βk, and the central portion is a region in which combustion vibration is less likely to occur, and the peripheral portion is a region in which combustion vibration is more likely to occur. The output destination is a display device such as a CRT or a printing device provided in the output unit 45.

As described above, according to the combustion vibration predicting device 40, the combustion vibration and the NOx emission generated in the combustor of the gas turbine are predicted by the mathematical model, and the combustion control is performed based on this. System adjustment can be facilitated, and equipment damage can be avoided, the equipment utilization rate can be improved, and safety can be improved. Therefore, by using the combustion vibration prediction device 40, for example, regarding the combustion control parameters of the gas turbine that have been conventionally adjusted based on the experience of a skilled adjuster, the NOx emission is equal to or less than the regulation value and the generation of combustion vibration Since the safety area that is difficult to perform is presented, for example, the on-site adjustment period of the gas turbine can be shortened, and the on-site adjustment can be easily performed even by a non-expert. Here, FIG. 12 specifically shows an example of an output result applicable to the adjustment of the combustion control system of the gas turbine.

In the above embodiment, the plant data and the weather data are inputted from the plant 50. However, these data may be directly inputted manually from a keyboard provided in the input means 44. In addition, the model structure is described as a linear linear expression, but may be a second-order or higher-order model, or may be a model to which a mutual term obtained by multiplying or dividing the operation amount or the state amount that cannot be operated is added. In addition, although described as a model formula using the operation amount or the state amount that cannot be operated input from the plant 50, a value converted based on physical characteristics or the like may be used.

Fourth Embodiment FIG. 13 shows the configuration of a fourth embodiment of the combustion vibration prediction device according to the present invention. FIG.
In the figure, reference numeral 60 indicates the entire combustion vibration prediction device. The combustion vibration prediction device 60 includes an internal pressure fluctuation characteristic grasping unit 41 that mathematically models the internal pressure fluctuation of the combustor, and a safe area estimation for obtaining a region where the NOx emission amount and the CO emission amount are equal to or less than the regulation values and in which the combustion oscillation hardly occurs. Means 62, a database 63 for storing plant data and weather data in time series, input means 64 for inputting plant data and weather data, internal pressure fluctuation limit values, NOx control values, CO control values, and the like. Output means 6 for outputting a region estimation result
5, NOx emission characteristic grasping means 46 for mathematically modeling NOx emissions, and CO for mathematically modeling CO emissions.
It has emission characteristic grasping means 67, and the plant 50 of which combustion vibration is to be predicted is connected to the input means 64.

The internal pressure fluctuation characteristic grasping means 41 and the NOx emission characteristic grasping means 46 are the same as those in the third embodiment. In the following, redundant description is omitted, and the third embodiment is omitted.
Only the points different from the above will be described.

The CO emission characteristic grasping means 67 uses the data stored in the database 63 to construct a mathematical model for explaining the CO emission. For example, the CO emission amount is modeled by a multiple regression model represented by the following equation (7). G = c ₀ + c ₁ × X ₁₁ + c ₂ × X ₁₂ + c ₃ × X ₂₁ + c ₄ × X ₂₂ (7)

Here, G: CO emission amount X ₁₁ : Value of manipulated variable 1 X ₁₂ : Value of manipulated variable 2 X ₂₁ : Value of inoperable state variable 1 X ₂₂ : Value of inoperable state variable 2 c ₀ , c ₁ , c ₂ , c ₃ , c ₄ : coefficient parameters

The CO emission characteristic grasping means 67 stores the CO emissions stored in the database 63 for each time,
The coefficient parameters c ₀ , c ₁ , c ₂ , c ₃ , and c ₄ of the above equation (7) are obtained using the operation amount and the state amount that cannot be operated, and are transmitted to the safe area estimation means 62. For solving the coefficient parameter, for example, the least squares method is used.

In the above description, for convenience of explanation, the operation amount is two variables and the state quantity that cannot be operated is two variables. However, the model expression is not limited to two variables.

The safety region estimating means 62 uses the mathematical model obtained by the internal pressure fluctuation characteristic grasping means 41, the NOx emission characteristic grasping means 46, and the CO emission characteristic grasping means 67 to determine both the NOx emission amount and the CO emission amount. A region that is equal to or less than the regulation value and in which combustion vibration is unlikely to occur is determined.

The method of finding the region where combustion vibration is less likely to occur, the region where combustion vibration is more likely to occur, the region where NOx is less likely to be generated and the region where NOx is more likely to occur is the same as in the third embodiment. Therefore, a method of obtaining an area where CO is less likely to be generated and an area where CO is more likely to be generated will be described below.

The safety region estimating means 62 determines whether combustion oscillation or NO
By the same processing as in the case of x, the regulated value H of the CO emission amount transmitted from the input means 64 and the gain γk (k = 1.
n ₃ ) and the values of the variables except for the specific two, and the coefficient parameter c ₀ transmitted from the CO emission characteristic grasping means 67,
From c ₁ , c ₂ , c ₃ , and c ₄ , an area where CO is less likely to be generated and an area where CO is more likely to be generated are obtained.

The safety region estimating means 62 calculates a region where combustion oscillation is hardly generated, a region where combustion vibration is easily generated, a region where NOx is hardly generated, a region where NOx is easily generated, a region where CO is hardly generated, and a region where COx is hardly generated. With respect to the region, a region in which both the NOx emission amount and the CO emission amount are equal to or less than the regulation value and in which combustion oscillation is hardly generated is obtained based on the procedure of the linear programming, and transmitted to the output means 65. FIG. 14 illustrates this. 14, if positive coefficient parameter c _2, upper straight lines is likely to occur region, hardly lower is generated region of CO. Conversely, if the coefficient parameter c ₂ is negative, the lower side of the straight line is an area where CO is easily generated, and the upper side is an area where CO is hardly generated.

The database 63 stores the internal pressure fluctuation value, NO
The x-emission amount, the CO-emission amount, the operation amount, and the inoperable state amount are arranged in time series and stored in chronological order, and when these data are transmitted from the input means 64, they are additionally stored in the database 63. Is done.

The input means 64 receives plant data and weather data transmitted from the plant 50 outside the combustion vibration prediction device 60 and transmits them to the database 63. Also,
From a device such as a keyboard or a touch screen provided in the input means 64, the limit value Z _ij (i = 1... N ₁ , j = 1... N ₂ ) of the internal pressure fluctuation in the j-th frequency band of the i-th combustor,
The control value F of the NOx emission amount, the control value H of the CO emission amount, the gains αk, βk, γk (k = 1... N ₃ ) and the values of the variables except for two specific values are input to the safety region estimation means 62. Send. The plant data and the weather data include the above-described internal pressure fluctuation value, NOx emission amount, CO emission amount, operation amount, and state amount that cannot be operated, and are described, for example, in the third embodiment with reference to FIG. This is the sum of various data and the amount of CO emissions.

The output means 65 outputs the estimation result transmitted from the safe area estimation means 62. In FIG. 14, the horizontal axis is X ₁₁ ,
The vertical axis X ₁₂ shows an output example of outputting the safe area. In this example, the safety region is expressed as a contour line for each of the gains αk, βk, and γk, and the central portion is a region in which combustion vibration is less likely to occur, and the peripheral region is more likely to occur.
The output destination is a display device such as a CRT or a printing device provided in the output unit 65.

As described above, according to the combustion vibration predicting device 60, the combustion vibration, NOx emissions, and CO emissions generated in the combustor of the gas turbine are predicted by a mathematical model. Based on this, it becomes possible to easily adjust the combustion control system, thereby avoiding equipment damage, improving the equipment utilization rate, and improving safety. Therefore, by using the combustion vibration predicting device 60, for example, the NOx control of the combustion control parameters of the gas turbine, which has been conventionally adjusted based on the experience of a skilled adjuster,
Since a safety region where the emission amount and the CO emission amount are equal to or less than the regulation values and in which combustion oscillation is hardly generated is presented, for example, it is possible to shorten the on-site adjustment period of the gas turbine, and it is possible to perform on-site adjustment even for a non-expert. Can be easily performed.

In the above embodiment, the plant data and the weather data are inputted from the plant 50. However, these data may be directly inputted manually from a keyboard provided in the input means 64. In addition, the model structure is described as a linear linear expression, but may be a second-order or higher-order model, or may be a model to which a mutual term obtained by multiplying or dividing the operation amount or the state amount that cannot be operated is added. In addition, although described as a model formula using the operation amount or the state amount that cannot be operated input from the plant 50, a value converted based on physical characteristics or the like may be used.

Fifth Embodiment FIG. 15 shows a combustion vibration predicting apparatus according to a fifth embodiment of the present invention.
Is shown. In FIG. 15, reference numeral 70 indicates the entire combustion vibration prediction device. Combustion vibration prediction device 7
0 is an internal pressure fluctuation characteristic grasping means 71 for modeling the internal pressure fluctuation of the combustor as a mathematical model; a safety region estimating means 72 for finding a region where the NOx emission amount and the CO emission amount are equal to or less than the regulation values and in which combustion oscillation is unlikely to occur; A database 63 for storing plant data and weather data in a time series; plant data and weather data; internal pressure fluctuation limit values; NOx regulation values;
An input unit 74 for inputting a regulation value of O, an output unit 65 for outputting a safety region estimation result, and an emission characteristic grasping unit 76 for numerically modeling the NOx emission amount and the CO emission amount
And a focus setting unit 78 for selecting data used for mathematical modeling. The input unit 74 is connected to the plant 50 whose combustion vibration is to be predicted.

The internal pressure fluctuation characteristic grasping means 71 has a function of selecting a database to be used for modeling a mathematical expression based on a selection result described in an appropriate area of the database 63. Other configurations, functions, and the like of the internal pressure fluctuation characteristic grasping means 71 are the same as those of the internal pressure fluctuation characteristic grasping means 41 of the third embodiment.

The emission characteristic grasping means 76 has a function of selecting a database to be used for mathematical modeling based on the selection result described in an appropriate area of the database 63. Other configurations, functions, and the like of the emission characteristic grasping means 76 have the respective configurations and functions of the NOx emission characteristic grasping means 46 of the third embodiment and the CO emission characteristic grasping means 67 of the fourth embodiment. Since it is the same as that of FIG.

The safety region estimating means 72 uses the mathematical model obtained by the internal pressure fluctuation characteristic grasping means 71 and the emission characteristic grasping means 76 to determine that both the NOx emission amount and the CO emission amount are equal to or less than the regulation value and that the occurrence of combustion oscillation. Find areas that are difficult to do.

The method of obtaining the region where combustion oscillation is unlikely to occur, the region where combustion oscillation is likely to occur, the region where NOx is unlikely to occur, the region where it is likely to generate CO, the region where CO is unlikely to occur, and the region where COX is likely to occur is any of: Since it is the same as the fourth embodiment, a duplicate description will be omitted.

The focus setting means 78 selects data corresponding to the focus setting information input by the input means 74 from the data stored in the database 63, and writes the selection result in an appropriate memory of the database 63. . Here, the focus setting information is not particularly limited, but is, for example, information such as an upper limit value and a lower limit value of each variable, a center point to be selected, and a maximum distance from the center point.

The input means 74 inputs focus setting information for selecting data used for mathematical expression modeling from a device such as a built-in keyboard or touch screen, and transmits the information to the focus setting means 78. Input means 74
Other configurations, functions, and the like of, except that the transmission destination of various data and the like input from a device such as a keyboard or a touch screen is the safety region estimation unit 72 (the safety region estimation unit 62 in the fourth embodiment). Since it is the same as the input means 64 of the fourth embodiment, a duplicate description will be omitted. Further, plant data and weather data are the same as in the fourth embodiment.

The output means 65 is the same as that of the fourth embodiment except that the supply source of the safety area estimation result is the safety area estimation means 72 (the safety area estimation means 62 in the fourth embodiment). , Overlapping description will be omitted. FIG. 16 shows an output example in which the safety area is output with the horizontal axis being X ₁₁ and the vertical axis being X ₁₂ . In this example, the safety region is expressed as a contour line for each of the gains αk, βk, and γk, and the central portion is a region in which combustion vibration is less likely to occur, and the peripheral region is more likely to occur.

As described above, according to the combustion vibration predicting apparatus 70, the combustion vibration, NOx emission and CO emission generated in the combustor of the gas turbine are predicted by the mathematical model, and the combustion vibration is hardly generated. It is necessary to obtain a safe area widely on a macro scale or with high precision on a micro scale. Therefore, it is easy to grasp the combustion vibration characteristics.
Further, by grasping the combustion oscillation characteristics, it is possible to easily adjust the combustion control system, and it is possible to avoid equipment damage, improve the equipment utilization rate, and improve safety.
For example, by using the combustion vibration prediction device 70, the NOx control of the combustion control parameters of the gas turbine conventionally adjusted based on the experience of a skilled adjuster
Since a safety region where the emission amount and the CO emission amount are equal to or less than the regulation values and in which combustion oscillation is hardly generated is presented, for example, it is possible to shorten the on-site adjustment period of the gas turbine, and it is possible to perform on-site adjustment even for a non-expert. Can be easily performed.

In the above-described embodiment, the plant data and the weather data are inputted from the plant 50. However, these data may be directly inputted manually from a keyboard provided in the input means 74. In addition, the model structure is described as a linear linear expression, but may be a second-order or higher-order model, or may be a model to which a mutual term obtained by multiplying or dividing the operation amount or the state amount that cannot be operated is added. In addition, although described as a model formula using the operation amount or the state amount that cannot be operated input from the plant 50, a value converted based on physical characteristics or the like may be used.

Sixth Embodiment FIG. 17 shows the configuration of a sixth embodiment of the combustion vibration prediction device according to the present invention. FIG.
In the figure, reference numeral 80 indicates the entire combustion vibration prediction device. The combustion vibration predicting device 80 includes an internal pressure fluctuation characteristic grasping unit 71 that mathematically models the internal pressure fluctuation of the combustor, and a safe area estimation for finding a region where the NOx emission amount and the CO emission amount are equal to or less than the regulation values and in which the combustion oscillation hardly occurs. Means 82, a database 63 for storing plant data and weather data in time series, input means 84 for inputting plant data and weather data, internal pressure fluctuation limit values, NOx control values, CO control values, and the like. Output means 85 for outputting the region estimation result and the points to be measured, NOx emission amount and C
Emission characteristics grasping means 76 for mathematically modeling O emissions
A focus determining means 88 for selecting data used for mathematical modeling, and an adjustment plan generating means 89 for obtaining a point to be measured next using the safety region estimation result. Target plant 50
Is connected.

Since the internal pressure fluctuation characteristic grasping means 71 and the discharge amount characteristic grasping means 76 are the same as those in the fifth embodiment, the duplicate description will be omitted.

The safety region estimating means 82 obtains a predicted optimum point where both the NOx emission amount and the CO emission amount are equal to or less than the regulation value and the combustion oscillation occurrence level is the smallest, based on the procedure of the linear programming, and determines the focus point. To the means 88 and the output means 85. Other configurations, functions, and the like of the safe area estimating means 82 are the same as those of the safe area estimating means 72 of the fifth embodiment, and thus redundant description will be omitted.

The method of obtaining the region in which combustion vibration is less likely to occur, the region in which combustion vibration is more likely to occur, the region in which NOx is less likely to be generated, the region in which COx is more likely to occur, the region in which CO is less likely to be generated, and the region in which COx is more likely to be determined, Since it is the same as the fourth embodiment, a duplicate description will be omitted.

The focus determining means 88 selects data corresponding to the initial focus setting information input by the input means 84 from the data stored in the database 63 at the initial stage of the adjustment, and stores the result of the selection in the database. Write in 63 appropriate memories. Here, the focus setting information is, for example, but not limited to, information such as an upper limit value and a lower limit value of each variable, and coordinates of a center point of focus.

The focus determining means 88 changes the initial focus setting information based on the predicted optimum point obtained by the safe area estimating means 82 to obtain new focus setting information. Further, the focus determining means 88
Makes a change to the current focus setting information based on the predicted optimum point obtained by the safe area estimation means 82, and sets the change as new focus setting information. Then, the focus determining means 88 selects data corresponding to the changed focus setting information from the data stored in the database 63, and records the selection result in an appropriate memory of the database 63. When the predicted optimum point is located outside the current focus, the focus moves in the direction of the predicted optimum point. Further, even if the predicted optimum point is located inside the focus, if the position is located at the periphery of the focus, the movement amount is reduced and the focus position is set again. FIG. 18 shows an example of the movement of the focus. In this example, the focus moves while partly overlapping, but the movement amount is not limited to this.

The adjustment plan generating means 89 searches the focus newly set by the focus determining means 88, for example, for a point at which data measurement of the actual gas turbine is sparse, and outputs the sparse point as a point for additional measurement. Transmit to the means 85. FIG. 18 shows this state. For example, it is assumed that the current focus is set in the area of focus 1 in FIG. 18, the data of the point indicated by the mark “x” has already been measured, and the predicted optimum point is the upper right direction in FIG. In this case, the focus moves in the direction of the expected optimum point (upper right direction in FIG. 18), and an area located in the upper right direction in FIG. At this time, the adjustment plan generation means 89 proposes to perform data measurement on a point newly included in the focus 2, for example, a point indicated by a triangle in FIG. 18.

The input means 84 inputs initial focus setting information for selecting data to be used for mathematical modeling from a device such as a built-in keyboard or touch screen, and transmits the information to the focus determining means 88. Input means 8
Other configurations, functions, and the like of the fourth embodiment except that the transmission destination of various data and the like input from a device such as a keyboard and a touch screen is the safety region estimation unit 82 (the safety region estimation unit 62 in the fourth embodiment). Since the input unit 64 is the same as the input unit 64 according to the fourth embodiment, a duplicate description will be omitted. Further, plant data and weather data are the same as in the fourth embodiment.

The output unit 85 outputs the estimation result transmitted from the safe area estimation unit 82, the focus area set by the focus determination unit 88, and the measurement points transmitted from the adjustment plan generation unit 89. The output destination is a display device such as a CRT or a printing device provided in the output unit 85. FIG. 18 shows an output example in which the horizontal axis is X ₁₁ and the vertical axis is X ₁₂ to output the safety region. In this example, the safety region is expressed as a contour line for each of the gains αk, βk, and γk, and the central portion is a region where combustion vibration is unlikely to occur,
This is an area that is more likely to occur in the peripheral area. In FIG. 18, rectangular areas indicated by focus 1, focus 2 and focus 3 show how the focus moves sequentially. In addition, crosses indicate points at which data has been measured at focus 1 and triangles and circles indicate data addition measurement points at focus 2 and focus 3, respectively.

As described above, according to the combustion vibration predicting device 80, the combustion vibration, NOx emissions, and CO emissions generated in the combustor of the gas turbine are predicted by the mathematical model, and the combustion vibration is hardly generated. A safe area is obtained with high accuracy, and a new measurement point is presented to search for a safer driving point. Therefore, when the NOx emissions and CO emissions are below the regulation values,
In addition, it is possible to obtain a point where combustion vibration is least likely to occur, that is, an optimum operation point. This facilitates adjustment of the combustion control system, and allows adjustment in a shorter period than before. In addition, it is possible to avoid equipment damage, improve the equipment utilization rate, and improve safety. For example, by using the combustion vibration prediction device 80, regarding the combustion control parameters of the gas turbine, which have been conventionally adjusted based on the experience of a skilled adjuster, the NOx emission and the CO emission are equal to or less than the regulation values and the combustion is controlled. Since the operating point where vibration is least likely to occur is presented, for example, the on-site adjustment period of the gas turbine can be shortened, and on-site adjustment can be easily performed even by a non-expert.

In the above embodiment, the plant data and the weather data are inputted from the plant 50. However, these data may be directly inputted manually from a keyboard provided in the input means 84. In addition, the model structure is described as a linear linear expression, but may be a second-order or higher-order model, or may be a model to which a mutual term obtained by multiplying or dividing the operation amount or the state amount that cannot be operated is added. In addition, although described as a model formula using the operation amount or the state amount that cannot be operated input from the plant 50, a value converted based on physical characteristics or the like may be used.

[0117]

As will be understood from the above description, according to the combustion vibration predicting apparatus (claim 1) of the present invention, combustion is performed based on a mathematical model describing internal pressure fluctuations constructed from plant data and weather data. An area where vibration is likely to occur and an area where vibration is unlikely to occur are obtained, and the results are output, so that it becomes possible to easily adjust the combustion control system, avoid equipment damage and improve equipment utilization, Safety can be improved.

Further, according to the combustion vibration predicting apparatus of the present invention, the internal pressure fluctuation characteristic grasping means forms a mathematical model of the internal pressure fluctuation of the combustor from the plant data and weather data inputted from the input means, and performs combustion. The vibration region estimating means applies the limit value of the internal pressure fluctuation to the mathematical model to obtain a region where combustion vibration is likely to occur, and the combustion vibration region estimation result is output to the output means, thereby facilitating adjustment of the combustion control system. This makes it possible to avoid equipment damage, increase the equipment utilization rate, and improve safety.

Further, according to the combustion vibration predicting apparatus of the present invention (claim 3), the plant data, weather data, and the like input by the input means are stored in a database in chronological order, and the internal pressure fluctuation characteristic grasping means is stored in the database. Since more data is acquired and the fluctuation of the internal pressure of the combustor is modeled as a mathematical model, a region where combustion oscillation is likely to occur is more accurately determined.

Further, according to the combustion vibration predicting device of the present invention (claim 4), the internal pressure fluctuation estimating means estimates the internal pressure fluctuation of the combustor from the plant data and the weather data input by the input means, and estimates the fluctuation. Since the internal pressure fluctuation estimation result is output to the output means, it is possible to determine whether the internal pressure fluctuation of the gas turbine combustor is at the expected level, it is possible to detect the occurrence of combustion vibration at an early stage, and the equipment is damaged. It is possible to avoid problems, improve the capacity factor, and improve safety.

Further, according to the combustion vibration predicting device (claim 5) of the present invention, the plant data, weather data, and the like input by the input means are stored in a database in chronological order. Since the expected value of the internal pressure fluctuation is estimated from the latest time data stored in the database, the prediction of the internal pressure fluctuation is accurately performed.

Further, according to the combustion vibration predicting apparatus of the present invention (claim 6), it is difficult to generate combustion vibration based on a mathematical model describing the internal pressure fluctuation and NOx emission amount constructed from plant data and weather data. The region and the region that is likely to occur are determined, and the result is output, so that it is possible to easily adjust the combustion control system, and to avoid equipment damage, improve equipment utilization, and improve safety. It becomes possible.

Further, according to the combustion vibration predicting apparatus of the present invention, the internal pressure fluctuation characteristic grasping means forms a mathematical model of the internal pressure fluctuation of the combustor from the plant data and the weather data inputted by the input means, and NOx The emission characteristic grasping means forms a mathematical model of the NOx emission from the plant data and the weather data inputted by the input means, and the safety region estimating means applies the internal pressure fluctuation limit value and the NOx regulation value to the mathematical model to obtain NOx. An area in which the amount of emission is less than the regulation value and in which combustion oscillation is unlikely to occur is obtained, and the safety area estimation result is output from the output means, so that it becomes possible to easily adjust the combustion control system, thereby avoiding equipment damage. It is possible to improve equipment utilization and safety.

Further, according to the combustion vibration predicting apparatus of the present invention (claim 8), the combustion vibration is estimated based on a mathematical model describing the internal pressure fluctuation, NOx emission and CO emission constructed from plant data and weather data. Area where the occurrence of
Since the area where the occurrence is likely to occur is obtained and the result is output, it is possible to easily adjust the combustion control system, and it is possible to avoid equipment damage, improve the equipment utilization rate, and improve safety. Become.

Further, according to the combustion vibration predicting apparatus of the present invention, the internal pressure fluctuation characteristic grasping means forms a mathematical model of the internal pressure fluctuation of the combustor from the plant data and the weather data inputted by the input means, and NOx The emission characteristic grasping means formulates the NOx emissions from the plant data and the weather data input by the input means, and the CO emission characteristic grasping means formulates the CO emissions from the plant data and the weather data input by the input means. The safety area estimating means calculates the limit value of the internal pressure fluctuation and the NO
By applying the regulated values of x and CO, a region where the NOx and CO emissions are less than the regulated value and in which combustion oscillation is unlikely to be generated is obtained, and the safety region estimation result is output from the output means. It is possible to avoid equipment damage, improve the equipment utilization rate, and improve safety.

Further, according to the combustion vibration predicting device of the present invention, the internal pressure fluctuation characteristic grasping means is provided with a combustor based on plant data and weather data inputted by the input means and selected by the focus setting means. Of the internal pressure fluctuation of the internal pressure fluctuation, the emission characteristic grasping means mathematically models the NOx and CO emissions from the plant data and the weather data selected by the focus setting means, and the safety region estimating means converts the internal pressure fluctuation into a mathematical model. Limit value and NO
Since the control values of x and CO are applied to obtain a region where the NOx and CO emissions are less than the control values and in which the occurrence of fuel-combustion oscillation is difficult, and the safety region estimation result is output from the output means, the combustion control system Adjustment can be facilitated, and equipment damage can be avoided, the equipment utilization rate can be improved, and safety can be improved.

Further, according to the combustion vibration predicting apparatus of the present invention (claim 11), the area specified by the input means,
Or, based on the setting mode, the focus setting means
Since the plant data and the weather data input from the input means are selected, it is possible to macro-widely obtain a safe area where combustion vibration is unlikely to occur, or to obtain microscopically high-precision.

Further, according to the combustion vibration predicting device of the present invention, the internal pressure fluctuation characteristic grasping means is provided with a combustor based on plant data and weather data inputted by the input means and selected by the focus determining means. Formula of the internal pressure fluctuation of the NOx and CO emissions from the plant data and the weather data selected by the focus determining means, and the safety region estimating means converts the internal pressure fluctuation into a mathematical model. Limit value and NO
Applying the regulated values of x and CO to obtain a region where NOx and CO emissions are less than or equal to the regulated values and in which combustion oscillation is unlikely to occur, and the adjustment plan generating means uses the safety area estimation result by the safety area estimating means. Then, the point to be measured next is obtained, and the safety area estimation result by the safety area estimation means and the point to be measured by the adjustment plan generation means are output from the output means, so that the combustion control system can be easily adjusted. As a result, it is possible to avoid equipment damage, improve the equipment utilization rate, and improve safety.

According to the combustion vibration predicting apparatus of the present invention (claim 13), the focus determination means is based on a mathematical model obtained based on the plant data and weather data selected by the previous focus determination. Since the next focus is determined, it is easy to search for the optimum driving point.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a combustion vibration prediction device according to a first embodiment of the present invention.

FIG. 2 is a principle diagram regarding a method of estimating a combustion oscillation region in the combustion oscillation prediction device according to the first embodiment of the present invention.

FIG. 3 is a graph showing an output example of a combustion vibration region in the combustion vibration prediction device according to the first embodiment of the present invention.

FIG. 4 is a configuration diagram of a combustion vibration prediction device according to a second embodiment of the present invention.

FIG. 5 is a graph showing an output example in the combustion vibration prediction device according to the second embodiment of the present invention.

FIG. 6 is a configuration diagram of a combustion vibration prediction device according to a third embodiment of the present invention.

FIG. 7 is a principle diagram relating to a method of estimating a safety region in a combustion vibration prediction device according to Embodiment 3 of the present invention.

FIG. 8 is a principle diagram relating to a method for estimating a safety region in a combustion vibration prediction device according to Embodiment 3 of the present invention.

FIG. 9 is a principle diagram regarding a method of estimating a NOx regulation level in a combustion vibration prediction device according to a third embodiment of the present invention.

FIG. 10 is a graph showing an output example of a safety region in the combustion vibration prediction device according to the third embodiment of the present invention.

FIG. 11 is a configuration diagram schematically illustrating an example of a system configuration of a gas turbine to which the present invention can be applied.

FIG. 12 is a graph showing a specific application example of the combustion vibration prediction device according to the third embodiment of the present invention.

FIG. 13 is a configuration diagram of a combustion vibration prediction device according to a fourth embodiment of the present invention.

FIG. 14 is a graph showing an output example of a safety region in the combustion vibration prediction device according to the fourth embodiment of the present invention.

FIG. 15 is a configuration diagram of a combustion vibration prediction device according to a fifth embodiment of the present invention.

FIG. 16 is a graph showing an output example of a safety region in the combustion vibration prediction device according to the fifth embodiment of the present invention.

FIG. 17 is a configuration diagram of a combustion vibration prediction device according to a sixth embodiment of the present invention.

FIG. 18 is a graph showing an output example of a safety region in the combustion vibration prediction device according to the sixth embodiment of the present invention.

FIG. 19 is a configuration diagram showing a conventional example.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Combustion vibration prediction apparatus 1 Internal pressure fluctuation characteristic grasping means 2 Combustion vibration area estimation means 3 Database 4 Input means 5 Output means 20 Combustion vibration prediction apparatus 12 Internal pressure fluctuation estimation means 13 Database 14 Input means 15 Output means 30 Plant 40 Combustion vibration prediction apparatus 41 Internal pressure fluctuation characteristic grasping means 42 Safety area estimating means 43 Database 44 Input means 45 Output means 46 NOx emission characteristic grasping means 50 Plant 60 Combustion vibration prediction device 62 Safety area estimating means 63 Database 64 Input means 65 Output means 67 CO emission Characteristic grasping means 70 Combustion vibration predicting device 71 Internal pressure fluctuation characteristic grasping means 72 Safety area estimating means 74 Input means 76 Emission characteristic grasping means 78 Focus setting means 80 Combustion vibration predicting device 82 Safety area estimating means 84 Input means 85 Output Means 88 focus determination unit 89 proposed adjustment generating unit

Claims (13)

[Claims] 1. A formula model for explaining internal pressure fluctuation is constructed from plant data and meteorological data, and a region where combustion oscillation is likely to occur and a region where combustion vibration is unlikely to occur are obtained and output based on the constructed formula model. Combustion vibration prediction device. 2. Input means for inputting plant data, weather data, a limit value of internal pressure fluctuation, etc., internal pressure fluctuation characteristic grasping means for mathematically modeling the internal pressure fluctuation of a combustor from the input plant data and weather data, A combustion oscillation region estimating unit that obtains a region where combustion oscillation is likely to occur by applying a limit value of the internal pressure variation to the mathematical model obtained by the internal pressure variation characteristic grasping unit; A combustion vibration predicting device, comprising: output means for outputting the following. 3. A data base for storing plant data, weather data, and the like input by the input means in time series, wherein the internal pressure fluctuation characteristic grasping means obtains data from the database and detects the internal pressure fluctuation of the combustor. The combustion vibration predicting apparatus according to claim 2, wherein the apparatus is modeled by a mathematical model. 4. An input means for inputting plant data or weather data, an internal pressure fluctuation estimating means for estimating an internal pressure fluctuation of a combustor from the input plant data or weather data, and an internal pressure fluctuation estimating means. A combustion vibration prediction device, comprising: output means for outputting an internal pressure fluctuation estimation result. 5. A database for storing plant data, weather data, and the like input by the input unit in a time series, wherein the internal pressure fluctuation estimating unit detects the internal pressure fluctuation based on the latest time data stored in the database. The combustion vibration prediction device according to claim 4, wherein the prediction value is estimated. 6. A mathematical model for explaining internal pressure fluctuations and NOx emissions is constructed from plant data and weather data, and a region in which combustion oscillation is less likely to occur and a region in which combustion vibration is more likely are determined and output based on the constructed mathematical model. A combustion vibration prediction device, characterized in that: 7. An input means for inputting plant data, weather data, a limit value of internal pressure fluctuation, a NOx control value, etc., and an internal pressure fluctuation for formulating an internal pressure fluctuation of a combustor from the input plant data and weather data. Characteristic grasping means; NOx emission characteristic grasping means for mathematically modeling NOx emissions from the input plant data and weather data; and internal pressure fluctuation limit value applied to the mathematical expression model obtained by the internal pressure fluctuation characteristic grasping means. A safety region estimating means for applying a NOx regulation value to the mathematical model obtained by the NOx emission characteristic grasping means to find a region where the NOx emission amount is equal to or less than the regulation value and in which combustion vibration is unlikely to occur; Output means for outputting a safety area estimation result obtained by the safety area estimation means. 8. A mathematical model for explaining internal pressure fluctuation, NOx, and CO emissions is constructed from plant data and weather data, and a region in which combustion oscillation is unlikely to occur and a region in which combustion vibration is likely to occur are determined based on the constructed mathematical model. A combustion vibration prediction device characterized by outputting. 9. An input means for inputting plant data, weather data, internal pressure fluctuation limit values, NOx, CO regulation values, and the like, and formulating an internal pressure fluctuation of a combustor from input plant data and weather data. Internal pressure fluctuation characteristic grasping means, NOx emission characteristic grasping means for mathematically modeling NOx emissions from input plant data and weather data, and CO for mathematically modeling CO emissions from input plant data and weather data Emission amount characteristic grasping means, a limit value of internal pressure fluctuation is applied to the mathematical model obtained by the internal pressure fluctuation characteristic grasping means, and a NOx regulation value is applied to the mathematical expression model obtained by the NOx emission characteristic grasping means. By applying the regulated value of CO to the mathematical model obtained by the CO emission characteristic grasping means, NOx and CO
A safety area estimating means for obtaining an area in which the emission amount is equal to or less than a regulation value and in which combustion oscillation is unlikely to occur, and an output means for outputting a safety area estimation result by the safety area estimating means. Combustion vibration prediction device. 10. Input means for inputting plant data, weather data, internal pressure fluctuation limit values, NOx, CO control values, etc., and a focus for selecting data to be used for mathematical modeling from the input plant data and weather data. Setting means; internal pressure fluctuation characteristic grasping means for mathematically modeling the internal pressure fluctuation of the combustor from the selected plant data and weather data; and emissions for mathematically modeling the NOx and CO emissions from the selected plant data and weather data. Volume characteristic grasping means, applying a limit value of internal pressure fluctuation to the mathematical model obtained by the internal pressure fluctuation characteristic grasping means, and applying NOx and CO regulation values to the mathematical expression model obtained by the emission characteristic grasping means. Safety area estimating means for finding an area where NOx and CO emissions are less than or equal to a regulation value and in which combustion oscillation is unlikely to occur; Combustion oscillation prediction apparatus characterized in that it has an output means for outputting the safety region estimation result by the region estimation means. 11. The apparatus according to claim 10, wherein said focus setting means selects plant data and weather data input from said input means based on an area designated by said input means or a setting mode. Combustion vibration prediction apparatus according to any one of the above. 12. Input means for inputting plant data, weather data, internal pressure fluctuation limit values, NOx, CO control values, etc., and a focus for selecting data to be used in mathematical modeling from the input plant data and weather data. Determining means; means for grasping the internal pressure fluctuation characteristics of the combustor from the selected plant data and weather data as a mathematical model; and emissions for modeling the NOx and CO emissions from the selected plant data and the weather data. Volume characteristic grasping means, applying a limit value of internal pressure fluctuation to the mathematical model obtained by the internal pressure fluctuation characteristic grasping means, and applying NOx and CO regulation values to the mathematical expression model obtained by the emission characteristic grasping means. Safety area estimating means for finding an area where NOx and CO emissions are less than or equal to a regulation value and in which combustion oscillation is unlikely to occur; Adjustment plan generating means for obtaining a point to be measured next using the safety area estimation result by the area estimating means, and a safety area estimation result by the safety area estimating means and a point to be measured by the adjustment plan generating means are output. A combustion vibration prediction device, comprising: output means. 13. The apparatus according to claim 12, wherein said focus determining means determines a next focus based on a mathematical model obtained based on plant data and weather data selected by a previous focus determination. Combustion vibration prediction apparatus according to any one of the above.