Optimization method and system for sealing structure of interstage containing cavity of gas compressor

Abstract

The invention relates to the technical field of simulation aided design, in particular to an optimization method and system for an interstage cavity sealing structure of a gas compressor, and the method comprises the steps: determining the modeling parameters of the cavity sealing structure; relevant modeling parameters are obtained; constructing a neural network, and training the neural network; corresponding relevant modeling parameters when the efficiency of the gas compressor is maximum and the wind resistance temperature rise of the containing cavity is minimum are obtained; and the optimized compressor interstage containing cavity sealing structure is constructed. The influence of the change of each local position in the cavity sealing structure on the leakage flow is comprehensively considered, so that the negative influence of the leakage flow on the performance of the gas compressor is improved to the greatest extent.

Description

Technical Field

[0001] This invention relates to the field of simulation-aided design technology, specifically to an optimization method and system for the sealing structure of the interstage cavity of a compressor. Background Technology

[0002] Between two relatively moving components inside an aero-engine, a certain gap always exists. Abnormal flow of gas or liquid within this gap can affect engine efficiency and normal operating time to some extent. The shrouded stator vane, with its good mechanical strength, is a commonly used structure in compressors. However, it faces severe end-area problems under high loads: a complex reverse cavity leakage flow forms between the stator inner ring and the rotor hub groove due to the large pressure difference before and after the stator vane. Although grate seals are designed to limit excessive leakage, some cavity leakage flow after work done by the rotating wall inevitably flows back to the main flow path from the front of the stator vane, affecting secondary flow at the stator root and thus impacting compressor performance. Therefore, reducing the negative impact of leakage flow on compressor performance is crucial.

[0003] Changes to the interstage cavity sealing structure can significantly impact the gas flow pattern within the cavity and the compressor blade passages. Furthermore, changes at different locations within the cavity can affect leakage flow and the main flow differently. Current optimization designs for cavity sealing structures are mostly localized, failing to adequately address the interactions between different local structures. This leads to excessively high cavity resistance and temperature rise, ultimately affecting compressor performance.

[0004] Therefore, there is a need to provide an optimized method and system for sealing the interstage cavity of a compressor to solve the above problems. Summary of the Invention

[0005] To address the problem that current optimization designs for compressor cavity sealing structures are mostly performed locally, failing to adequately match the mutual influence between different local structures, resulting in excessively high cavity resistance and temperature rise, thus affecting compressor performance, this invention provides an optimization method and system for compressor stage cavity sealing structures to solve the existing problems.

[0006] The first aspect of this invention provides an optimization method for the sealing structure of an interstage compressor cavity, the method employing the following technical solution, including: Determine the design parameters of the cavity sealing structure; Based on the preset design parameter range, obtain several cavity sealing structure designs; obtain the compressor efficiency and cavity resistance temperature rise under each cavity sealing structure design; perform correlation analysis on the design parameters and compressor efficiency, and the design parameters and cavity resistance temperature rise to obtain relevant design parameters in the design parameters. Based on the relevant design parameters of each cavity sealing structure, as well as the corresponding compressor efficiency and cavity drag temperature rise, a database is constructed; a neural network is constructed, and the target neural network is trained based on the database. Using the target neural network as a proxy model, with compressor efficiency and cavity resistance temperature rise as optimization objectives, a multi-objective optimization algorithm is used to iteratively optimize the relevant shape parameters in the database. The database is updated based on the shape parameters generated after each iteration and their corresponding compressor efficiency and cavity resistance temperature rise, until the relevant shape parameters corresponding to the maximum compressor efficiency and the minimum cavity resistance temperature rise are obtained. The relevant design parameters corresponding to the maximum compressor efficiency and the minimum cavity resistance temperature rise are taken as the optimal design parameters, and the optimized compressor interstage cavity sealing structure is constructed based on the optimal design parameters.

[0007] A further technical solution of the present invention includes the following shape parameters of the cavity sealing structure: the distance between the side wall of the upstream rectangular cavity away from the stationary blade and the side wall of the upstream leakage channel away from the stationary blade; the distance between the side wall of the upstream rectangular cavity near the stationary blade and the side wall of the upstream leakage channel near the stationary blade; the distance between the side wall of the downstream rectangular cavity away from the stationary blade and the side wall of the downstream leakage channel away from the stationary blade; the distance between the side wall of the downstream rectangular cavity near the stationary blade and the side wall of the downstream leakage channel near the stationary blade; the distance between two adjacent grating teeth; the distance between the side wall of the upstream rectangular cavity near the stationary blade and the grating teeth near the upstream rectangular cavity; the distance between the side wall of the downstream rectangular cavity near the stationary blade and the grating teeth near the downstream rectangular cavity; the distance between the tooth tip of the grating tooth and the bottom of the stationary blade; the tooth tip width of the grating tooth; the bevel length of the grating tooth; the distance from the tooth tip of each grating tooth to the rotation center; the distance from the upper wall surface of the upstream rectangular cavity to the rotation center; and the distance from the upper wall surface of the downstream rectangular cavity to the rotation center.

[0008] A further technical solution of the present invention, the step of obtaining several cavity sealing structure shapes according to a preset shape parameter range is as follows: performing Latin hypercube sampling on the shape parameters within the preset shape parameter range to generate several cavity sealing structure shapes.

[0009] A further technical solution of the present invention, the steps for obtaining the compressor efficiency and cavity resistance temperature rise under each cavity sealing structure design are as follows: Construct a simulation model corresponding to the cavity sealing structure; Numerical simulation calculations were performed on the simulation model to obtain the compressor efficiency and cavity resistance temperature rise corresponding to the sealing structure of each cavity.

[0010] A further technical solution of the present invention includes the following related shape parameters: the distance between the sidewall of the upstream rectangular cavity away from the stationary blade and the sidewall of the upstream leakage channel away from the stationary blade; the distance between the sidewall of the upstream rectangular cavity near the stationary blade and the sidewall of the upstream leakage channel near the stationary blade; the distance between the sidewall of the downstream rectangular cavity away from the stationary blade and the sidewall of the downstream leakage channel away from the stationary blade; the distance between two adjacent grating teeth; the distance between the sidewall of the downstream rectangular cavity near the stationary blade and the grating teeth near the downstream rectangular cavity; the distance between the tooth tip of the grating tooth and the bottom of the stationary blade; the tooth tip width of the grating tooth; the length of the inclined side of the grating tooth; the distance from the upper wall surface of the upstream rectangular cavity to the rotation center; and the distance from the upper wall surface of the downstream rectangular cavity to the rotation center.

[0011] A further technical solution of the present invention involves performing a correlation analysis on the design parameters and compressor efficiency, and on the design parameters and cavity resistance temperature rise, to obtain the relevant design parameters from the design parameters. The steps are as follows: Using the Spearman correlation analysis algorithm, the significance levels of the relationships between the design parameters and compressor efficiency, and between the design parameters and cavity drag temperature rise, were obtained respectively. The styling parameters whose significance level is less than the preset significance level threshold are used as the relevant styling parameters.

[0012] In a further technical solution of the present invention, the neural network adopts a radial basis function neural network.

[0013] A further technical solution of the present invention is that the multi-objective optimization algorithm adopts the NSGA-II genetic algorithm.

[0014] A second aspect of the present invention provides an optimized system for the sealing structure of an interstage compressor cavity, comprising: The styling parameter module is used to determine the styling parameters of the cavity sealing structure; The relevant styling parameter acquisition module is used to acquire several cavity sealing structure shapes according to a preset styling parameter range; acquire the compressor efficiency and cavity resistance temperature rise under each cavity sealing structure shape; perform correlation analysis on the styling parameters and compressor efficiency, and the styling parameters and cavity resistance temperature rise to acquire relevant styling parameters in the styling parameters; The neural network module is used to construct a database based on the relevant design parameters of each cavity sealing structure, as well as the corresponding compressor efficiency and cavity drag temperature rise; construct a neural network, and train the neural network based on the database to obtain a trained target neural network; The target optimization module uses the target neural network as a proxy model, with compressor efficiency and cavity resistance temperature rise as optimization objectives. It uses a multi-objective optimization algorithm to iteratively optimize the relevant shape parameters in the database, and updates the database based on the shape parameters generated after each iteration and their corresponding compressor efficiency and cavity resistance temperature rise, until the relevant shape parameters corresponding to the maximum compressor efficiency and minimum cavity resistance temperature rise are obtained. The modeling module is used to select the relevant modeling parameters corresponding to the maximum compressor efficiency and the minimum cavity resistance temperature rise as the optimal modeling parameters, and to construct the optimized compressor interstage cavity sealing structure based on the optimal modeling parameters.

[0015] The beneficial effects of this invention are: This invention optimizes the entire interstage cavity sealing structure of the compressor by performing global optimization based on the parameterization of the interstage cavity sealing structure. This yields an interstage cavity sealing structure that improves leakage flow characteristics and compressor performance. The advantage of this method is that it comprehensively considers the impact of changes in various local positions within the cavity seal on leakage flow, thereby minimizing the negative impact of leakage flow on compressor performance. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0017] Figure 1 This is a flowchart illustrating an optimization method for an interstage cavity sealing structure of a compressor according to the present invention. Figure 2 This is a detailed flowchart of an optimization method for an interstage cavity sealing structure of a compressor according to an embodiment of the present invention; Figure 3 This is a parameterized schematic diagram of the interstage cavity sealing structure in an embodiment of the present invention; Figure 4 These are the correlation analysis results in the embodiments of the present invention; Figure 5 This is a schematic diagram showing the optimized result of the compressor interstage cavity sealing structure in an embodiment of the present invention; Figure 6 The relative change in cavity leakage characteristics of the cavity sealing structure optimized using the optimization method in this embodiment of the invention compared to the cavity sealing structure before optimization. Detailed Implementation

[0018] 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.

[0019] Current research on the optimization design of cavity sealing structures mainly focuses on improvements in local areas, such as adjustments to the shape of the cavity inlet, outlet, or specific walls. While these local optimization measures improve the characteristics of cavity leakage flow to some extent, the interactions between structural changes at different locations within the cavity mean that local optimization often fails to fully consider the negative impacts of leakage flow on end-region flow. Structural changes at different locations within the cavity have significantly different effects on the flow rate, direction, and interaction with the mainstream flow of the leakage flow. For example, adjusting the shape of the cavity inlet may alter the initial momentum of the leakage flow, while changes in the internal geometry of the cavity will affect the flow path and turbulence characteristics of the leakage flow. Therefore, local optimization of only a few parts of the cavity is insufficient to achieve optimal matching between the leakage flow and the mainstream flow. To minimize the negative impact of cavity leakage flow on the mainstream and fully leverage its potential positive effects on end-region flow, it is necessary to perform global optimization of the cavity structure. Global optimization design not only needs to consider the geometric parameters of each part of the cavity, but also needs to comprehensively consider the interaction mechanism between the leakage flow and the mainstream flow. Through multi-objective optimization methods, the optimal cavity shape is sought. Therefore, this invention provides an embodiment of an optimization method for the interstage cavity sealing structure of a compressor, such as... Figure 1 and Figure 2 As shown, it includes: S1. Determine the design parameters of the cavity sealing structure; For example, in one specific embodiment, since the three-dimensional cavity sealing structure is obtained by rotation, to facilitate parameterization, the process in this embodiment is performed on a two-dimensional structure. First, the two-dimensional cavity sealing structure is discretized into several feature points. The two-dimensional cavity sealing structure is constructed by connecting the feature points. Then, the three-dimensional cavity sealing structure is obtained by rotation. The specific parameterization result is as follows: Figure 3 As shown, the blue frame line represents the cavity sealing structure, and parameterization involves constructing the two-dimensional blue frame line. Figure 3 The shape parameters represent the distances between related feature points. By changing these shape parameters, the distances between feature points can be controlled, thereby altering the entire cavity sealing structure. Therefore, in this embodiment, the shape parameters of the cavity sealing structure include the distance between the sidewall of the upstream rectangular cavity away from the stator vane and the sidewall of the upstream leakage channel away from the stator vane. G 1. The distance between the sidewall of the upstream rectangular cavity near the stator vane and the sidewall of the upstream leakage channel near the stator vane. G2. The distance between the sidewall of the downstream rectangular cavity away from the stator vane and the sidewall of the downstream leakage channel away from the stator vane. G 3. The distance between the sidewall of the downstream rectangular cavity near the stator vane and the sidewall of the downstream leakage channel near the stator vane. G 4. Distance between two adjacent teeth B 1. The distance between the sidewall of the upstream rectangular cavity near the stator blade and the tooth of the upstream rectangular cavity. B 2. The distance between the sidewall of the downstream rectangular cavity near the stator blade and the grate teeth near the downstream rectangular cavity. B 3. Distance between the top of the comb teeth and the bottom of the stationary blade c The width of the tooth tip of the comb e The length of the hypotenuse of the comb teeth H The distance from the upper wall of the downstream rectangular cavity (the wall furthest from the center of rotation) to the center of rotation. R 5. The distance from the upper wall surface (the wall surface furthest from the center of rotation) of the upstream rectangular cavity to the center of rotation. R 6. The cavity sealing structure can be controlled by changing the size of the shape parameters of the cavity sealing structure.

[0020] It should be noted that, as Figure 3 As shown, in this embodiment, the distance between the sidewall of the upstream rectangular cavity away from the stator vane and the sidewall of the upstream leakage channel away from the stator vane is changed. G 1. The distance between the sidewall of the upstream rectangular cavity near the stator vane and the sidewall of the upstream leakage channel near the stator vane. G 2. The distance between the sidewall of the downstream rectangular cavity away from the stator vane and the sidewall of the downstream leakage channel away from the stator vane. G 3. The distance between the sidewall of the downstream rectangular cavity near the stator vane and the sidewall of the downstream leakage channel near the stator vane. G The size of 4 can change the cavity width of the cavity sealing structure; it can also change the distance between two adjacent teeth. B 1. The distance between the sidewall of the upstream rectangular cavity near the stator blade and the tooth of the upstream rectangular cavity. B 2. The distance between the sidewall of the downstream rectangular cavity near the stator and the teeth of the downstream rectangular cavity. B 3. This allows control over the position of the three grates in the cavity sealing structure; it also changes the distance between the grate tips and the bottom of the stationary blades. c The gap between the tooth tips of the grate can be controlled; the tooth tip width e can be controlled; the height of the grate can be controlled by changing the length H of the inclined side; and the maximum distance between the downstream rectangular cavity and the rotation axis can be changed. R 5. The maximum distance between the upstream rectangular cavity and the rotation axis R 6 can change the cavity height of the cavity sealing structure.

[0021] For example, in one specific embodiment, the coordinates of the unknown feature point are obtained through a self-compiled function relational program, specifically expressed as:

[0022] In the formula, i Indicates the first i One known feature point: j Indicates the first j One unknown feature point; using the coordinates and shape parameters of the known feature points, through a function F xj and F yj To obtain the x and y coordinates of the unknown point, i.e. x j and y j The cavity shape can be constructed based on the coordinates of all feature points and the shape parameters.

[0023] S2. Obtain relevant styling parameters; Specifically, based on the preset range of design parameters, several cavity sealing structure designs are obtained; the compressor efficiency and cavity resistance temperature rise under each cavity sealing structure design are obtained; correlation analysis is performed on the design parameters and compressor efficiency, and the design parameters and cavity resistance temperature rise to obtain relevant design parameters in the design parameters.

[0024] For example, in one specific embodiment, the step of obtaining several cavity sealing structure shapes according to the preset shape parameter range is as follows: Latin hypercube sampling is performed on the shape parameters within the preset shape parameter range to generate several cavity sealing structure shape samples. The variation range of the 12 shape parameters is defined under the premise of maintaining the basic structure of the cavity sealing. The specific preset shape parameter range is shown in Table 1, and there are 100 cavity sealing structure shape samples.

[0025] Table 1

[0026] For example, in one specific embodiment, the steps for obtaining the compressor efficiency and cavity resistance temperature rise under each cavity sealing structure design are as follows: constructing a simulation model corresponding to the cavity sealing structure design; performing numerical simulation calculations on the simulation model to obtain the compressor efficiency and cavity resistance temperature rise corresponding to each cavity sealing structure design. It should be noted that, since cavity resistance temperature rise, as one of the performance parameters, can take into account the safety issues caused by excessive temperature rise inside the compressor, and compressor efficiency directly reflects the compressor performance, this embodiment uses compressor efficiency and cavity resistance temperature rise as compressor performance parameters.

[0027] For example, in one specific embodiment, the step of performing correlation analysis on the design parameters and compressor efficiency, and on the design parameters and cavity resistance temperature rise, to obtain the relevant design parameters is as follows: using the Spearman correlation analysis algorithm, the corresponding significance level values ​​between the design parameters and compressor efficiency, and between the design parameters and cavity resistance temperature rise are obtained respectively. P ; the significance level value P The styling parameters that meet the correlation criteria are considered as relevant styling parameters (i.e., at the significance level). P When the significance level is less than a preset significance level threshold, the significance level value satisfies the correlation condition. Specifically, in this embodiment, the significance level threshold is 0.05, that is, when the significance level value is less than a preset significance level threshold, the significance level value satisfies the correlation condition. P When the value is less than 0.05, it indicates that the two are significantly correlated. The results of the correlation analysis are shown in Table 2.

[0028] Table 2

[0029] As shown in Table 2, correlation analysis reveals that among the styling parameters... c , B 1 、B 3. G 1. G 3 is significantly correlated with compressor efficiency; among the design parameters G 1. G 2. G 3. B 3. c, e , H , R 5. R 6. This is significantly related to the compressor cavity resistance and temperature rise. Therefore, in this embodiment, the relevant design parameters are: the distance between the sidewall of the upstream rectangular cavity away from the stationary vane and the sidewall of the upstream leakage channel away from the stationary vane. G 1. The distance between the sidewall of the upstream rectangular cavity near the stator vane and the sidewall of the upstream leakage channel near the stator vane. G 2. The distance between the sidewall of the downstream rectangular cavity away from the stator vane and the sidewall of the downstream leakage channel away from the stator vane. G 3. The distance between two adjacent teeth B 1. The distance between the sidewall of the downstream rectangular cavity near the stator and the tooth of the downstream rectangular cavity. B 3. Distance between the top of the comb teeth and the bottom of the stationary blade c The width of the tooth tip of the comb e The length of the hypotenuse of the comb teeth H The distance from the upper wall of the downstream rectangular cavity to the center of rotation R 5. The distance from the upper wall of the upstream rectangular cavity to the center of rotation. R 6. The correlation coefficients between relevant design parameters and compressor performance parameters are as follows: Figure 4As shown, a positive correlation coefficient indicates a positive correlation between the two parameters, meaning the compressor performance parameters increase with increasing related design parameters; a negative correlation coefficient indicates a negative correlation, meaning the compressor performance parameters decrease with increasing related design parameters. Among these related design parameters... c , B 3. G 1. G 3 is negatively correlated with compressor efficiency, that is c , B 3. G 1. G When 3 decreases, the compressor efficiency increases. c , B 3. G 1. G 3. When the compressor efficiency increases, the compressor efficiency decreases. B 1 is positively correlated with compressor efficiency, that is B As the value increases by 1, the compressor efficiency also increases. (This refers to the relevant design parameters.) G 1. G 2. G 3. H , R 5 is positively correlated with the temperature rise of the cavity air resistance, that is G 1. G 2. G 3. H , R When 5 increases, the temperature rise of the cavity air resistance also increases; B 3. c, e , R 6 is negatively correlated with the temperature rise of the cavity air resistance, that is B 3. c, e , R When 6 increases, the temperature rise of the cavity air resistance decreases.

[0030] S3. Construct a neural network and train it; Specifically, a database is constructed based on the relevant design parameters of each cavity sealing structure, as well as the corresponding compressor efficiency and cavity drag temperature rise; a neural network is constructed, and the target neural network is trained based on the database.

[0031] For example, in one specific embodiment, the neural network adopts a radial basis function (RBF) neural network. That is, in this embodiment, the Gaussian radial basis function (RBF) is used as the activation function of the hidden layer of the RBF neural network, and its mathematical expression is: (1) In the formula, For the first j Output values ​​of radial basis neurons; X For the relevant modeling parameters input,c j For the first radial basis function neural network j The center of each hidden layer neuron, ε j is a shape parameter used to control the width of the Gaussian radial basis function.

[0032] The output layer of the radial basis function neural network uses a linearly weighted combination, and the expression for predicting the objective function value is as follows: (2) In the formula, This represents the neural network's prediction of the objective function. ω j For the weights of the output layer, b For bias terms, m This represents the number of neurons in the hidden layer.

[0033] In the radial basis function neural network training process, orthogonal least squares (OLS) is used to determine the center of neurons in the hidden layer of the RBF. c j Adjusted by the nearest neighbor method ε j The gradient descent method is used to solve for the weights. ω j To minimize the prediction error, the expression is: (3) In the formula, σ This is the output matrix of the hidden layer. This is the true target value; This is the weight vector.

[0034] S4. Obtain the relevant design parameters corresponding to the maximum compressor efficiency and the minimum temperature rise of the cavity resistance; Specifically, the target neural network is used as a surrogate model, with compressor efficiency and cavity resistance temperature rise as optimization objectives. A multi-objective optimization algorithm is used to iteratively optimize the relevant shape parameters in the database. The database is updated based on the shape parameters generated after each iteration and their corresponding compressor efficiency and cavity resistance temperature rise, until the relevant shape parameters corresponding to the maximum compressor efficiency and minimum cavity resistance temperature rise are obtained.

[0035] For example, in one specific embodiment, the step of iteratively optimizing the relevant modeling parameters of the database using a multi-objective optimization algorithm is as follows: In this embodiment, the multi-objective optimization algorithm adopts the NSGA-II genetic algorithm, that is, in this embodiment, the NSGA-II genetic algorithm is embedded in the target neural network in step S3 for iterative optimization. The iterative optimization process is as follows: S41. Initialize the population: The initial population was generated by sampling within a set range of modeling parameters using the Latin hypercube sampling method. P 0, to ensure that the samples are evenly distributed in the decision space.

[0036] S42. Non-dominated sorting and crowding distance calculation: Non-dominated ranking: Pareto stratification is performed on the initial population based on the predictions of the target neural network to calculate the dominance relationship of individuals.

[0037] Crowding distance: Measures the distribution density of solutions in the target space, avoiding local clustering. To maintain the diversity of the solution set, the algorithm calculates the crowding distance for solutions within each Pareto level. Here, the solutions are the initial population divided into different Pareto levels. The expression for crowding distance is: (4) In the formula, For the first i+ One solution is in the first place. k Function values ​​on each target; For the first i The crowding distance of each solution; K is the number of objective functions; For the first i- One solution is in the first place. k Function values ​​on each target.

[0038] S43. Selection, Crossover, and Mutation: Tournament selection: Select high-quality individuals based on non-dominant ranking and crowding distance.

[0039] Simulated Binary Crossover (SBX): Generates offspring individuals; crossover operator formula: (5) (6) In the formula, For cross-distribution parameters, The first offspring produced after crossover; This is the second offspring produced after crossover; The first parent generation to participate in the crossover; The second parent generation to participate in the crossover; The calculation expression is: (7) In the formula, μ c The crossover index is used to control the similarity between offspring and parents. These are random numbers uniformly distributed in [0,1].

[0040] Multinomial mutation: Introducing random perturbations enhances diversity. The mutation formula is: (8) In the formula, The solution after mutation; This is the original solution; The variable; It represents the upper bound of the variable; It represents the lower bound of the variable.

[0041] S44, Elite Retention Strategy: By merging the parent and offspring populations and preserving non-dominated front solutions, the convergence and diversity of Pareto optimal solutions are ensured.

[0042] S45. Update the database and continue iterative optimization until the optimization converges or the set number of iterations is reached: Numerical calculations are performed on the non-dominated front solution in step S44 to obtain the compressor efficiency and cavity drag temperature rise of the generated shape parameters after each iteration optimization. This realizes the iterative optimization of the relevant shape parameters in the database using a multi-objective optimization algorithm, obtaining the generated shape parameters and their corresponding compressor efficiency and cavity drag temperature rise after each iteration optimization. The database is dynamically updated using the generated shape parameters and their corresponding compressor efficiency and cavity drag temperature rise after each iteration optimization until the compressor efficiency is maximized and the cavity drag temperature rise is minimized, or the iteration optimization reaches the required number of cycles, at which point the iteration optimization ends.

[0043] It should be noted that in the iterative optimization of this embodiment, the initial population size of the NSGA-II genetic algorithm is 20, the number of generations is 100, and the maximum number of iterations is 50. The optimal solution given in each round (i.e., the optimized design parameters and the corresponding compressor efficiency and cavity drag temperature rise) is added to the database. The final optimization result is as follows: Figure 5 As shown, from Figure 5 As can be seen, the optimization method of this invention can achieve a cavity sealing structure that improves compressor performance. That is, it improves compressor efficiency while controlling cavity resistance and temperature rise.

[0044] S5. Construct an optimized compressor interstage cavity sealing structure; Specifically, the relevant design parameters corresponding to the maximum compressor efficiency and minimum cavity resistance temperature rise are taken as the optimal design parameters, and an optimized compressor interstage cavity sealing structure is constructed based on the optimal design parameters; the relative change in cavity leakage characteristics of the cavity sealing structure optimized according to the optimization method of this invention compared to the unoptimized cavity sealing structure is as follows: Figure 6 As shown, from Figure 6As can be seen, optimized design C can reduce cavity leakage flow by 41.82% while keeping the cavity resistance temperature rise change less than 1%, and optimized design D can reduce cavity leakage flow by 58.19% while increasing cavity resistance by only 11.84%. This indicates that optimizing the cavity sealing design can effectively reduce leakage flow.

[0045] An optimization system for compressor stage cavity sealing structures includes: a shape parameter module, a related shape parameter acquisition module, a neural network module, a target optimization module, and a shape module. The shape parameter module is used to determine the shape parameters of the cavity sealing structure. The related shape parameter acquisition module is used to acquire several cavity sealing structure shapes according to a preset shape parameter range; acquire the compressor efficiency and cavity drag temperature rise under each cavity sealing structure shape; perform correlation analysis on the shape parameters and compressor efficiency, and the shape parameters and cavity drag temperature rise to acquire related shape parameters. The neural network module is used to construct a database based on the related shape parameters and the corresponding compressor efficiency and cavity drag temperature rise under each cavity sealing structure shape. The process involves: constructing a neural network and training it on a database to obtain a trained target neural network; using the target neural network as a surrogate model, optimizing compressor efficiency and cavity resistance temperature rise as optimization objectives, employing a multi-objective optimization algorithm to iteratively optimize the relevant modeling parameters in the database, and updating the database based on the modeling parameters generated after each iteration and their corresponding compressor efficiency and cavity resistance temperature rise, until the relevant modeling parameters corresponding to the maximum compressor efficiency and minimum cavity resistance temperature rise are obtained; and using the modeling module as the optimal modeling parameters corresponding to the maximum compressor efficiency and minimum cavity resistance temperature rise, constructing an optimized compressor interstage cavity sealing structure based on the optimal modeling parameters.

[0046] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An optimization method for the sealing structure of an interstage cavity in a compressor, characterized in that, include: Determine the design parameters of the cavity sealing structure; Based on the preset range of styling parameters, several cavity sealing structure designs are obtained; Obtain the compressor efficiency and cavity drag temperature rise under each cavity sealing structure design; perform correlation analysis on the design parameters and compressor efficiency, and the design parameters and cavity drag temperature rise to obtain relevant design parameters. Based on the relevant design parameters of each cavity sealing structure, as well as the corresponding compressor efficiency and cavity drag temperature rise, a database is constructed; a neural network is constructed, and the target neural network is trained based on the database. Using the target neural network as a proxy model, with compressor efficiency and cavity resistance temperature rise as optimization objectives, a multi-objective optimization algorithm is used to iteratively optimize the relevant shape parameters in the database. The database is updated based on the shape parameters generated after each iteration and their corresponding compressor efficiency and cavity resistance temperature rise, until the relevant shape parameters corresponding to the maximum compressor efficiency and the minimum cavity resistance temperature rise are obtained. The relevant design parameters corresponding to the maximum compressor efficiency and the minimum cavity resistance temperature rise are taken as the optimal design parameters, and the optimized compressor interstage cavity sealing structure is constructed based on the optimal design parameters.

2. The method for optimizing the sealing structure of an interstage cavity in a compressor according to claim 1, characterized in that, The design parameters of the cavity sealing structure include: the distance between the sidewall of the upstream rectangular cavity away from the stationary blade and the sidewall of the upstream leakage channel away from the stationary blade; the distance between the sidewall of the upstream rectangular cavity near the stationary blade and the sidewall of the upstream leakage channel near the stationary blade; the distance between the sidewall of the downstream rectangular cavity away from the stationary blade and the sidewall of the downstream leakage channel away from the stationary blade; the distance between the sidewall of the downstream rectangular cavity near the stationary blade and the sidewall of the downstream leakage channel near the stationary blade; the distance between two adjacent grating teeth; the distance between the sidewall of the upstream rectangular cavity near the stationary blade and the grating teeth near the upstream rectangular cavity; the distance between the sidewall of the downstream rectangular cavity near the stationary blade and the grating teeth near the downstream rectangular cavity; the distance between the tooth tip of the grating tooth and the bottom of the stationary blade; the tooth tip width of the grating tooth; the bevel length of the grating tooth; the distance from the upper wall surface of the upstream rectangular cavity to the rotation center; and the distance from the upper wall surface of the downstream rectangular cavity to the rotation center.

3. The method for optimizing the sealing structure of an interstage compressor cavity according to claim 1, characterized in that, The steps to obtain several cavity sealing structure shapes according to the preset shape parameter range are as follows: perform Latin hypercube sampling on the shape parameters within the preset shape parameter range to generate several cavity sealing structure shapes.

4. The method for optimizing the sealing structure of an interstage cavity in a compressor according to claim 1, characterized in that, The steps to obtain the compressor efficiency and cavity drag temperature rise under each cavity sealing structure design are as follows: Construct a simulation model corresponding to the shape of the cavity sealing structure; Numerical simulation calculations were performed on the simulation model to obtain the compressor efficiency and cavity resistance temperature rise corresponding to the sealing structure of each cavity.

5. The method for optimizing the sealing structure of an interstage compressor cavity according to claim 1, characterized in that, The relevant design parameters are as follows: the distance between the sidewall of the upstream rectangular cavity away from the stator and the sidewall of the upstream leakage channel away from the stator; the distance between the sidewall of the upstream rectangular cavity near the stator and the sidewall of the upstream leakage channel near the stator; the distance between the sidewall of the downstream rectangular cavity away from the stator and the sidewall of the downstream leakage channel away from the stator; the distance between two adjacent grates; the distance between the sidewall of the downstream rectangular cavity near the stator and the grates near the downstream rectangular cavity; the distance between the grate tooth tip and the bottom of the stator; the grate tooth tip width; the grate tooth bevel length; the distance from the upper wall surface of the upstream rectangular cavity to the rotation center; and the distance from the upper wall surface of the downstream rectangular cavity to the rotation center.

6. The method for optimizing the sealing structure of an interstage cavity in a compressor according to claim 1, characterized in that, The steps for performing correlation analysis between design parameters and compressor efficiency, and between design parameters and cavity resistance temperature rise, to obtain relevant design parameters are as follows: Using the Spearman correlation analysis algorithm, the significance levels of the relationships between the design parameters and compressor efficiency, and between the design parameters and cavity drag temperature rise, were obtained respectively. The styling parameters whose significance level is less than the preset significance level threshold are used as the relevant styling parameters.

7. The method for optimizing the sealing structure of an interstage cavity in a compressor according to claim 1, characterized in that, The neural network used is a radial basis function neural network.

8. The method for optimizing the sealing structure of an interstage cavity in a compressor according to claim 1, characterized in that, The multi-objective optimization algorithm uses the NSGA-II genetic algorithm.

9. An optimized system for sealing the interstage cavity of a compressor, characterized in that, include: The styling parameter module is used to determine the styling parameters of the cavity sealing structure; The relevant styling parameter acquisition module is used to acquire several cavity sealing structure shapes according to a preset styling parameter range; acquire the compressor efficiency and cavity resistance temperature rise under each cavity sealing structure shape; perform correlation analysis on the styling parameters and compressor efficiency, and the styling parameters and cavity resistance temperature rise to acquire relevant styling parameters in the styling parameters; The neural network module is used to construct a database based on the relevant design parameters of each cavity sealing structure, as well as the corresponding compressor efficiency and cavity drag temperature rise; construct a neural network, and train the neural network based on the database to obtain a trained target neural network; The target optimization module uses the target neural network as a proxy model, with compressor efficiency and cavity resistance temperature rise as optimization objectives. It uses a multi-objective optimization algorithm to iteratively optimize the relevant shape parameters in the database, and updates the database based on the shape parameters generated after each iteration and their corresponding compressor efficiency and cavity resistance temperature rise, until the relevant shape parameters corresponding to the maximum compressor efficiency and minimum cavity resistance temperature rise are obtained. The modeling module is used to select the relevant modeling parameters corresponding to the maximum compressor efficiency and the minimum cavity resistance temperature rise as the optimal modeling parameters, and to construct the optimized compressor interstage cavity sealing structure based on the optimal modeling parameters.

Images