Gas turbine compressor variable guide vane segmented modeling method and system

Abstract

The application relates to the technical field of gas turbine control, and discloses a gas turbine compressor adjustable guide vane segmented modeling method and system. The method comprises the following steps: dividing the compressor into initial calculation sections according to cooling air extraction positions, establishing a mechanism model and calculating a cumulative error contribution rate; identifying mismatched sections and splitting the mismatched sections into front and rear sub-sections, and re-distributing pressure ratio coefficients; setting a virtual buffer layer, calculating a dynamic weighting coefficient according to the cumulative error contribution rate, and obtaining buffer layer output parameters through weighted average of boundary parameters of the front and rear sub-sections; updating cooling air extraction flow and mass flow, and globally optimizing pressure ratio distribution coefficients to obtain a target segmented model. The application improves the prediction accuracy and numerical stability of the gas turbine compressor adjustable guide vane segmented modeling under variable working conditions.

Description

Technical Field

[0001] This application relates to the field of gas turbine control technology, and in particular to a method and system for segmented modeling of adjustable guide vanes in a gas turbine compressor. Background Technology

[0002] The segmented modeling method for adjustable guide vanes in gas turbine compressors is a core foundation for the research and design of gas turbine control systems. Existing technologies typically employ a modular modeling approach, dividing the compressor into several computational segments based on its physical structure. Each segment corresponds to a certain number of compression stages, and the air compression process is described by establishing thermodynamic mechanism models for each segment. The adjustable guide vanes, as key regulating components at the compressor inlet, affect the airflow and compression characteristics upon entry. Existing segmented modeling methods usually set fixed IGV (Inlet Gas Vault) opening correction coefficients to reflect this effect. Simultaneously, the segment boundaries are determined based on the location of cooling extraction within the compressor. Mathematical models for each segment are established using the laws of conservation of mass and energy. The dynamic characteristics of the entire compressor are calculated by cascading the segment models and considering the distribution ratio of cooling extraction. While this modeling method has a certain level of accuracy under design or steady-state conditions, it still has significant shortcomings in practical applications.

[0003] The main problem with existing technologies is that the segmentation scheme is fixed at the initial stage of modeling, making it impossible to dynamically adjust based on accuracy issues revealed during actual operation. When there is a significant deviation between the mechanism model of a certain calculation segment and the actual physical process, the accuracy can only be improved by adjusting the correction parameters of that segment. However, such global parameter adjustments often lead to a decrease in the prediction accuracy of other segments, resulting in a "whack-a-mole" phenomenon. In addition, the IGV opening correction coefficient and cooling extraction flow rate in existing methods usually use fixed proportions or simple linear corrections, which cannot accurately reflect the attenuation law of IGV influence under different operating conditions and the dynamic matching relationship between turbine cooling demand and compressor gas supply capacity. This leads to a significant decrease in the prediction accuracy of the model under varying operating conditions, making it difficult to meet the modeling accuracy requirements for gas turbine control system development and performance testing. Summary of the Invention

[0004] This application provides a method and system for segmented modeling of adjustable guide vanes in a gas turbine compressor. It is used to achieve accurate positioning of mismatched segments by establishing an error propagation tracing matrix, solve the boundary discontinuity problem caused by segment refinement by setting a dynamically weighted virtual buffer layer, and achieve the unity of local accuracy improvement and global constraint satisfaction by constructing a multi-objective optimization function. It solves the technical problems of existing fixed segmentation schemes being unable to adaptively adjust and local optimization destroying global consistency, and improves the prediction accuracy and numerical stability of segmented modeling of adjustable guide vanes in gas turbine compressors under variable operating conditions.

[0005] In a first aspect, this application provides a method for segmented modeling of adjustable guide vanes in a gas turbine compressor, the method comprising: Step S1: Divide the compressor into initial calculation segments according to the cooling and extraction positions, establish the mechanism model of each initial calculation segment and calculate the cumulative error contribution rate; Step S2: Identify the mismatch segment with the largest cumulative error contribution rate, split the mismatch segment into a front sub-segment and a rear sub-segment, and redistribute the pressure ratio coefficient; Step S3: Set up a virtual buffer layer between the front sub-segment and the rear sub-segment. Calculate a dynamic weighting coefficient based on the cumulative error contribution rate of the front sub-segment and the rear sub-segment. Use the dynamic weighting coefficient to perform a weighted average of the outlet temperature of the front sub-segment and the expected inlet temperature of the rear sub-segment to obtain the buffer layer output temperature. Use the dynamic weighting coefficient to perform a weighted average of the outlet pressure of the front sub-segment and the expected inlet pressure of the rear sub-segment to obtain the buffer layer output pressure. Use the buffer layer output temperature and the buffer layer output pressure as the actual input parameters of the rear sub-segment. Step S4: Update the cooling air extraction flow rate and mass flow rate of each initial calculation segment, and perform global optimization adjustment of the pressure ratio distribution coefficient to obtain the target segmented model.

[0006] Secondly, this application provides a segmented modeling system for adjustable guide vanes of a gas turbine compressor, the system comprising: The partitioning module is used to divide the compressor into initial calculation segments according to the cooling and extraction positions, establish the mechanism model of each initial calculation segment, and calculate the cumulative error contribution rate. The splitting module is used to identify the mismatch segment with the largest cumulative error contribution rate, split the mismatch segment into a front sub-segment and a rear sub-segment, and redistribute the pressure ratio coefficient. The calculation module is used to set a virtual buffer layer between the front sub-segment and the rear sub-segment, calculate a dynamic weighting coefficient based on the cumulative error contribution rate of the front sub-segment and the rear sub-segment respectively, use the dynamic weighting coefficient to perform a weighted average of the outlet temperature of the front sub-segment and the expected inlet temperature of the rear sub-segment to obtain the buffer layer output temperature, use the dynamic weighting coefficient to perform a weighted average of the outlet pressure of the front sub-segment and the expected inlet pressure of the rear sub-segment to obtain the buffer layer output pressure, and use the buffer layer output temperature and the buffer layer output pressure as the actual input parameters of the rear sub-segment. The update module is used to update the cooling air extraction flow rate and mass flow rate of each initial calculation segment, and to perform global optimization and adjustment of the pressure ratio distribution coefficient to obtain the target segmented model.

[0007] The technical solution provided in this application divides the compressor into initial calculation segments according to the cooling extraction position and establishes a mechanistic model for each initial calculation segment. This ensures that the air mass flow rate within each calculation segment remains continuous, avoiding the modeling difficulties caused by sudden changes in mass flow rate when the cooling extraction position is placed within a segment. Simultaneously, by calculating the cumulative error contribution rate, a quantitative assessment of the prediction accuracy of each initial calculation segment is achieved, providing a clear data foundation for subsequent adaptive adjustments. By identifying the mismatch segment with the largest cumulative error contribution rate and splitting it into a pre-segment and a post-segment, this application achieves precise optimization of the bottleneck link in modeling accuracy, avoiding the mutual interference problem caused by global parameter adjustments in existing technologies. Furthermore, the splitting point is determined based on the cooling extraction position or the attenuation boundary of the adjustable guide vane influence, giving the split segments clear physical meaning. The pressure ratio coefficient is redistributed based on the adiabatic compression work distribution principle, ensuring the thermodynamic rationality of the refined model and overcoming the shortcomings of existing fixed-segment schemes that cannot dynamically adjust based on accuracy issues exposed during actual operation. Attached Figure Description

[0008] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a schematic diagram of an embodiment of the segmented modeling method for adjustable guide vanes of a gas turbine compressor in this application. Figure 2 This is a schematic diagram of an embodiment of the gas turbine compressor adjustable guide vane segment modeling system in this application. Figure 3 This is a schematic block diagram of the structure of the adjustable guide vane segment modeling device for a gas turbine compressor in an embodiment of the present invention. Detailed Implementation

[0010] This application provides a method and system for segmented modeling of adjustable guide vanes in a gas turbine compressor. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0011] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the segmented modeling method for adjustable guide vanes of a gas turbine compressor in this application includes: Step S1: Divide the compressor into initial calculation segments according to the cooling and extraction positions, establish the mechanism model of each initial calculation segment and calculate the cumulative error contribution rate; Specifically, the initial calculation segments of the compressor are divided according to the cooling extraction location because cooling extraction causes abrupt changes in air mass flow rate before and after the segment. Placing the extraction location within the segment would increase modeling complexity. The mechanistic models of each initial calculation segment are established based on the first law of thermodynamics and the law of conservation of mass. The pressure ratio and efficiency relationship under different speeds and guide vane openings are determined through compressor characteristic curves. The calculation of the cumulative error contribution rate needs to consider the prediction deviations of both temperature and pressure dimensions. The relative errors of each segment are normalized and then summed to make the errors of different physical quantities comparable, thereby accurately locating the segment with the worst modeling accuracy.

[0012] Step S2: Identify the mismatch segment with the largest cumulative error contribution rate, split the mismatch segment into a front sub-segment and a rear sub-segment, and redistribute the pressure ratio coefficient; Specifically, after identifying the mismatched segment, it is necessary to determine whether the number of compression stages it contains meets the splitting conditions, because too few compression stages cannot be further refined. The determination of the splitting point follows the principle of prioritizing physical meaning. The cooling extraction position is a natural dividing point, while the adjustable guide vane influence attenuation boundary is calculated through the influence attenuation coefficient. This boundary divides the compressor into a strong IGV influence zone and a weak IGV influence zone. The recalculation of the pressure ratio distribution coefficient is based on the adiabatic compression work distribution principle, taking into account the increase in single-stage compression work caused by the increased air density in the later stages of the compressor, and is corrected through the compression work distribution index.

[0013] Step S3: Set up a virtual buffer layer between the front sub-segment and the back sub-segment. Calculate the dynamic weighting coefficient based on the cumulative error contribution rate of the front and back sub-segments. Use the dynamic weighting coefficient to perform a weighted average of the outlet temperature of the front sub-segment and the expected inlet temperature of the back sub-segment to obtain the buffer layer output temperature. Use the dynamic weighting coefficient to perform a weighted average of the outlet pressure of the front sub-segment and the expected inlet pressure of the back sub-segment to obtain the buffer layer output pressure. Use the buffer layer output temperature and buffer layer output pressure as the actual input parameters of the back sub-segment. Specifically, the dynamic weighting coefficients are calculated based on the cumulative error contribution rate of each segment. Segments with a smaller cumulative error contribution rate indicate higher prediction accuracy and should receive a larger weight in boundary calculations. This dynamic weighting allocation mechanism avoids unreasonable boundary parameters caused by using fixed weights. The weighted averaging process is performed independently for both temperature and pressure parameters. The resulting buffer layer output parameters must satisfy the thermodynamic monotonicity constraint, i.e., pressure and temperature increase monotonically along the compressor flow direction.

[0014] Step S4: Update the cooling air extraction flow rate and mass flow rate of each initial calculation segment, and perform global optimization adjustment of the pressure ratio distribution coefficient to obtain the target segmented model.

[0015] Specifically, the mass flow rate is updated segment by segment, starting from the compressor inlet. The outlet flow rate of each segment equals the inlet flow rate minus the cooling pump flow rate of that segment. Global optimization is performed using a multi-objective optimization function, which includes three sub-objectives: local accuracy, energy conservation, and mass conservation. By setting different weight coefficients to balance the importance of each objective, the gradient descent method is used to coordinately fine-tune the pressure ratio allocation coefficients of all segments, so that the improvement of local accuracy and the satisfaction of global constraints are unified.

[0016] In one specific embodiment, step S1 includes: Obtain the design parameters of the compressor and divide the compressor into initial calculation sections according to the internal cooling and extraction locations; Mechanism models are established for each initial calculation segment. The input parameters of the mechanism models include inlet air temperature, inlet pressure, adjustable guide vane opening correction coefficient, and unit speed. The output parameters include outlet temperature, outlet pressure, compressor power consumption, and cooling air mass flow rate. The actual operating measurement data of the compressor is collected, and the predicted outlet temperature and outlet pressure are obtained by forward calculation using the mechanism model. The actual outlet temperature and actual outlet pressure of each initial calculation segment are inferred by back-calculation based on the measured parameters using the interpolation method. The predicted outlet temperature and pressure are compared with the actual outlet temperature and pressure obtained by reverse calculation to obtain the temperature prediction error and pressure prediction error. The cumulative error contribution rate of each initial calculation segment is calculated based on the temperature prediction error and pressure prediction error.

[0017] Specifically, the compressor design parameters include the total number of stages, the total pressure ratio under design conditions, the design speed, and the design opening of the adjustable guide vanes. These parameters determine the basic structure of the initial segmentation scheme. The reason for dividing the initial calculation segments based on the cooling extraction location is that cooling extraction causes a step change in air mass flow rate before and after that location. If the extraction point were placed within a segment, a mathematical description of the abrupt change in mass flow rate would need to be established within that segment, increasing modeling complexity and computational instability. Therefore, using the extraction location as the segment boundary ensures that the mass flow rate within each segment remains continuous. The calculation of the adjustable guide vane opening correction coefficient considers the physical characteristic that the influence of the IGV on subsequent compression stages decreases progressively. It is determined by the influence attenuation coefficient and the position sequence of the segment in the compressor. The earlier the segment, the more significant the impact of the IGV opening change; the later the segment, the weaker the impact.

[0018] The interpolation method for inferring the actual outlet parameters of each initial calculation segment is based on the physical characteristic that the temperature and pressure of the compressor increase monotonically along the flow direction. Using measured values ​​at the compressor inlet and outlet, and the pressure ratio distribution of each segment under design conditions, a parameter distribution curve from inlet to outlet is constructed through polynomial interpolation. The temperature and pressure at the boundaries of each segment are extracted from this curve as estimates of the actual values. Calculating the cumulative error contribution rate requires normalizing the errors of temperature and pressure, which have two different dimensions. Specifically, the relative temperature and pressure errors of each segment are added together and then divided by the sum of the relative errors of all segments, ensuring that the sum of the error contribution rates of each segment equals 1. This normalization process allows for direct comparison of the relative magnitudes of the errors in each segment, identifying the mismatch segments that have the greatest impact on the overall prediction accuracy, and providing a clear target for subsequent dynamic refinement operations.

[0019] In one specific embodiment, step S2 includes: Identify the mismatch segment corresponding to the maximum value in the cumulative error contribution rate; Determine whether the number of compression levels contained in the mismatch segment is greater than the preset minimum number of segments. If it is greater, then perform the segment splitting operation. The split point is determined based on whether there is a cooling evacuation position within the mismatched segment. If there is a cooling evacuation position, the split point is taken as the cooling evacuation position. If there is no cooling evacuation position, the split point is determined based on the attenuation boundary of the adjustable guide vane, and the mismatched segment is split into a front sub-segment and a rear sub-segment. Based on the principle of adiabatic compression work distribution, the pressure ratio distribution coefficients of the front and rear sections are recalculated, and the adjustable guide vane opening correction coefficients of the front and rear sections are updated.

[0020] Specifically, the mismatched segment with the largest cumulative error contribution rate indicates that the mechanistic model prediction accuracy of that segment is the worst, becoming a bottleneck restricting the overall modeling accuracy. Therefore, structural optimization of this segment should be prioritized. The purpose of determining whether the number of compression stages is greater than the minimum number of stages is to avoid over-refinement leading to too few compression stages in a single segment. Too few compression stages will result in too small a range of thermodynamic parameter changes in that segment, making it difficult to accurately describe its dynamic characteristics through the mechanistic model. At the same time, over-refinement increases the total number of segments and computational complexity. The minimum number of stages is usually set to 4 to 6, ensuring that the refined sub-segments have sufficient physical meaning while controlling the reasonable scale of the segmentation scheme.

[0021] The determination of the split point follows the principle of prioritizing physical meaning. The cooling extraction location is where the mass flow rate of air inside the compressor changes abruptly. Using this as the split point gives the boundary conditions of the preceding and following segments a clear physical meaning. The outlet flow rate of the preceding segment needs to be reduced by the cooling extraction flow rate to obtain the inlet flow rate of the following segment. If there is no cooling extraction location within the mismatched segment, the split point is determined based on the adjustable guide vane influence attenuation boundary. This boundary is calculated using the influence attenuation coefficient and represents the boundary where the influence of IGV opening changes on the compression stage transitions from a significant region to a weak influence region. Splitting the segment at this boundary allows the preceding segment to primarily reflect the compression characteristics under strong IGV influence, while the following segment primarily reflects the compression characteristics under weak IGV influence. The physical mechanisms of the two segments are significantly different, and modeling them separately can improve local accuracy. The adiabatic compression work distribution principle considers the increase in single-stage compression work caused by the increased air density in the later stages of the compressor. It uses a compression work distribution index to weight the pressure ratio distribution coefficients of the preceding and following stages. This index typically ranges from 1.1 to 1.3, reflecting the physical law that the pressure ratio contribution of the later stages is greater than that of the earlier stages in a real compressor. The update of the adjustable guide vane opening correction coefficient needs to be recalculated based on the relative positions of the preceding and following stages within the overall compressor. The correction coefficient for the preceding stage retains the original mismatched segment correction coefficient, while the correction coefficient for the following stage is determined based on the degree of IGV influence attenuation corresponding to its initial position, ensuring that the refined stage model still conforms to the physical characteristic of the gradual attenuation of IGV influence.

[0022] In one specific embodiment, step S3, which involves setting a virtual buffer layer between the preceding and following sub-segments, includes: The outlet temperature and outlet pressure of the preceding segment are obtained as the input parameters of the preceding segment of the virtual buffer layer; The expected inlet temperature and expected inlet pressure of the subsequent segment are obtained by reverse calculation based on the mechanism model of the subsequent segment, and the expected inlet temperature and expected inlet pressure are used as the input parameters of the subsequent segment of the virtual buffer layer. The dynamic weighting coefficient is calculated based on the ratio of the cumulative error contribution rate of the preceding sub-segment to the cumulative error contribution rate of the following sub-segment. The output temperature of the buffer layer is obtained by weighting the outlet temperature of the front segment and the expected inlet temperature of the rear segment using dynamic weighting coefficients. The output pressure of the buffer layer is obtained by weighting the outlet pressure of the front segment and the expected inlet pressure of the rear segment using dynamic weighting coefficients.

[0023] Specifically, the outlet temperature and pressure of the first segment are calculated forward from the mechanistic model of that segment, representing the predicted boundary state of the first segment. However, due to modeling errors, there is a deviation between this predicted value and the actual inlet conditions required by the subsequent segment. The expected inlet temperature and pressure of the subsequent segment are obtained through reverse calculation. Specifically, given the outlet state that the subsequent segment needs to achieve, the inlet state is inferred using the pressure ratio and adiabatic efficiency of the subsequent segment, through the inverse process formula of adiabatic compression. This inferred value represents the ideal conditions that the inlet should possess to ensure that the output of the subsequent segment meets the global pressure ratio constraint. The difference between these two sets of parameters reflects the boundary discontinuity problem caused by segmentation refinement.

[0024] The dynamic weighting coefficients are calculated based on the cumulative error contribution rates of the preceding and following sub-segments. The weight of the preceding sub-segment is obtained by dividing its cumulative error contribution rate by the sum of the cumulative error contribution rates of the preceding and following sub-segments. The weight of the following sub-segment is 1 minus the weight of the preceding sub-segment. The physical meaning of this calculation method is that the smaller the cumulative error contribution rate, the higher the prediction accuracy of the sub-segment, and the closer its predicted boundary parameters are to the true values. Therefore, it should be given a larger weight during boundary fusion. Conversely, the larger the cumulative error contribution rate, the lower the prediction reliability of the sub-segment, and its influence on the determination of boundary parameters should be reduced. The weighted average processing is performed independently for the temperature and pressure parameters. The output temperature of the buffer layer is equal to the outlet temperature of the previous sub-segment multiplied by the dynamic weighting coefficient, plus the expected inlet temperature of the subsequent sub-segment multiplied by 1 minus the dynamic weighting coefficient. The output pressure of the buffer layer is calculated in the same way. Through this dynamic weighting mechanism, the sub-segment with high prediction accuracy automatically gains greater weight in the boundary calculation, while the influence of the sub-segment with low prediction accuracy is weakened accordingly. This achieves a reasonable compromise when there is a conflict between the prediction values ​​of the previous and subsequent sub-segments. The resulting output parameters of the buffer layer take into account both the calculation results of the previous sub-segment and the actual needs of the subsequent sub-segment, avoiding parameter abrupt changes and calculation divergence caused by directly using the output of the previous sub-segment or directly using the expected input of the subsequent sub-segment.

[0025] In one specific embodiment, the dynamic weighting coefficient is calculated based on the ratio of the cumulative error contribution rate of the preceding sub-segment to the cumulative error contribution rate of the following sub-segment, including: The ratio of the cumulative error contribution rate of the preceding sub-segment to the sum of the cumulative error contribution rates of the preceding and following sub-segments is used as a dynamic weighting coefficient, so that sub-segments with smaller cumulative error contribution rates receive greater weight in boundary calculations.

[0026] Specifically, the dynamic weighting coefficient is calculated by dividing the cumulative error contribution rate of the preceding sub-segment by the sum of the cumulative error contribution rates of the preceding and following sub-segments. This ratio ranges from 0 to 1. When the cumulative error contribution rate of the preceding sub-segment is much smaller than that of the following sub-segment, the ratio is close to 0, indicating that the prediction accuracy of the preceding sub-segment is high while that of the following sub-segment is low. In this case, the dynamic weighting coefficient is small, causing the output parameters of the buffer layer to be closer to the expected input value of the following sub-segment. This is equivalent to making a significant correction to the prediction result of the preceding sub-segment to match the requirements of the following sub-segment. Conversely, when the cumulative error contribution rate of the preceding sub-segment is much larger than that of the following sub-segment, the dynamic weighting coefficient is close to 1, and the output parameters of the buffer layer are closer to the output value of the preceding sub-segment. This is equivalent to adjusting the expected input of the following sub-segment to adapt to the actual output of the preceding sub-segment. This mechanism of dynamically allocating weights based on prediction accuracy ensures that the sub-segment with higher prediction reliability always dominates the boundary parameter fusion process, avoiding excessive influence of low-precision prediction on the boundary state, thereby achieving a physically reasonable smooth transition when there is parameter mismatch between the preceding and following sub-segments.

[0027] In one specific embodiment, step S3 further includes: Perform physical constraint checks on the output pressure and output temperature of the buffer layer; If the output pressure of the buffer layer is less than the outlet pressure of the preceding segment or the output temperature of the buffer layer is less than the outlet temperature of the preceding segment, the output pressure of the buffer layer is corrected to the product of the outlet pressure of the preceding segment and the minimum pressure ratio increment coefficient, and the output temperature of the buffer layer is corrected to the product of the outlet temperature of the preceding segment and the minimum pressure ratio increment coefficient, to ensure the monotonically increasing characteristics of pressure and temperature.

[0028] Specifically, the purpose of physical constraint checks is to ensure that the output parameters of the buffer layer conform to the fundamental thermodynamic law that the pressure and temperature of the compressor increase monotonically along the flow direction. When the output pressure of the buffer layer obtained by dynamic weighted averaging is lower than the outlet pressure of the preceding section or the output temperature of the buffer layer is lower than the outlet temperature of the preceding section, it indicates that the weighting result violates the second law of thermodynamics. In this case, forced correction is required. The correction method is to set the output pressure of the buffer layer to the outlet pressure of the preceding section multiplied by the minimum pressure ratio increment coefficient, and the output temperature of the buffer layer to the outlet temperature of the preceding section multiplied by the minimum pressure ratio increment coefficient. This coefficient is usually taken as 1.001, representing the minimum pressure ratio increment of a single calculation section of the compressor. This correction mechanism ensures the strict monotonicity of pressure and temperature, minimizes the deviation between the corrected parameters and the original values, avoids excessive artificial disturbances introduced by constraint correction, and ensures the physical feasibility and numerical stability of the buffer layer output.

[0029] In one specific embodiment, step S4 includes: Obtain the real-time measurement value of the gas turbine inlet temperature, and calculate the cooling extraction correction coefficient of each initial calculation section based on the turbine inlet temperature, the outlet temperature of each initial calculation section, and the pressure ratio. The cooling extraction mass flow rate of each initial calculation segment is updated based on the cooling extraction correction coefficient, and the inlet mass flow rate and outlet mass flow rate of each initial calculation segment are recursively calculated using the mass flow balance equation. A multi-objective optimization function is constructed, which includes a local accuracy objective, a global energy conservation objective, and a global mass conservation objective. The local accuracy objective is the weighted sum of the prediction errors of each initial calculation segment, the global energy conservation objective is the prediction error of the total power consumption of the compressor, and the global mass conservation objective is the prediction error of the outlet mass flow rate. The gradient descent method is used to iteratively adjust the pressure ratio allocation coefficient of each initial calculation segment until the multi-objective optimization function converges, thus obtaining the target piecewise model.

[0030] Specifically, the turbine inlet temperature directly determines the cooling requirements of the turbine blades. The higher the temperature, the more cooling air flow is needed to maintain the blade surface temperature within the allowable range of the material. The calculation of the cooling extraction correction factor takes into account the deviation of the turbine inlet temperature from the design value and the quality differences of the cooling air in each initial calculation section. The quality is quantified by the product of the outlet temperature and pressure ratio of the section. The cooling air extracted by the section with high outlet temperature and high pressure ratio has a higher temperature and pressure, resulting in a better cooling effect. Therefore, the extraction volume of the section can be reduced accordingly under the same cooling requirements. The corrected cooling extraction mass flow rate is equal to the inlet mass flow rate of the section multiplied by the cooling extraction correction factor. The recursive calculation of the mass flow balance equation starts from the compressor inlet. The inlet mass flow rate of the first segment is calculated based on the inlet air density, inlet cross-sectional area, and axial velocity. The outlet mass flow rate of this segment is equal to the inlet mass flow rate minus the cooling extraction mass flow rate of this segment. This outlet flow rate is used as the inlet flow rate of the next segment and the calculation continues segment by segment until the compressor outlet. After the recursion is completed, the calculated outlet mass flow rate is compared with the turbine inlet flow rate requirement. If the deviation exceeds the set threshold, the inlet mass flow rate of the first segment is iteratively adjusted until the global mass balance converges.

[0031] The construction of the multi-objective optimization function linearly combines three sub-objectives—local accuracy, energy conservation, and mass conservation—through weighted coefficients. The local accuracy objective calculates the sum of squares of the temperature and pressure prediction errors of each initial calculation segment and multiplies it by the segment weighting coefficient. For refined segments, the weighting coefficient is set to 1.5 times that of unrefined segments to strengthen the accuracy requirements of the problem region. The global energy conservation objective is the square of the difference between the sum of the compressor power consumption of each segment and the theoretical total power consumption calculated by the energy equation based on the inlet and outlet parameters. The global mass conservation objective is the square of the difference between the recursively calculated outlet mass flow rate and the turbine inlet flow rate requirement. The weighting coefficients of the three sub-objectives are set to 0.5, 0.3, and 0.2, respectively, prioritizing local accuracy while taking into account global constraints. When solving using the gradient descent method, the optimization variable is the pressure ratio allocation coefficient of each segment. The gradient is calculated using a numerical differentiation method, that is, after applying a small increment to a pressure ratio allocation coefficient, the objective function value is recalculated, and the difference between the two function values ​​divided by the increment is the gradient of that variable. The iterative update formula is the current pressure ratio allocation coefficient minus the learning rate multiplied by the gradient. The learning rate is initially set to 0.01. When the objective function does not decrease after several consecutive iterations, the learning rate is decayed to 0.5 times the original value. During the iteration process, it is necessary to ensure that all pressure ratio allocation coefficients are greater than 0 and the normalized sum is equal to the initial value. The iteration is terminated when the objective function is less than the convergence threshold or the number of iterations exceeds the upper limit. The optimized pressure ratio allocation coefficients are used to update the pressure ratio of each segment and the mechanism model parameters to form the final segmented model after global consistency verification.

[0032] The above describes the segmented modeling method for adjustable guide vanes of a gas turbine compressor in the embodiments of this application. The following describes the segmented modeling system for adjustable guide vanes of a gas turbine compressor in the embodiments of this application. Please refer to [link to relevant documentation]. Figure 2 One embodiment of the gas turbine compressor adjustable guide vane segment modeling system in this application includes: The partitioning module is used to divide the compressor into initial calculation segments according to the cooling and extraction positions, establish the mechanism model of each initial calculation segment, and calculate the cumulative error contribution rate. The splitting module is used to identify the mismatch segment with the largest cumulative error contribution rate, split the mismatch segment into a front sub-segment and a rear sub-segment, and redistribute the pressure ratio coefficient. The calculation module is used to set a virtual buffer layer between the front sub-segment and the rear sub-segment, calculate a dynamic weighting coefficient based on the cumulative error contribution rate of the front sub-segment and the rear sub-segment respectively, use the dynamic weighting coefficient to perform a weighted average of the outlet temperature of the front sub-segment and the expected inlet temperature of the rear sub-segment to obtain the buffer layer output temperature, use the dynamic weighting coefficient to perform a weighted average of the outlet pressure of the front sub-segment and the expected inlet pressure of the rear sub-segment to obtain the buffer layer output pressure, and use the buffer layer output temperature and the buffer layer output pressure as the actual input parameters of the rear sub-segment. The update module is used to update the cooling air extraction flow rate and mass flow rate of each initial calculation segment, and to perform global optimization and adjustment of the pressure ratio distribution coefficient to obtain the target segmented model.

[0033] above Figure 2 The adjustable guide vane segment modeling system for gas turbine compressor in this embodiment of the invention is described in detail from the perspective of modular functional entities. The adjustable guide vane segment modeling equipment for gas turbine compressor in this embodiment of the invention is described in detail from the perspective of hardware processing.

[0034] Reference Figure 3 This invention also provides a segmented modeling device for adjustable guide vanes of a gas turbine compressor. This device can be a server, and its internal structure can be as follows: Figure 3 As shown, the gas turbine compressor adjustable guide vane segment modeling device includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor in this computer design provides computational and control capabilities. The memory of the gas turbine compressor adjustable guide vane segment modeling device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the gas turbine compressor adjustable guide vane segment modeling device stores the data corresponding to this embodiment. The network interface of the gas turbine compressor adjustable guide vane segment modeling device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.

[0035] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the gas turbine compressor adjustable guide vane segment modeling device to which the present invention is applied.

[0036] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the gas turbine compressor adjustable guide vane segment modeling method.

[0037] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0038] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a gas turbine compressor adjustable guide vane segment modeling device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0039] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for segmented modeling of adjustable guide vanes in a gas turbine compressor, characterized in that, The method includes: Step S1: Divide the compressor into initial calculation segments according to the cooling and extraction positions, establish the mechanism model of each initial calculation segment and calculate the cumulative error contribution rate; Step S2: Identify the mismatch segment with the largest cumulative error contribution rate, split the mismatch segment into a front sub-segment and a rear sub-segment, and redistribute the pressure ratio coefficient; Step S3: Set up a virtual buffer layer between the front sub-segment and the rear sub-segment. Calculate a dynamic weighting coefficient based on the cumulative error contribution rate of the front sub-segment and the rear sub-segment. Use the dynamic weighting coefficient to perform a weighted average of the outlet temperature of the front sub-segment and the expected inlet temperature of the rear sub-segment to obtain the buffer layer output temperature. Use the dynamic weighting coefficient to perform a weighted average of the outlet pressure of the front sub-segment and the expected inlet pressure of the rear sub-segment to obtain the buffer layer output pressure. Use the buffer layer output temperature and the buffer layer output pressure as the actual input parameters of the rear sub-segment. Step S4: Update the cooling air extraction flow rate and mass flow rate of each initial calculation segment, and perform global optimization adjustment of the pressure ratio distribution coefficient to obtain the target segmented model.

2. The method for segmented modeling of adjustable guide vanes for gas turbine compressors according to claim 1, characterized in that, Step S1 includes: Obtain the design parameters of the compressor and divide the compressor into initial calculation sections according to the internal cooling and extraction locations; A mechanistic model is established for each initial calculation segment. The input parameters of the mechanistic model include inlet air temperature, inlet pressure, adjustable guide vane opening correction coefficient and unit speed. The output parameters include outlet temperature, outlet pressure, compressor power consumption and cooling air mass flow rate. The actual operating measurement data of the compressor is collected, and the predicted outlet temperature and outlet pressure are obtained by forward calculation using the mechanism model. The actual outlet temperature and actual outlet pressure of each initial calculation segment are inferred by back-calculation based on the measured parameters using the interpolation method. The predicted outlet temperature and outlet pressure are compared with the actual outlet temperature and outlet pressure obtained by reverse calculation to obtain the temperature prediction error and pressure prediction error. The cumulative error contribution rate of each initial calculation segment is calculated based on the temperature prediction error and the pressure prediction error.

3. The method for segmented modeling of adjustable guide vanes for gas turbine compressors according to claim 1, characterized in that, Step S2 includes: Identify the mismatch segment corresponding to the maximum value in the cumulative error contribution rate; Determine whether the number of compression levels contained in the mismatch segment is greater than the preset minimum number of segments; if it is greater, then perform a segmentation operation. The splitting point is determined based on whether there is a cooling evacuation position within the mismatched segment. If there is a cooling evacuation position, the splitting point is taken as the cooling evacuation position. If there is no cooling evacuation position, the splitting point is determined based on the attenuation boundary of the adjustable guide vane. The mismatched segment is then split into a front sub-segment and a rear sub-segment. Based on the principle of adiabatic compression work distribution, the pressure ratio distribution coefficients of the front and rear sections are recalculated, and the adjustable guide vane opening correction coefficients of the front and rear sections are updated.

4. The method for segmented modeling of adjustable guide vanes for gas turbine compressors according to claim 1, characterized in that, Step S3, which involves setting a virtual buffer layer between the preceding sub-segment and the following sub-segment, includes: The outlet temperature and outlet pressure of the preceding sub-segment are obtained as the input parameters of the preceding segment of the virtual buffer layer; The desired inlet temperature and desired inlet pressure of the subsequent segment are obtained by reverse calculation based on the mechanism model of the subsequent segment, and the desired inlet temperature and desired inlet pressure are used as the input parameters of the subsequent segment of the virtual buffer layer. The dynamic weighting coefficient is calculated based on the ratio of the cumulative error contribution rate of the preceding sub-segment to the cumulative error contribution rate of the following sub-segment. The output temperature of the buffer layer is obtained by weighting the outlet temperature of the front sub-segment and the expected inlet temperature of the rear sub-segment using the dynamic weighting coefficient. The output pressure of the buffer layer is obtained by weighting the outlet pressure of the front sub-segment and the expected inlet pressure of the rear sub-segment using the dynamic weighting coefficient.

5. The method for segmented modeling of adjustable guide vanes for a gas turbine compressor according to claim 4, characterized in that, The calculation of the dynamic weighting coefficient based on the ratio of the cumulative error contribution rate of the preceding sub-segment to the cumulative error contribution rate of the following sub-segment includes: The ratio of the cumulative error contribution rate of the preceding sub-segment to the sum of the cumulative error contribution rates of the preceding sub-segment and the following sub-segment is used as a dynamic weighting coefficient, so that the sub-segment with a smaller cumulative error contribution rate receives a larger weight in the boundary calculation.

6. The method for segmented modeling of adjustable guide vanes for a gas turbine compressor according to claim 4, characterized in that, Step S3 further includes: Physical constraint checks are performed on the output pressure and output temperature of the buffer layer. If the output pressure of the buffer layer is less than the outlet pressure of the preceding segment or the output temperature of the buffer layer is less than the outlet temperature of the preceding segment, then the output pressure of the buffer layer is corrected to the product of the outlet pressure of the preceding segment and the minimum pressure ratio increment coefficient, and the output temperature of the buffer layer is corrected to the product of the outlet temperature of the preceding segment and the minimum pressure ratio increment coefficient, to ensure the monotonically increasing characteristics of pressure and temperature.

7. The method for segmented modeling of adjustable guide vanes for gas turbine compressors according to claim 1, characterized in that, Step S4 includes: The real-time measured value of the gas turbine inlet temperature is obtained, and the cooling extraction correction coefficient of each initial calculation section is calculated based on the turbine inlet temperature, the outlet temperature of each initial calculation section and the pressure ratio. The cooling extraction mass flow rate of each initial calculation segment is updated according to the cooling extraction correction coefficient, and the inlet mass flow rate and outlet mass flow rate of each initial calculation segment are recursively calculated through the mass flow rate balance equation. A multi-objective optimization function is constructed, which includes a local accuracy objective, a global energy conservation objective, and a global mass conservation objective. The local accuracy objective is the weighted sum of the prediction errors of each initial calculation segment, the global energy conservation objective is the prediction error of the total power consumption of the compressor, and the global mass conservation objective is the prediction error of the outlet mass flow rate. The gradient descent method is used to iteratively adjust the pressure ratio allocation coefficient of each initial calculation segment until the multi-objective optimization function converges, thus obtaining the target segmented model.

8. A segmented modeling system for adjustable guide vanes of a gas turbine compressor, characterized in that, For implementing the segmented modeling method for adjustable guide vanes of a gas turbine compressor as described in any one of claims 1-7, the segmented modeling system for adjustable guide vanes of a gas turbine compressor comprises: The partitioning module is used to divide the compressor into initial calculation segments according to the cooling and extraction positions, establish the mechanism model of each initial calculation segment, and calculate the cumulative error contribution rate. The splitting module is used to identify the mismatch segment with the largest cumulative error contribution rate, split the mismatch segment into a front sub-segment and a rear sub-segment, and redistribute the pressure ratio coefficient. The calculation module is used to set a virtual buffer layer between the front sub-segment and the rear sub-segment, calculate a dynamic weighting coefficient based on the cumulative error contribution rate of the front sub-segment and the rear sub-segment respectively, use the dynamic weighting coefficient to perform a weighted average of the outlet temperature of the front sub-segment and the expected inlet temperature of the rear sub-segment to obtain the buffer layer output temperature, use the dynamic weighting coefficient to perform a weighted average of the outlet pressure of the front sub-segment and the expected inlet pressure of the rear sub-segment to obtain the buffer layer output pressure, and use the buffer layer output temperature and the buffer layer output pressure as the actual input parameters of the rear sub-segment. The update module is used to update the cooling air extraction flow rate and mass flow rate of each initial calculation segment, and to perform global optimization and adjustment of the pressure ratio distribution coefficient to obtain the target segmented model.

9. A segmented modeling device for adjustable guide vanes of a gas turbine compressor, characterized in that, It includes a memory and a processor, the memory storing a computer program that can run on the processor, and the processor executing the computer program to implement the segmented modeling method for the adjustable guide vanes of the gas turbine compressor as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the segmented modeling method for the adjustable guide vanes of the gas turbine compressor as described in any one of claims 1 to 7.

Images