A confidence matrix-based gas turbine simulation model design method

Abstract

The application discloses a gas turbine simulation model design method based on a confidence matrix, comprising the following steps: defining a confidence matrix: creating a confidence matrix with a size of n*n, wherein n represents the number of gas working medium flow parameters, and each element of the confidence matrix represents the confidence or correlation strength between two gas working medium flow parameters; generating relevant gas working medium flow: according to the confidence matrix, relevant gas working medium flow is generated, for each gas working medium flow parameter i, a random number generator is selected, and the values of other parameters related to the parameter i in the confidence matrix are used to generate relevant gas working medium flow; introducing uncertainty and variability: introducing randomness or disturbance when generating relevant gas working medium flow; repeating data generation: repeating steps 2 and 3. The gas working medium flow data generated by the application will match the distribution characteristics and correlation of the actual data, and can be used for evaluation, optimization and analysis of the gas turbine simulation model.

Description

Technical Field

[0001] This invention relates to the field of simulation technology, and in particular to a method for designing a gas turbine simulation model based on a confidence matrix. Background Technology

[0002] When generating a simulation model of a gas turbine, many parameters need to be considered, such as: Thermal system parameters: including working fluid flow rate, working fluid temperature, working fluid pressure, inlet and outlet temperatures, pressure loss, etc. These parameters are key to describing the gas turbine's operating state and thermodynamic cycle. Gas turbine configuration and geometry parameters: including hub diameter, number of blades, blade geometry, blade angles, etc. These parameters determine the gas turbine's structure and performance characteristics, significantly impacting energy conversion efficiency and output power. Material properties: including the material strength, thermal conductivity, and heat resistance of blades and other components. These parameters are crucial for evaluating the gas turbine's reliability and durability. Control system parameters: including the response speed, sensitivity, and control strategy of regulating valves, turbine governors, pressure controllers, etc. These parameters affect the gas turbine's operational stability and responsiveness. Efficiency parameters: including mechanical efficiency, thermal efficiency, and power generation efficiency. These parameters are used to evaluate the gas turbine's energy conversion efficiency and performance. Environmental conditions: including ambient temperature, altitude, humidity, etc. These parameters affect the thermodynamic cycle and heat dissipation of gas turbines, especially in high-temperature or high-altitude regions. However, current technology does not provide a clear description of the working gas flow rate of gas turbines. Summary of the Invention

[0003] To address the above problems, this invention provides a method for designing a gas turbine simulation model based on a confidence matrix.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0005] A method for designing a gas turbine simulation model based on a confidence matrix includes the following steps:

[0006] Step 1, Define the confidence matrix: Create a confidence matrix of size n×n, where n represents the number of parameters for gas working fluid flow rate, and each element of the confidence matrix represents the confidence or correlation strength between two gas working fluid flow rate parameters;

[0007] Step 2, Generate relevant gas working fluid flow rates: Based on the confidence matrix, generate relevant gas working fluid flow rates. For each gas working fluid flow rate parameter i, select a random number generator and use the values ​​of other parameters related to parameter i in the confidence matrix to generate the relevant gas working fluid flow rates.

[0008] Step 3, introduce uncertainty and variability: introduce randomness or disturbance when generating the relevant gas working fluid flow rate;

[0009] Step 4, Repeat Data Generation: Repeat steps 2 and 3 to generate multiple sets of relevant gas working fluid flow data, using the same confidence matrix and random number generator settings each time.

[0010] Further: Step 1 includes:

[0011] Step 101: Collect gaseous working fluid flow data;

[0012] Step 102: Using the collected gas working fluid flow rate data samples, calculate the covariance matrix between the gas working fluid flow rate parameters;

[0013] Step 103: Standardize the covariance matrix;

[0014] Step 104: Fill the standardized correlation coefficients into the corresponding positions in the confidence matrix.

[0015] Further: Step 2 includes:

[0016] Step 201: Based on the existing gas working fluid flow rate data samples, calculate the mean vector and covariance matrix of the relevant gas working fluid flow rates. The mean vector contains the average value of each gas working fluid flow rate parameter, and the covariance matrix represents the covariance between the parameters.

[0017] Step 202: Based on the mean vector and covariance matrix of the relevant gas working fluid flow rate, generate a set of random standard normal distribution samples. Each sample is a vector corresponding to the number of gas working fluid flow rate parameters, where each element is a random number drawn from the standard normal distribution.

[0018] Step 203: Perform a linear transformation to convert the random standard normal distribution samples generated in step 202 into multivariate normal distribution samples that conform to the expected mean and covariance.

[0019] Step 204: By applying the inverse function of the multivariate Gaussian distribution, the multivariate normal distribution sample is transformed into the actual gas working fluid flow rate value;

[0020] Step 205: Repeat steps 202 to 204 to generate multiple sets of relevant gas working fluid flow data. Each time, the same mean vector and covariance matrix settings are used to simulate the diversity of uncertainty and variability.

[0021] Furthermore, step 3 includes:

[0022] Step 301: Determine the standard deviation of the error or variation;

[0023] Step 302: Use a Gaussian distribution random number generator to generate a random number that follows a standard normal distribution;

[0024] Step 303: Perform a linear transformation on the random numbers generated in step 302 to make them conform to the expected mean and standard deviation;

[0025] Step 304: Add the random number obtained in step 303 to the generated relevant gas working fluid flow rate value to simulate the error or variation in the actual measurement;

[0026] Step 305: Repeat steps 302 to 304 to generate multiple sets of relevant gas working fluid flow data with errors or variations. Each time, use the same standard deviation and random number generator settings to simulate the diversity of uncertainty and variability.

[0027] The technological advancements achieved by this invention compared to existing technologies are as follows:

[0028] This method can describe the correlation between parameters: a confidence matrix can be used to represent the correlation or dependency between parameters. By setting the values ​​of the elements in the matrix, the confidence level or correlation strength between parameters can be determined. For example, if some parameters have high confidence or correlation in the matrix, then the changes between these parameters will be more consistent or correlated when generating simulation data.

[0029] This method can simulate uncertainty and variability: the operating state and performance of a gas turbine can be affected by a variety of uncertain factors. By introducing uncertainty or randomness into the confidence matrix, these effects can be simulated, and parameter variability can be taken into account when generating simulation data. This allows for a more accurate simulation of uncertainties in actual operation. Attached Figure Description

[0030] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0031] In the attached diagram:

[0032] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0033] The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of the present invention will now be described with reference to the accompanying drawings.

[0034] like Figure 1 As shown, a method for designing a gas turbine simulation model based on a confidence matrix is ​​as follows:

[0035] Step 1: Define the confidence matrix

[0036] First, create a confidence matrix of size n×n, where n represents the number of parameters related to the gas flow rate. Each element of the confidence matrix represents the confidence level or correlation strength between two gas flow rate parameters. The value of each element can be determined using correlation coefficients, covariance, or other correlation measures. High values ​​indicate a strong correlation, while low values ​​indicate a weak correlation or lack of correlation.

[0037] Step 2: Generate the relevant working gas flow rate

[0038] Based on the confidence matrix, relevant gas flow rates are generated. For each gas flow rate parameter i, a random number generator is selected, and the values ​​of other parameters related to parameter i in the confidence matrix are used to generate the relevant gas flow rates. Linear models, Gaussian models, or other correlation models can be used to generate the relevant gas flow rate values. In this way, by considering correlation, the generated gas flow rates will be consistent with or correlated with the values ​​of other parameters.

[0039] Step 3: Introduce uncertainty and variability

[0040] To account for the uncertainty and variability of the working gas flow rate, randomness or perturbation can be introduced when generating the relevant working gas flow rate. This can be achieved by adding a random error term or random perturbation during the generation process. For example, a random value can be obtained from a random number generator that conforms to a certain distribution (such as a Gaussian distribution) and added to the generated relevant working gas flow rate value to simulate errors or variations in actual measurements.

[0041] Step 4: Repeat the data generation

[0042] As needed, steps 2 and 3 can be repeated multiple times to generate multiple sets of relevant gaseous working fluid flow data. Each time, the same confidence matrix and random number generator settings are used, but different gaseous working fluid flow data are obtained to simulate the diversity of uncertainty and variability.

[0043] Through the above steps, a confidence matrix can be used to describe the correlation, simulation uncertainty, and variability among gas working fluid flow parameters. The generated correlated gas working fluid flow data can more accurately reflect the correlation and variability in actual operation, thereby supporting the evaluation, optimization, and analysis of gas turbine simulation models.

[0044] Specifically, step 1 includes:

[0045] Step 101: Collect gaseous working fluid flow data

[0046] First, it is necessary to collect some data samples related to the flow rates of the working gas. These data samples can be actual observations or known simulation data. Collecting sufficient data samples will help to accurately estimate the correlation between the flow rates of the working gas.

[0047] Step 102: Calculate the covariance matrix of the gas working fluid flow parameters.

[0048] Using the collected samples of gaseous working fluid flow rate data, the covariance matrix between the gaseous working fluid flow rate parameters is calculated. The covariance matrix is ​​a symmetric matrix whose elements represent the covariance values ​​between corresponding gaseous working fluid flow rate parameters. Covariance measures the strength of the linear relationship between two variables; positive values ​​indicate a positive correlation, negative values ​​indicate a negative correlation, and zero indicates no correlation.

[0049] Step 103: Standardize the covariance matrix

[0050] Since the value of covariance is affected by the scale of the variables, the covariance matrix needs to be standardized in order to compare and normalize the covariances of different variables. Standardization can be achieved by calculating the correlation coefficient, which is obtained by dividing the covariance by the standard deviation of the corresponding variable.

[0051] Step 104: Fill the standardized correlation coefficients into the confidence matrix.

[0052] The standardized correlation coefficients obtained in step 103 are then filled into the corresponding positions in the confidence matrix. Each element represents the correlation strength between the two gaseous working fluid flow parameters, and the value range is typically between -1 and 1. 1 indicates a perfect positive correlation, -1 indicates a perfect negative correlation, and 0 indicates no correlation.

[0053] By following the steps above, covariance can be used to determine the element values ​​of the confidence matrix between the gas working fluid flow parameters. This confidence matrix can be used to generate relevant gas working fluid flow data, simulate the correlation between parameters, and account for uncertainties and variability in gas turbine simulation models.

[0054] Specifically, step 2 includes:

[0055] Step 201: Calculate the mean and covariance matrix of the relevant gas working fluid flow rates.

[0056] Based on existing samples of gaseous working fluid flow rates, calculate the mean vector and covariance matrix for the relevant gaseous working fluid flow rates. The mean vector contains the average value of each gaseous working fluid flow rate parameter, and the covariance matrix represents the covariance between the parameters.

[0057] Step 202: Generate random standard normal distribution samples

[0058] Based on the mean vector and covariance matrix of the relevant gas working fluid flow rates, a set of random standard normal distribution samples is generated. Each sample is a vector corresponding to the number of gas working fluid flow rate parameters, where each element is a random number drawn from a standard normal distribution (mean 0, variance 1).

[0059] Step 203: Perform a linear transformation

[0060] The random standard normal distribution samples generated in step 2 are transformed into multivariate normal distribution samples that conform to the expected mean and covariance through a linear transformation. This can be achieved by multiplying the random samples by the square root of the covariance matrix and adding the mean vector.

[0061] Step 204: Apply the inverse function of the Gaussian distribution

[0062] By applying the inverse function of the multivariate Gaussian distribution, a multivariate normally distributed sample can be transformed into a real gaseous working fluid flow value. This can be achieved by using the inverse function of the cumulative distribution function (CDF) of the multivariate Gaussian distribution (also known as the percentile function or inverse Gaussian function).

[0063] Step 205: Repeatedly generate data

[0064] As needed, steps 202 through 204 can be repeated multiple times to generate multiple sets of relevant gaseous working fluid flow data. Each time, the same mean vector and covariance matrix settings are used, but different gaseous working fluid flow data are obtained to simulate the diversity of uncertainty and variability.

[0065] By following the steps above, a Gaussian model can be used to generate relevant gas working fluid flow rates. This generated gas working fluid flow rate data will match the distribution characteristics and correlations of actual data, and can be used for the evaluation, optimization, and analysis of gas turbine simulation models.

[0066] Specifically, step 3 includes:

[0067] Step 301: Determine the standard deviation of the error or variation.

[0068] First, it is necessary to determine the standard deviation of the error or variation, which is a measure of the magnitude of the error or variation. The larger the standard deviation, the greater the magnitude of the error or variation.

[0069] Step 302: Generate random numbers using a Gaussian distributed random number generator.

[0070] Use a Gaussian distribution random number generator to generate a random number that follows a standard normal distribution (mean 0, standard deviation 1).

[0071] Step 303: Perform a linear transformation

[0072] The random numbers generated in step 302 are linearly transformed to conform to the desired mean and standard deviation. This can be achieved by multiplying the generated random number by the standard deviation and adding the desired mean.

[0073] Step 304: Add to the relevant working gas flow rate value

[0074] The random number obtained in step 303 is added to the generated relevant gas working fluid flow rate value to simulate errors or variations in actual measurements. Adding the random number to the relevant gas working fluid flow rate value introduces uncertainty and variability, making the generated data closer to reality.

[0075] Step 305: Repeatedly generate data

[0076] As needed, steps 302 through 304 can be repeated multiple times to generate multiple sets of relevant gaseous working fluid flow data with errors or variations. Each time, the same standard deviation and random number generator settings are used, but different gaseous working fluid flow data are obtained to simulate the diversity of uncertainty and variability.

[0077] By following the steps above, a Gaussian-distributed random number generator can be used to simulate errors or variations in actual measurements. The resulting gaseous working fluid flow data will contain randomness, more realistically reflecting measurement errors or variations in parameters under actual conditions.

[0078] Specifically, step 4 includes:

[0079] Step 401: Determine the number of times to generate repeatedly

[0080] First, determine the number of times the data needs to be generated, which depends on how many sets of relevant gas working fluid flow data you want to generate.

[0081] Step 402: Set the seed for the random number generator

[0082] Before starting repeated generation, a seed needs to be set for the random number generator. The seed is an initial value that determines the random sequence generated by the random number generator. By setting the same seed, you can ensure that the same random number sequence is obtained each time.

[0083] Step 403: Repeatedly generate data

[0084] For each repeated generation, follow these steps:

[0085] 4031 Reset Random Number Generator

[0086] Before each generation, the random number generator needs to be reset to its initial state. This ensures that each generation starts from the same point.

[0087] 4032 generates related gas working fluid flow rate

[0088] Using a confidence matrix and a random number generator, a set of relevant gaseous working fluid flow data is generated as described previously. This may involve steps such as generating relevant gaseous working fluid flow values, introducing uncertainties and variability.

[0089] 4033 stores the generated data

[0090] The generated relevant gaseous working fluid flow data can be stored using suitable data structures such as lists, arrays, and files.

[0091] 4034 Update random number generator state

[0092] After each generation, the state of the random number generator is updated so that a different random sequence can be generated in the next generation. Ensure that the state of the random number generator is updated before the next repeated generation.

[0093] Step 404: Repeat the data generation until the specified number of times is reached.

[0094] Repeat step 403 to generate data a specified number of times. The same confidence matrix and random number generator settings are used each time, but different gas working fluid flow data are obtained each time because the state of the random number generator is reset each time.

[0095] By following the steps above, relevant gaseous working fluid flow data can be generated repeatedly to simulate the diversity of uncertainties and variability. Each generated data point can be stored and analyzed for the evaluation, optimization, and analysis of the gas turbine simulation model.

[0096] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for designing a gas turbine simulation model based on a confidence matrix, characterized in that, Includes the following steps: Step 1, Define the confidence matrix: Create a confidence matrix of size n×n, where n represents the number of parameters for gas working fluid flow rate, and each element of the confidence matrix represents the confidence or correlation strength between two gas working fluid flow rate parameters; Step 2, Generate relevant gas working fluid flow rates: Based on the confidence matrix, generate relevant gas working fluid flow rates. For each gas working fluid flow rate parameter i, select a random number generator and use the values ​​of other parameters related to parameter i in the confidence matrix to generate the relevant gas working fluid flow rates. Step 3, introduce uncertainty and variability: introduce randomness or disturbance when generating the relevant gas working fluid flow rate; Step 4, Repeat Data Generation: Repeat steps 2 and 3 to generate multiple sets of relevant gas working fluid flow data, using the same confidence matrix and random number generator settings each time. Step 3 includes: Step 301: Determine the standard deviation of the error or variation; Step 302: Use a Gaussian distribution random number generator to generate a random number that follows a standard normal distribution; Step 303: Perform a linear transformation on the random numbers generated in step 302 to make them conform to the expected mean and standard deviation; Step 304: Add the random number obtained in step 303 to the generated relevant gas working fluid flow rate value to simulate the error or variation in the actual measurement; Step 305: Repeat steps 302 to 304 to generate multiple sets of relevant gas working fluid flow data with errors or variations. Each time, use the same standard deviation and random number generator settings to simulate the diversity of uncertainty and variability.

2. The method for designing a gas turbine simulation model based on a confidence matrix according to claim 1, characterized in that, Step 1 includes: Step 101: Collect gaseous working fluid flow data; Step 102: Using the collected gas working fluid flow rate data samples, calculate the covariance matrix between the gas working fluid flow rate parameters; Step 103: Standardize the covariance matrix; Step 104: Fill the standardized correlation coefficients into the corresponding positions in the confidence matrix.

3. The method for designing a gas turbine simulation model based on a confidence matrix according to claim 1, characterized in that, Step 2 includes: Step 201: Based on the existing gas working fluid flow rate data samples, calculate the mean vector and covariance matrix of the relevant gas working fluid flow rates. The mean vector contains the average value of each gas working fluid flow rate parameter, and the covariance matrix represents the covariance between the parameters. Step 202: Based on the mean vector and covariance matrix of the relevant gas working fluid flow rate, generate a set of random standard normal distribution samples. Each sample is a vector corresponding to the number of gas working fluid flow rate parameters, where each element is a random number drawn from the standard normal distribution. Step 203: Perform a linear transformation to convert the random standard normal distribution samples generated in step 202 into multivariate normal distribution samples that conform to the expected mean and covariance. Step 204: By applying the inverse function of the multivariate Gaussian distribution, the multivariate normal distribution sample is transformed into the actual gas working fluid flow rate value; Step 205: Repeat steps 202 to 204 to generate multiple sets of relevant gas working fluid flow data. Each time, the same mean vector and covariance matrix settings are used to simulate the diversity of uncertainty and variability.

Images