A method, system, and electronic equipment for determining the confidence level of a gas turbine simulation model.

Abstract

This invention discloses a method, system, and electronic equipment for determining the confidence level of a gas turbine simulation model, relating to the field of confidence level assessment technology. The method includes: acquiring experimental data of the gas turbine and corresponding simulation data of the gas turbine simulation model; the experimental data includes experimental data under multiple steady-state operating conditions and multiple dynamic operating conditions, and the simulation data includes simulation data under multiple steady-state operating conditions and multiple dynamic operating conditions; one experimental data point corresponds to one simulation data point; using a spatial cosine similarity evaluation algorithm, the confidence level between each experimental data point and its corresponding simulation data is calculated, thereby obtaining multiple bottom-level confidence levels; weighting each bottom-level confidence level is performed based on the analytic hierarchy process (AHP), and the confidence level of the gas turbine simulation model is determined based on all weighted bottom-level confidence levels. This invention realizes the assessment of the confidence level of a gas turbine simulation model, quantifying the accuracy of the gas turbine simulation model in characterizing the gas turbine.

Description

Technical Field

[0001] This invention relates to the field of confidence assessment technology, and in particular to a method, system and electronic equipment for determining the confidence of a gas turbine simulation model. Background Technology

[0002] Before a gas turbine is put into service, it undergoes extensive testing, including safety, reliability, fault analysis, and structural optimization. Testing with a physical gas turbine requires collecting multiple sets of data, potentially leading to long testing cycles and low accuracy of measurements. Furthermore, shipboard gas turbines have high power outputs, making testing extremely costly and requiring significant investment of manpower and resources, resulting in substantial waste and environmental pollution. Therefore, simulation modeling has emerged as a solution. It uses a mathematical model of the gas turbine to simulate its operation using computer technology. This method overcomes the drawbacks of long testing cycles and high costs, and simulation modeling can be performed as long as the gas turbine's operation and the characteristics of its components are understood. However, the actual operation of a gas turbine inevitably involves uncertainties that cannot be described by a mathematical model, causing errors between the simulation model and reality. For example, under sudden load changes, the turbine speed in the model will inevitably deviate from the actual operating speed. Therefore, it is very important to evaluate the confidence level of the simulation model. The confidence level obtained after the evaluation directly determines whether the gas turbine model can be applied to engineering practice. However, there is a lack of methods for evaluating the confidence level of the gas turbine model. Summary of the Invention

[0003] The purpose of this invention is to provide a method, system, and electronic equipment for determining the confidence level of a gas turbine simulation model, which realizes the confidence level of the gas turbine simulation model and quantifies the accuracy of the gas turbine simulation model in depicting the gas turbine entity.

[0004] To achieve the above objectives, the present invention provides the following solution:

[0005] A method for determining the confidence level of a gas turbine simulation model includes:

[0006] Acquire test data of the gas turbine and simulation data of the corresponding gas turbine simulation model; the test data includes test data under multiple steady-state operating conditions and test data under multiple dynamic operating conditions, and the simulation data includes simulation data under multiple steady-state operating conditions and simulation data under multiple dynamic operating conditions; one set of test data corresponds to one set of simulation data;

[0007] The confidence scores between each experimental data and the corresponding simulation data are calculated using the spatial cosine similarity evaluation algorithm, thereby obtaining multiple underlying confidence scores.

[0008] The confidence levels of each bottom layer are weighted using the analytic hierarchy process (AHP), and the confidence level of the gas turbine simulation model is determined based on all weighted bottom layer confidence levels.

[0009] Optionally, after acquiring the test data of the gas turbine and the simulation data of the corresponding gas turbine simulation model, the method further includes:

[0010] The experimental data and the simulation data are sequentially subjected to outlier removal, interval normalization, noise filtering, and sensor correction.

[0011] Optionally, acquire the test data of the gas turbine and the corresponding simulation data of the gas turbine simulation model, including:

[0012] During the steady-state process, the parameters of the gas turbine under each steady-state condition are obtained, resulting in test data under multiple steady-state conditions. The parameters under the steady-state conditions include: high-pressure compressor speed, low-pressure compressor speed, power turbine speed, exhaust temperature, and load.

[0013] During the steady-state process, the indicators of the gas turbine simulation model under each steady-state condition are obtained, and simulation data under multiple steady-state conditions are obtained.

[0014] During the dynamic process, the indicators of the gas turbine under each dynamic operating condition are acquired, and test data under multiple operating conditions are obtained. The indicators under the dynamic operating conditions include: the power turbine speed change diagram during the load reduction process, the power turbine speed change diagram during the loading process, and the power turbine speed change diagram during the load shedding process.

[0015] During the dynamic process, the indicators of the gas turbine simulation model under each dynamic operating condition are obtained, and simulation data under multiple dynamic operating conditions are obtained.

[0016] Optionally, the confidence levels of each of the lower-level confidence levels are weighted based on the analytic hierarchy process (AHP), and the confidence level of the gas turbine simulation model is determined based on all weighted lower-level confidence levels. Specifically, this includes:

[0017] Based on the time proportion of each steady-state condition in a preset time period, a judgment matrix for the steady-state sub-criteria layer is constructed.

[0018] Based on the time proportion of each dynamic working condition in a preset time period, a judgment matrix for the dynamic sub-criteria layer is constructed.

[0019] Construct the judgment matrix of the criterion layer;

[0020] Consistency checks are performed on the judgment matrices of the steady-state sub-criteria layer and the dynamic sub-criteria layer respectively, and the judgment matrices that fail the consistency check are re-established.

[0021] The weights of each indicator are determined based on the judgment matrix of the criterion layer, the judgment matrix of the steady-state sub-criterion layer that has passed the consistency test, and the judgment matrix of the dynamic sub-criterion layer that has passed the consistency test.

[0022] The confidence level of the gas turbine simulation model is calculated based on the weights of all indicators and their corresponding underlying confidence levels.

[0023] Optionally, the experimental data and the simulation data are sequentially subjected to outlier removal, interval normalization, noise filtering, and sensor correction, specifically including:

[0024] The 3σ criterion was used to remove outliers from the experimental data and the simulation data, respectively, to obtain the outlier-removed experimental data and the outlier-removed simulation data.

[0025] The maximum-minimum normalization method was used to perform interval normalization on the experimental data and simulation data after outlier removal, respectively, to obtain the normalized experimental data and normalized simulation data.

[0026] Kalman filtering was used to filter noise from the normalized experimental data and the normalized simulation data, respectively, to obtain the filtered experimental data and the filtered simulation data.

[0027] The median method was used to perform sensor correction on the filtered experimental data and the filtered simulation data, respectively, to obtain the sensor-corrected experimental data and the sensor-corrected simulation data.

[0028] Optionally, the confidence level of the gas turbine simulation model is calculated based on the weights of all indicators and their corresponding underlying confidence levels, specifically including:

[0029] Multiply the weight of each indicator by its corresponding underlying confidence level to obtain multiple weighted underlying confidence levels;

[0030] The confidence level of the gas turbine simulation model is obtained by summing all the weighted underlying confidence levels.

[0031] A confidence determination system for a gas turbine simulation model includes:

[0032] The data acquisition module is used to acquire test data of the gas turbine and simulation data of the corresponding gas turbine simulation model; the test data includes test data under multiple steady-state operating conditions and test data under multiple dynamic operating conditions, and the simulation data includes simulation data under multiple steady-state operating conditions and simulation data under multiple dynamic operating conditions; one set of test data corresponds to one set of simulation data;

[0033] The underlying confidence calculation module is used to calculate the confidence between each experimental data and the corresponding simulation data using the spatial cosine similarity evaluation algorithm, thereby obtaining multiple underlying confidence levels.

[0034] The confidence level determination module is used to assign weights to each of the underlying confidence levels based on the analytic hierarchy process (AHP), and to determine the confidence level of the gas turbine simulation model based on all the weighted underlying confidence levels.

[0035] An electronic device includes a memory and a processor, the memory storing a computer program, and the processor running the computer program to enable the electronic device to perform the above-described method for determining the confidence level of a gas turbine simulation model.

[0036] Optionally, the memory is a readable storage medium.

[0037] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0038] This invention discloses a method, system, and electronic equipment for determining the confidence level of a gas turbine simulation model. It comprehensively considers the confidence levels of steady-state and dynamic operating conditions, thereby obtaining the confidence level of the gas turbine simulation model and quantifying the accuracy of the gas turbine simulation model in depicting the gas turbine entity. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a schematic diagram of the confidence determination method for a gas turbine simulation model provided in Embodiment 1 of the present invention;

[0041] Figure 2 This is a flowchart illustrating the Analytic Hierarchy Process (AHP).

[0042] Figure 3 A schematic diagram of the intelligent confidence assessment process for a three-axis gas turbine simulation model;

[0043] Figure 4 The simulation diagram shows the change in rotational speed of the power turbine during the load reduction process;

[0044] Figure 5 The simulation diagram shows the change in rotational speed of the power turbine during the loading process;

[0045] Figure 6 A simulation diagram of the turbine speed change during load shedding;

[0046] Figure 7 This is a graph showing the change in turbine speed during the unloading process.

[0047] Figure 8 The graph shows the turbine speed variation during the loading process.

[0048] Figure 9 The graph shows the speed change during the load shedding turbine test.

[0049] Figure 10 This is a diagram illustrating the overall indicators;

[0050] Figure 11 This is a comparison chart of the low-pressure compressor speed before and after filtering;

[0051] Figure 12 This is a comparison chart of the low-pressure compressor speed before and after sensor correction. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] The purpose of this invention is to provide a method, system, and electronic device for determining the confidence level of a gas turbine simulation model, aiming to realize the confidence level of the gas turbine simulation model and quantify the accuracy of the gas turbine simulation model in depicting the gas turbine entity.

[0054] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0055] Example 1

[0056] Figure 1 This is a schematic flowchart of the method for determining the confidence level of a gas turbine simulation model provided in Embodiment 1 of the present invention. Figure 1 As shown, the method for determining the confidence level of the gas turbine simulation model in this embodiment includes:

[0057] Step 101: Obtain the test data of the gas turbine and the simulation data of the corresponding gas turbine simulation model.

[0058] The test data includes test data under multiple steady-state conditions and test data under multiple dynamic conditions, and the simulation data includes simulation data under multiple steady-state conditions and simulation data under multiple dynamic conditions; one test data corresponds to one simulation data.

[0059] As an optional implementation, step 101 includes:

[0060] During the steady-state process, the parameters of the gas turbine under each steady-state condition are obtained, resulting in test data under multiple steady-state conditions. The parameters under the steady-state conditions include: high-pressure compressor speed, low-pressure compressor speed, power turbine speed, exhaust temperature, and load.

[0061] During the steady-state process, the indicators of the gas turbine simulation model under each steady-state condition are obtained, resulting in simulation data under multiple steady-state conditions.

[0062] During the dynamic process, the indicators of the gas turbine under each dynamic operating condition are acquired, and test data under multiple operating conditions are obtained. The indicators under the dynamic operating conditions include: the power turbine speed change diagram during the load reduction process, the power turbine speed change diagram during the loading process, and the power turbine speed change diagram during the load shedding process.

[0063] During the dynamic process, the indicators of the gas turbine simulation model under each dynamic operating condition are obtained, and simulation data under multiple dynamic operating conditions are obtained.

[0064] As an optional implementation, after step 101, the method further includes:

[0065] The experimental and simulation data were sequentially subjected to outlier removal, interval normalization, noise filtering, and sensor correction. Interval normalization makes the evaluation data more standardized and the evaluation results more reliable; outlier removal eliminates unreasonable data; sensor correction addresses the problem of sensor measurement errors; and noise filtering filters out system noise and measurement noise.

[0066] As an optional implementation, the experimental data and simulation data are sequentially subjected to outlier removal, interval normalization, noise filtering, and sensor correction, specifically including:

[0067] The 3σ criterion was used to remove outliers from both the experimental and simulation data, resulting in outlier-removed experimental and simulation data.

[0068] The maximum-minimum normalization method was used to perform interval normalization on the experimental data and simulation data after outlier removal, respectively, to obtain normalized experimental data and normalized simulation data.

[0069] Kalman filtering was used to filter noise from both the normalized experimental data and the normalized simulation data, resulting in filtered experimental data and filtered simulation data.

[0070] The median method was used to perform sensor correction on the filtered experimental data and the filtered simulation data, respectively, to obtain the sensor-corrected experimental data and the sensor-corrected simulation data.

[0071] Step 102: Calculate the confidence scores between each experimental data point and the corresponding simulation data using the spatial cosine similarity evaluation algorithm, thereby obtaining multiple underlying confidence scores.

[0072] Specifically, spatial cosine similarity is defined as follows: The parameters of the two objects to be evaluated are constructed as two spatial vectors, and the relative deviation between the two objects is measured by the cosine value of the angle between these two vectors. The larger the cosine value of the angle between the two spatial vectors, the closer the angle is to zero degrees, and the more similar the two objects are. Conversely, the smaller the cosine value of the angle between the two spatial vectors, the lower the similarity between the two objects is considered.

[0073] Step 103: Assign weights to the confidence levels of each bottom layer based on the analytic hierarchy process (AHP), and determine the confidence level of the gas turbine simulation model based on all weighted bottom layer confidence levels.

[0074] As an optional implementation method, such as Figure 2 As shown, step 103 specifically includes:

[0075] Based on the time proportion of each steady-state condition in a preset time period, a judgment matrix for the steady-state sub-criteria layer is constructed.

[0076] Based on the time proportion of each dynamic working condition in a preset time period, a judgment matrix for the dynamic sub-criteria layer is constructed.

[0077] Construct the judgment matrix for the criterion layer.

[0078] Consistency checks are performed on the judgment matrices of the steady-state sub-criteria layer and the dynamic sub-criteria layer respectively, and the judgment matrices that fail the consistency check are re-established.

[0079] The weights of each indicator are determined based on the judgment matrix of the criterion layer, the judgment matrix of the steady-state sub-criterion layer that has passed the consistency test, and the judgment matrix of the dynamic sub-criterion layer that has passed the consistency test.

[0080] The confidence level of the gas turbine simulation model is calculated based on the weights of all indicators and their corresponding underlying confidence levels.

[0081] Specifically, the judgment matrix is ​​constructed based on the relative importance of the indicators. A nine-level scaling method is used to determine the importance of the indicators. The nine-level scaling method is shown in Table 1.

[0082] Table 1. Nine-level scale method

[0083] <![CDATA[b ij ]]> The relative importance of indicator i compared to indicator j 1 Equally important 3 Slightly important 5 Obviously important 7 Strongly important 9 Extremely important 2、4、6、8 The median of the above adjacent judgments reciprocal <![CDATA[b ji =1 / b ij ]]>

[0084] Where i represents the i-th indicator, j represents the j-th indicator, and b represents the importance value. ijb represents a scale indicating the relative importance of the i-th indicator to the j-th indicator; ji This represents a scale indicating the importance of the j-th indicator relative to the i-th indicator.

[0085] Consistency check: First, calculate the consistency index C. I :

[0086] Where, λ max is the largest eigenvalue of the judgment matrix; n is the order of the judgment matrix.

[0087] Secondly, calculate the consistency ratio.

[0088] Among them, R I The average random consistency index, Rn, is related to the order n of the judgment matrix. I As shown in Table 2.

[0089] Table 2. Relationship between average random consistency index and order.

[0090] n 1 2 3 4 5 6 7 8 9 <![CDATA[R I ]]> 0 0 0.52 0.89 1.12 1.26 1.36 1.41 1.46

[0091] The weights are determined using the arithmetic mean method, specifically:

[0092] The normalization calculation formula is:

[0093]

[0094] in, b represents a scale indicating the relative importance of the i-th indicator to the j-th indicator after normalization; ij This represents a scale indicating the relative importance of the i-th indicator to the j-th indicator;

[0095] The formula for calculating the indicator weight is:

[0096]

[0097] Among them, w i represents the weight of the i-th indicator, and n represents the dimension of the judgment matrix.

[0098] As an optional implementation method, the confidence level of the gas turbine simulation model is calculated based on the weights of all indicators and their corresponding underlying confidence levels, specifically including:

[0099] The weights of each indicator are multiplied by their corresponding underlying confidence scores to obtain multiple weighted underlying confidence scores.

[0100] The confidence scores of the gas turbine simulation model are obtained by summing all the weighted underlying confidence scores.

[0101] The following section uses a certain type of three-shaft gas turbine as an example (with a load of 22030 kW under full load) to conduct intelligent confidence assessment of the gas turbine simulation model. The process is as follows: Figure 3 As shown.

[0102] (1) Collect experimental and simulation data of the steady-state process of the gas turbine.

[0103] Test data is collected through gas turbine commissioning, and simulation data is collected by running a gas turbine simulation model.

[0104] The stop time for the gas turbine simulation model is set to 165 seconds, with a step size of 0.01 seconds. The compressor inlet pressure P2 is 101325 Pa, and the guide vane angle is 0 degrees.

[0105] The steady-state process was divided into six steady-state operating conditions, with corresponding loads of 22.18MW, 21.36MW, 17.01MW, 12.85MW, 8.62MW, and 4.34MW. Five parameters were selected as evaluation parameters for each operating condition: high-pressure compressor speed, low-pressure compressor speed, power turbine speed, exhaust temperature, and load. Tables 3 and 4 show the experimental and simulation data for the steady-state operating conditions, respectively.

[0106] Table 3. Test data under steady-state conditions

[0107]

[0108] Table 4 Simulation data under steady-state conditions

[0109]

[0110] Where MW stands for megawatt, r / min for revolutions per minute, and K for Kelvin.

[0111] The dynamic process includes three aspects: sudden load reduction, sudden load increase, and load shedding. Specifically, it is divided into 11 dynamic operating conditions. Sudden load reduction includes decreasing from 100% load to 80% load, from 80% load to 60% load, from 60% load to 40% load, from 40% load to 20% load, and from 20% load to 10% load. Sudden load increase includes increasing from 10% load to 20% load, from 20% load to 40% load, from 40% load to 60% load, from 60% load to 80% load, and from 80% load to 100% load. Load shedding is decreasing from 100% load to 10% load. Furthermore, the single-dimensional time series of the power turbine speed is selected as the evaluation parameter for each dynamic operating condition. Figures 4-9 As shown, the dynamic operating conditions of the gas turbine and its simulation model are obtained during the dynamic process, thus yielding the following results: Figure 10The figure shown is a total index that includes both steady-state and dynamic data.

[0112] (2) Data preprocessing of experimental and simulation data

[0113] An automated program based on MATLAB was developed for data preprocessing. Data preprocessing includes outlier removal, interval normalization, noise filtering, and sensor correction.

[0114] Because the data sampling step is short and the amount of data in the database is huge, 400 points are selected for each dynamic working condition: 100 points are selected before the mutation point and 300 points are selected after the mutation point.

[0115] Data preprocessing methods include:

[0116] First, outlier removal is performed using the 3σ criterion. Since the experimental data is collected via sensors, it may contain random errors, systematic errors, sampling biases, sensor failures, and other interference factors. Therefore, the collected data needs to undergo quality checks to provide accurate input for subsequent data preprocessing. The formula for the 3σ criterion is: μ - 3σ <x i <μ+3σ.

[0117] Where μ is the mean of a set of data, and σ is the standard deviation of a set of data. If a data point is outside this range, it is discarded.

[0118] Outlier removal was performed on the data in this example, and it was found that the data length of all groups remained unchanged, indicating that the experimental data and simulation data were collected well and there was no unreasonable data.

[0119] Secondly, the data is normalized using the Minimax normalization (MMN) method. Normalization is a necessary prerequisite for all subsequent data preprocessing. The MMN method formula is:

[0120] Where max is the maximum value in the set of data, min is the minimum value in the set of data, and x * This is the normalized value of data x.

[0121] Next, Kalman filtering is applied to the data. The Kalman filter processes the system's state vector through two processes: "prediction" and "verification," obtaining the optimal estimate. For each state vector, three values ​​are obtained: the prior state value, the optimal estimate, and the true state value. The Kalman gain is the core of the filtering process; its essence is the filter's bias towards the accuracy of the predicted and measured values. If the Kalman gain approaches zero, the predicted state value is trusted; conversely, the measured value is trusted. The Kalman gain is important because it connects these three quantities, continuously correcting the prior value through state transition calculations.

[0122] The five important equations of Kalman filtering are as follows:

[0123]

[0124]

[0125]

[0126]

[0127]

[0128] in, These are the posterior state estimates at time k and time k-1, respectively. P represents the prior state estimate at time k; k Let be the posterior estimated covariances at time k; Let be the prior estimated covariances at time k+1 and time k, respectively; K is the Kalman gain; A is the state transition matrix; B is the control variable matrix; u k z is the state control vector; k H is the measurement vector; H is the transformation matrix from the state vector to the measurement vector; R is the covariance matrix of the measurement noise; Q is the system noise covariance matrix.

[0129] Taking a sudden change in load from 100% to 80% as an example, the comparison before and after filtering is as follows: Figure 11 , Figure 11The horizontal axis represents time, and the vertical axis represents the low-pressure compressor speed. The broken line represents the low-pressure compressor speed versus time before filtering, and the dotted line represents the low-pressure compressor speed versus time after filtering. The graph before filtering shows a large negative slope after load reduction, while the graph after filtering shows a slightly slower rate of descent at the corresponding point. In actual operation, sudden load changes do not instantly cause changes in the power turbine speed; there is a lag. Therefore, the graph after filtering is closer to actual operating conditions. During the curve's ascent, the slopes of both graphs are not significantly different, and the time to reach the steady-state point is roughly the same. The shapes of the graphs before and after filtering are very similar, so the filtered graph is closer to the true data value.

[0130] Finally, sensor correction is performed. The median method, while correcting, can protect the signal edges, preventing them from becoming blurred. Taking a sudden change in load from 100% to 80% as an example, the comparison before and after correction is as follows: Figure 12 , Figure 12 The horizontal axis represents time, the vertical axis represents the low-pressure compressor speed, the broken line represents the low-pressure compressor speed-time line before correction, and the dotted line represents the low-pressure compressor speed-time line after correction.

[0131] The output is the median of the window array. By utilizing the advantages of the median method, overshoot data in the simulation data can be corrected, making the dynamic data curve smoother and closer to the actual situation.

[0132] (3) Calculate the steady-state confidence of the gas turbine simulation model based on the spatial cosine similarity method.

[0133] The spatial cosine similarity evaluation algorithm is developed using MATLAB, and the data used for evaluation comes from preprocessed data.

[0134] The formula for calculating spatial cosine similarity is:

[0135]

[0136] In the steady-state assessment, the simulation data and experimental data are respectively composed into two sets of spatial vectors, where 'a' represents the simulation data and 'b' represents the experimental data. a = (low-pressure compressor speed of the gas turbine simulation model, high-pressure compressor speed of the gas turbine simulation model, power turbine speed of the gas turbine simulation model, exhaust temperature of the gas turbine simulation model, load of the gas turbine simulation model), b = (low-pressure compressor speed of the gas turbine, high-pressure compressor speed of the gas turbine, power turbine speed of the gas turbine, exhaust temperature of the gas turbine, load of the gas turbine).

[0137] In dynamic evaluation, vector a represents dynamic simulation data, x1, x2…x p This represents a one-dimensional time series of a parameter in a simulation process. Vector b represents the dynamic experimental data, y1, y2, ..., y3. pA one-dimensional time series for a parameter (the same as the simulation parameter) in a single experimental process.

[0138] Substituting the two sets of vectors into the spatial cosine similarity calculation formula, the steady-state confidence assessment results are obtained, as shown in Tables 5 and 6.

[0139] Table 5. Confidence Assessment Results of Steady-State State

[0140] Operating conditions (load) Confidence 22.17734688MW 0.997185938805103 21.35868273MW 0.997852588695416 17.01052632MW 0.998980394465831 12.85263158MW 0.999492141870793 8.621052632MW 0.999655741445505 4.336842105MW 0.999712770329707

[0141] Table 6. Results of Dynamic Confidence Assessment

[0142]

[0143]

[0144] (4) The steady-state confidence score and dynamic confidence score are weighted based on the analytic hierarchy process (AHP) to obtain the overall confidence score of the gas turbine simulation model after weighting.

[0145] First, construct a judgment matrix (based on the time percentage of each working condition throughout the year).

[0146] The judgment matrix of the steady-state sub-criteria layer is constructed, and the resulting judgment matrix is ​​shown in Table 7.

[0147] Table 7. Judgment Matrix of Steady-State Sub-Criterion Layer

[0148]

[0149] The judgment matrix of the dynamic sub-criteria layer is constructed, and the resulting judgment matrix is ​​shown in Table 8.

[0150] Table 8 Judgment Matrix of Dynamic Sub-Criterion Layer

[0151]

[0152]

[0153] The judgment matrix of the criterion layer is constructed, and the resulting judgment matrix is ​​shown in Table 9.

[0154] Table 9 Judgment Matrix of Criterion Layer

[0155]

[0156]

[0157] Second, check the consistency of the judgment matrix.

[0158] Steady-state sub-criterion layer: λ=6.0898, n=6, C1=0.018, CR=0.0143<1, accepted.

[0159] Dynamic sub-criteria layer: λ=11.5262, n=11, C1=0.0526, CR=0.0346<1, Accept.

[0160] Criterion layer: a second-order matrix, which is a positive reciprocal matrix, and does not require consistency testing.

[0161] Third, determine the weight of each indicator based on the judgment matrix.

[0162] The weights of operating conditions 1-6 in the steady-state sub-criteria layer are shown in Table 10.

[0163] Table 10 Weights of the Steady-State Sub-Criterion Layer

[0164] Operating conditions 1 2 3 4 5 6 Weight 0.0525 0.0525 0.0525 0.2783 0.4139 0.1503

[0165] The weights of operating conditions 1-11 in the dynamic sub-criteria layer are shown in Table 11.

[0166] Table 11 Weights of the Steady-State Sub-Criterion Layer

[0167]

[0168] The weights of the criteria layer are shown in Table 12.

[0169] Table 12 Weight Table of Criterion Layer

[0170] steady state dynamic Weight 0.8333 0.1667

[0171] Overall weight calculation:

[0172] The overall weights of the steady-state sub-criteria layer are shown in Table 13.

[0173] Table 13 Comprehensive Weight Table of Steady-State Sub-Criterion Layer

[0174] Operating conditions 1 2 3 4 5 6 Weight 0.0437 0.0437 0.0437 0.2319 0.3449 0.1253

[0175] The comprehensive weights of the dynamic sub-criteria layer are shown in Table 14.

[0176] Table 14 Comprehensive Weight Table of Dynamic Sub-Criterion Layer

[0177]

[0178] Fourth, calculate the overall confidence level based on the evaluation model.

[0179] The calculated overall confidence level is 0.9970. Therefore, the confidence level of the simulation model of a certain type of three-axis gas turbine is 0.9970.

[0180] Example 2

[0181] The confidence determination system for the gas turbine simulation model in this embodiment includes:

[0182] The data acquisition module is used to acquire test data of the gas turbine and simulation data of the corresponding gas turbine simulation model. The test data includes test data under multiple steady-state conditions and test data under multiple dynamic conditions, and the simulation data includes simulation data under multiple steady-state conditions and simulation data under multiple dynamic conditions. One test data corresponds to one simulation data.

[0183] The underlying confidence calculation module is used to calculate the confidence between each experimental data and the corresponding simulation data using the spatial cosine similarity evaluation algorithm, thereby obtaining multiple underlying confidence scores.

[0184] The confidence level determination module is used to assign weights to the confidence levels of each bottom layer based on the analytic hierarchy process (AHP), and to determine the confidence level of the gas turbine simulation model based on all weighted bottom layer confidence levels.

[0185] Example 3

[0186] An electronic device includes a memory and a processor. The memory stores a computer program, and the processor runs the computer program to enable the electronic device to perform the above-described method for determining the confidence level of a gas turbine simulation model.

[0187] As an optional implementation, the memory is a readable storage medium.

[0188] The advantages of this invention are:

[0189] (1) The Analytic Hierarchy Process (AHP) is good at handling complex decision-making problems: The comprehensive evaluation of gas turbine simulation models needs to consider multiple factors, and the AHP can hierarchically and structurally process these factors, which helps to reduce the complexity of the problem.

[0190] (2) It can quantify the relationship between different factors: In the analytic hierarchy process, by constructing a hierarchical structure and comparison matrix, the relative importance and influence relationship between different factors can be quantified, which is conducive to more objectively weighing and comparing each factor.

[0191] (3) It has good visualization effect: The analytic hierarchy process can intuitively show the relative importance and influence of each factor by drawing a hierarchical structure diagram and calculating weight vectors, making the decision results easier to understand and accept.

[0192] (4) High reliability: The method based on the comparative consistency index can correct the decision-maker's subjective bias and improve the stability and reliability of the decision results.

[0193] (5) Wide range of applications: The analytic hierarchy process is not only applicable to single-criteria decision problems, but also to multi-criteria decision problems and situations with high uncertainty.

[0194] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0195] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for determining the confidence level of a gas turbine simulation model, characterized in that, The method includes: Acquire test data of the gas turbine and simulation data of the corresponding gas turbine simulation model; the test data includes test data under multiple steady-state operating conditions and test data under multiple dynamic operating conditions, and the simulation data includes simulation data under multiple steady-state operating conditions and simulation data under multiple dynamic operating conditions; one set of test data corresponds to one set of simulation data; The confidence scores between each experimental data and the corresponding simulation data are calculated using the spatial cosine similarity evaluation algorithm, thereby obtaining multiple underlying confidence scores. The confidence scores of each bottom layer are weighted using the analytic hierarchy process (AHP), and the confidence score of the gas turbine simulation model is determined based on all weighted bottom layer confidence scores. Obtain test data of the gas turbine and simulation data of the corresponding gas turbine simulation model, including: During the steady-state process, the parameters of the gas turbine under each steady-state condition are obtained, resulting in test data under multiple steady-state conditions. The parameters under the steady-state conditions include: high-pressure compressor speed, low-pressure compressor speed, power turbine speed, exhaust temperature, and load. During the steady-state process, the indicators of the gas turbine simulation model under each steady-state condition are obtained, and simulation data under multiple steady-state conditions are obtained. During the dynamic process, the indicators of the gas turbine under each dynamic operating condition are acquired, and test data under multiple operating conditions are obtained. The indicators under the dynamic operating conditions include: the power turbine speed change diagram during the load reduction process, the power turbine speed change diagram during the loading process, and the power turbine speed change diagram during the load shedding process. During the dynamic process, the indicators of the gas turbine simulation model under each dynamic operating condition are obtained, and simulation data under multiple dynamic conditions are obtained. The confidence levels of each bottom layer are weighted using the analytic hierarchy process (AHP), and the confidence level of the gas turbine simulation model is determined based on all weighted bottom layer confidence levels. Specifically, this includes: Based on the time proportion of each steady-state condition in a preset time period, a judgment matrix for the steady-state sub-criteria layer is constructed. Based on the time proportion of each dynamic working condition in a preset time period, a judgment matrix for the dynamic sub-criteria layer is constructed. Construct the judgment matrix of the criterion layer; Consistency checks are performed on the judgment matrices of the steady-state sub-criteria layer and the dynamic sub-criteria layer respectively, and the judgment matrices that fail the consistency check are re-established. The weights of each indicator are determined based on the judgment matrix of the criterion layer, the judgment matrix of the steady-state sub-criterion layer that has passed the consistency test, and the judgment matrix of the dynamic sub-criterion layer that has passed the consistency test. The confidence level of the gas turbine simulation model is calculated based on the weights of all indicators and their corresponding underlying confidence levels.

2. The method for determining the confidence level of a gas turbine simulation model according to claim 1, characterized in that, After obtaining the test data of the gas turbine and the simulation data of the corresponding gas turbine simulation model, the following is also included: The experimental data and the simulation data are sequentially subjected to outlier removal, interval normalization, noise filtering, and sensor correction.

3. The method for determining the confidence level of a gas turbine simulation model according to claim 2, characterized in that, The experimental data and the simulation data are sequentially subjected to outlier removal, interval normalization, noise filtering, and sensor correction, specifically including: The 3σ criterion was used to remove outliers from the experimental data and the simulation data, respectively, to obtain the outlier-removed experimental data and the outlier-removed simulation data. The maximum-minimum normalization method was used to perform interval normalization on the experimental data and simulation data after outlier removal, respectively, to obtain the normalized experimental data and normalized simulation data. Kalman filtering was used to filter noise from the normalized experimental data and the normalized simulation data, respectively, to obtain the filtered experimental data and the filtered simulation data. The median method was used to perform sensor correction on the filtered experimental data and the filtered simulation data, respectively, to obtain the sensor-corrected experimental data and the sensor-corrected simulation data.

4. The method for determining the confidence level of a gas turbine simulation model according to claim 1, characterized in that, Based on the weights of all indicators and their corresponding underlying confidence levels, the confidence level of the gas turbine simulation model is calculated, specifically including: Multiply the weight of each indicator by its corresponding underlying confidence level to obtain multiple weighted underlying confidence levels; The confidence level of the gas turbine simulation model is obtained by summing all the weighted underlying confidence levels.

5. A confidence determination system for a gas turbine simulation model, characterized in that, The system includes: The data acquisition module is used to acquire test data of the gas turbine and simulation data of the corresponding gas turbine simulation model; the test data includes test data under multiple steady-state operating conditions and test data under multiple dynamic operating conditions, and the simulation data includes simulation data under multiple steady-state operating conditions and simulation data under multiple dynamic operating conditions; one set of test data corresponds to one set of simulation data; The underlying confidence calculation module is used to calculate the confidence between each experimental data and the corresponding simulation data using the spatial cosine similarity evaluation algorithm, thereby obtaining multiple underlying confidence levels. The confidence level determination module is used to assign weights to each of the underlying confidence levels based on the analytic hierarchy process (AHP), and to determine the confidence level of the gas turbine simulation model based on all the weighted underlying confidence levels. Obtain test data of the gas turbine and simulation data of the corresponding gas turbine simulation model, including: During the steady-state process, the parameters of the gas turbine under each steady-state condition are obtained, resulting in test data under multiple steady-state conditions. The parameters under the steady-state conditions include: high-pressure compressor speed, low-pressure compressor speed, power turbine speed, exhaust temperature, and load. During the steady-state process, the indicators of the gas turbine simulation model under each steady-state condition are obtained, and simulation data under multiple steady-state conditions are obtained. During the dynamic process, the indicators of the gas turbine under each dynamic operating condition are acquired, and test data under multiple operating conditions are obtained. The indicators under the dynamic operating conditions include: the power turbine speed change diagram during the load reduction process, the power turbine speed change diagram during the loading process, and the power turbine speed change diagram during the load shedding process. During the dynamic process, the indicators of the gas turbine simulation model under each dynamic operating condition are obtained, and simulation data under multiple dynamic conditions are obtained. The confidence levels of each bottom layer are weighted using the analytic hierarchy process (AHP), and the confidence level of the gas turbine simulation model is determined based on all weighted bottom layer confidence levels. Specifically, this includes: Based on the time proportion of each steady-state condition in a preset time period, a judgment matrix for the steady-state sub-criteria layer is constructed. Based on the time proportion of each dynamic working condition in a preset time period, a judgment matrix for the dynamic sub-criteria layer is constructed. Construct the judgment matrix of the criterion layer; Consistency checks are performed on the judgment matrices of the steady-state sub-criteria layer and the dynamic sub-criteria layer respectively, and the judgment matrices that fail the consistency check are re-established. The weights of each indicator are determined based on the judgment matrix of the criterion layer, the judgment matrix of the steady-state sub-criterion layer that has passed the consistency test, and the judgment matrix of the dynamic sub-criterion layer that has passed the consistency test. The confidence level of the gas turbine simulation model is calculated based on the weights of all indicators and their corresponding underlying confidence levels.

6. An electronic device, characterized in that, The device includes a memory and a processor, the memory being used to store a computer program, and the processor running the computer program to cause the electronic device to perform the confidence determination method for a gas turbine simulation model according to any one of claims 1 to 4.

7. An electronic device according to claim 6, characterized in that, The memory is a readable storage medium.

Images