Method for stabilizing compressor based on cluster modal cluster energy transient growth rate regulation

Abstract

The application discloses a kind of based on cluster modal cluster energy transient growth rate regulation and control compressor expansion stability method, comprising the following steps: by centrifugal compressor unsteady numerical simulation to obtain flow field data, power mode decomposition is extracted modal feature, modal cluster is obtained using cluster analysis and energy is calculated, define modal cluster energy transient growth rate, determine design variable and variation range, generate sample set and calculate its modal cluster energy transient growth rate, build surrogate model, use non-dominated sorting genetic algorithm for multi-objective optimization, finally obtain optimal design scheme.The application realizes the performance improvement of centrifugal compressor under different working conditions by the optimization method based on cluster modal cluster energy transient growth rate regulation and control, significantly improves the isentropic efficiency and total pressure ratio of compressor, and enhances its operating stability.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of optimal design of impeller machines, and particularly relates to a method for stabilizing expansion of a compressor based on cluster modal cluster energy transient growth rate regulation. BACKGROUND

[0002] As a kind of efficient energy conversion machinery, centrifugal compressor can convert mechanical energy into pressure energy through blades, and is widely used in turbocharged internal combustion engine, small gas turbine of ship, aero-engine and industrial compression device. Therefore, the stability expansion and efficiency improvement of centrifugal compressor has always been the focus of research in this field.

[0003] In the early stage, due to insufficient understanding of flow mechanism, the traditional trial and error method often needs multiple iterative design cycles, which is long and difficult to obtain a global optimal solution. At the same time, this method is heavily dependent on the experience of designers, and the design cycle of those with insufficient experience is longer, and even they cannot get effective results. Therefore, the traditional trial and error method has been difficult to meet the design requirements of modern high-performance compressors. In recent years, with the progress of computational fluid dynamics, mathematical optimization theory and computer hardware, the rapid automatic optimization method based on numerical simulation has been widely used in the aerodynamic optimization of centrifugal compressor inlet guide vane, impeller, diffuser, refluxer, volute and casing treatment. However, this method only optimizes the macroscopic aerodynamic performance index, and cannot actively control the specific flow structure according to the flow mechanism.

[0004] Recently, a new method of centrifugal compressor optimization based on modal energy transfer theory and taking modal energy as objective function is gradually emerging. However, after the modal decomposition of centrifugal compressor unsteady flow field data, there may be a wide frequency modal group with close modal energy proportion and close modal frequency, making it difficult to select the target modal.

[0005] In view of the above problems, it is urgent to propose a method for stabilizing expansion of a compressor based on cluster modal cluster energy transient growth rate regulation. SUMMARY

[0006] To solve the above technical problems, the application provides a method for stabilizing expansion of a compressor based on cluster modal cluster energy transient growth rate regulation.

[0007] The application provides a method for stabilizing expansion of a compressor based on cluster modal cluster energy transient growth rate regulation, which comprises the following steps:

[0008] Obtaining unsteady flow field data based on centrifugal compressor unsteady numerical simulation;

[0009] Performing dynamic modal decomposition based on flow field data, and extracting spatial and time-varying characteristics of the modal;

[0010] Based on the spatial and time-varying characteristics of the modal, modal clusters are obtained by cluster analysis and modal cluster energy is calculated;

[0011] Based on the modal cluster energy, a function relationship of the modal cluster energy transient growth rate is preset;

[0012] Based on the influence of the modal cluster energy transient growth rate on the compressor, the compressor design variables and the change range are determined;

[0013] Based on the compressor design variables, a sample set is generated, and for each sample, an unsteady numerical simulation is performed, and based on the function relationship of the modal cluster energy transient growth rate, the modal cluster energy transient growth rate of each sample is calculated;

[0014] Based on the calculation results, a surrogate model is constructed using a generalized neural network;

[0015] Based on the surrogate model, a non-dominated sorting genetic algorithm is used to implement multi-objective optimization of the compressor, and the design scheme with the optimal transient growth rate is obtained.

[0016] Optionally, the process of obtaining unsteady flow field data based on unsteady numerical simulation of the centrifugal compressor comprises:

[0017] A structured grid is used to generate the grid of the inlet section, the impeller and the diffuser, wherein the O-type grid is used for the impeller blades and the diffuser blades, and the H-type, J-type, C-type and L-type grids are used for the impeller and diffuser passages;

[0018] Based on the grid, numerical simulation is performed using an unsteady Reynolds average method, an SST k-ω model is selected as the turbulence model, a high-order difference format is used for spatial discretization, and a second-order backward Euler format is used for time discretization; the inlet boundary condition is set as total pressure and total temperature, the outlet boundary condition is set as static pressure or mass flow rate, the solid wall surface is set as an adiabatic no-slip boundary condition, and the dynamic and static interface is set as a transient rotor-stator method, to obtain the unsteady flow field of the centrifugal compressor.

[0019] Optionally, the process of performing dynamic modal decomposition based on the flow field data and extracting the spatial and time-varying characteristics of the modal comprises:

[0020] The unsteady flow field data of the centrifugal compressor impeller spanwise section are selected as the analysis object, the dynamic modal decomposition method is used to perform modal decomposition on the selected flow field data, the spatial characteristics of the modal are extracted, and the spatial structure of the modal is represented by the first five-order central moments of the modal; meanwhile, the time-varying characteristics of the modal are extracted, and the time-varying characteristics of the modal are represented by the increase and decrease of the frequency and time coefficient.

[0021] Optionally, the function relationship of the modal cluster energy transient growth rate is:

[0022] The preset total energy of the inherent or characteristic modal cluster before optimization is a first energy value, and the total energy of the inherent or characteristic modal cluster after optimization is a second energy value, and the modal cluster energy transient growth rate is a difference between the second energy value and the first energy value divided by the first energy value.

[0023] Optionally, the process of determining the compressor design variables and the variation range based on the influence of the modal cluster energy transient growth rate on the compressor includes:

[0024] The influence of a single design variable on the inherent modal cluster energy transient growth rate and the characteristic modal cluster energy transient growth rate in the variation range is calculated, and according to the calculation result, a key design variable that has an influence on the modal cluster energy transient growth rate exceeding a preset threshold in the processing of the impeller, the diffuser and the casing is selected as the design space.

[0025] Optionally, the process of generating a sample set based on the compressor design variables, performing unsteady numerical simulation on each sample, and calculating the modal cluster energy transient growth rate of each sample based on the functional relationship of the modal cluster energy transient growth rate includes:

[0026] Based on the compressor design variables, an optimal Latin hypercube sampling method is used to generate a sample set, so that the sample points are uniformly distributed in the entire design space. For each sample in the sample set, an unsteady numerical simulation method is used to obtain unsteady flow field data of the compressor under the highest efficiency condition and the near stall condition. Based on the functional relationship of the modal cluster energy transient growth rate, the modal cluster energy obtained by the dynamic modal decomposition is used to calculate the inherent modal cluster energy transient growth rate and the characteristic modal cluster energy transient growth rate of each sample under the highest efficiency condition and the near stall condition.

[0027] Optionally, the process of constructing a surrogate model based on the calculation result using a generalized neural network includes:

[0028] The inherent modal cluster energy transient growth rate and the characteristic modal cluster energy transient growth rate data of the compressor under different conditions calculated based on the unsteady numerical simulation are collected, a generalized neural network is used to construct a surrogate model, and a mapping relationship between the compressor modal cluster energy transient growth rate and the key design variables is established through the surrogate model.

[0029] Optionally, the process of implementing multi-objective optimization on the compressor based on the surrogate model using a non-dominated sorting genetic algorithm to obtain a design scheme with the optimal transient growth rate includes:

[0030] Based on the mapping relationship between the modal cluster energy transient growth rate and the design variables established by the agent model, a non-dominated sorting genetic algorithm is used for multi-objective optimization of the compressor, the objective functions are to maximize the inherent modal cluster energy transient growth rate at the highest efficiency condition and the near stall condition of the compressor, and to minimize the characteristic modal cluster energy transient growth rate; through iterative calculation of the algorithm, a Pareto front solution set is generated, and the design variable parameters corresponding to the optimal solution are screened out, numerical simulation calculation is carried out on the parameters, and finally the design scheme with the optimal transient growth rate is determined.

[0031] The application also provides a computer device, comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to realize the steps of the method.

[0032] The application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the steps of the method.

[0033] Compared with the prior art, the application has the following advantages and technical effects:

[0034] 1. The power mode decomposition method is used for modal decomposition of the compressor, and the spatial and time-varying characteristics of the mode are extracted, the K-means++ algorithm is introduced for cluster analysis of the modal characteristics, the modal cluster containing similar modes is obtained, and the modal cluster energy is calculated, so that the difficulty that the dominant mode of the compressor stability cannot be correctly selected due to the existence of similar modes is solved, and the rationality and interpretability of the target mode selection are improved.

[0035] 2. The application defines the modal cluster energy transient growth rate, compared with the modal cluster energy proportion, the modal cluster energy transient growth rate is suitable for comparison of modal clusters between different working conditions and different geometric models, and the unreasonable problem of the optimization result of the modal cluster energy proportion as the target function is solved by taking the modal cluster energy transient growth rate as the target function, and high-efficiency and accurate multi-objective optimization is realized. BRIEF DESCRIPTION OF DRAWINGS

[0036] The accompanying drawings, which form a part of this application, are included to provide a further understanding of the application and are incorporated in and constitute a part of this application. The embodiments of this application and their description are used to explain the application and are not intended to limit the application. In the drawings:

[0037] Figure 1 A modal cluster schematic diagram for the highest efficiency condition of the embodiment of the application;

[0038] Figure 2 A modal cluster schematic diagram for the near stall condition of the embodiment of the application;

[0039] Figure 3 A self-circulation casing geometry schematic diagram of an embodiment of the present application;

[0040] Figure 4 A comparison diagram of isentropic efficiency curves before and after optimization of an embodiment of the present application;

[0041] Figure 5 A comparison diagram of total pressure ratio curves before and after optimization of an embodiment of the present application. DETAILED DESCRIPTION

[0042] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0043] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a group of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown herein.

[0044] Embodiment one

[0045] In this embodiment, a compressor stability expansion method based on cluster modal cluster energy transient growth rate regulation is provided, comprising the following steps:

[0046] Unsteady flow field data is obtained based on unsteady numerical simulation of a centrifugal compressor;

[0047] Based on the flow field data, dynamic modal decomposition is performed, and the spatial and time-varying characteristics of the modal are extracted;

[0048] Based on the spatial and time-varying characteristics of the modal, modal clusters are obtained by clustering analysis, and the modal cluster energy is calculated;

[0049] Based on the modal cluster energy, a functional relationship of the modal cluster energy transient growth rate is preset;

[0050] Based on the influence of the modal cluster energy transient growth rate on the compressor, the compressor design variables and the change range are determined;

[0051] Based on the compressor design variables, a sample set is generated, unsteady numerical simulation is performed on each sample, and the modal cluster energy transient growth rate of each sample is calculated based on the functional relationship of the modal cluster energy transient growth rate;

[0052] Based on the calculation results, a surrogate model is constructed using a generalized neural network;

[0053] Based on the surrogate model, a non-dominated sorting genetic algorithm is used to implement multi-objective optimization of the compressor, and the design scheme with the optimal transient growth rate is obtained.

[0054] The process of obtaining the unsteady flow field data based on the centrifugal compressor unsteady numerical simulation can be implemented, and includes the following steps:

[0055] The structured grid is used to generate the grid of the inlet section, the impeller and the diffuser, wherein the O-type grid is used for the impeller blade and the diffuser blade, and the H-type, J-type, C-type and L-type grid is used for the impeller and the diffuser channel; based on the grid, the numerical simulation is performed by using the unsteady Reynolds average method, the SST k-ω model is selected as the turbulence model, the high-order difference format is used for spatial discretization, and the second-order backward Euler format is used for time discretization; the total pressure and the total temperature are set as the inlet boundary condition, the static pressure or the mass flow is set as the outlet boundary condition, the solid wall surface is set as the adiabatic no-slip boundary condition, the transient rotor-stator method is set for the dynamic and static interface, and the centrifugal compressor unsteady flow field is obtained.

[0056] As a specific embodiment, the structured grid is used to generate the grid of the inlet section, the impeller and the diffuser, wherein the O-type grid is used for the impeller blade and the diffuser blade, the H / J / C / L-type grid is used for the impeller and the diffuser channel, the non-structured grid is used for the volute grid, the y+ value of all the grids is less than 5, and the grid density meets the grid independence requirement. In addition, the self-circulation casing is used for the compressor casing treatment, the parameterized modeling is used for the casing, the structured grid is used for the casing grid, and the y+ and the grid setting are consistent with the impeller. The numerical simulation is performed by using the unsteady Reynolds average method, the SST k-ω model is used as the turbulence model, the high-order difference format is used for spatial discretization, and the second-order backward Euler format is used for time discretization. The total pressure is set as 1 atm and the total temperature is set as 298 K as the inlet boundary condition, the static pressure outlet is set as the outlet boundary condition under the large flow condition, and the mass flow outlet is set as the outlet boundary condition under the small flow condition, the solid wall surface is set as the adiabatic no-slip boundary condition, the transient rotor-stator method is set for the interface between the inlet section and the impeller and the interface between the impeller and the diffuser, the frozen rotor method is set for the interface between the diffuser and the volute and the interface between the casing and the inlet section and the impeller, and the centrifugal compressor unsteady flow field is obtained based on the above numerical simulation method.

[0057] The process of performing the dynamic mode decomposition based on the flow field data and extracting the spatial and time-varying characteristics of the mode can be implemented, and includes the following steps:

[0058] The unsteady flow field data of the impeller spanwise section of the centrifugal compressor is selected as the analysis object, the dynamic mode decomposition method is used to perform the mode decomposition on the selected flow field data, the spatial characteristics of the mode are extracted, and the spatial structure of the mode is represented by the first five order central moments of the mode; meanwhile, the time-varying characteristics of the mode are extracted, and the time variation characteristics of the mode are represented by the increase and decrease of the frequency and the time coefficient.

[0059] As a specific embodiment, the compressor unsteady flow field dynamic mode decomposition and mode cluster analysis are implemented. First, the unsteady flow field data of the compressor impeller spanwise section are selected, and the mode decomposition is performed using the dynamic mode decomposition method. Second, the mode space and time-varying characteristics are extracted, the mode space structure is represented by the first five order central moments of the mode, and the mode time-varying characteristics are represented by the frequency and time coefficient increase and decrease. Then, according to the mode characteristics, the K-means++ algorithm is used to cluster the modes to obtain the mode clusters containing similar modes. Finally, according to the impeller flow field analysis, the inherent mode cluster and the characteristic mode cluster are determined, the inherent mode cluster represents the self-excited disturbance and dynamic-static interference structure in the compressor flow field, and the mode space structure and frequency contained in the inherent mode cluster generally do not change with the change of the compressor flow rate, the characteristic mode cluster represents the vortex excitation disturbance structure in the compressor flow field, and the mode space structure and frequency contained in the characteristic mode cluster change with the change of the compressor flow rate, and the total energy of the inherent mode cluster and the total energy of the characteristic mode cluster are calculated.

[0060] The function relationship of the mode cluster energy transient growth rate that can be implemented is:

[0061] The preset total energy of the inherent mode cluster before optimization is a first energy value, and the total energy of the inherent mode cluster after optimization is a second energy value, and the mode cluster energy transient growth rate is the difference between the second energy value and the first energy value divided by the first energy value.

[0062] As a specific embodiment, the mode cluster energy transient growth rate is defined. Assuming that the total energy of the inherent (characteristic) mode cluster before optimization is E1, and the total energy of the inherent (characteristic) mode cluster after optimization is E2, the inherent (characteristic) mode cluster energy transient growth rate is defined as

[0063] The process of determining the compressor design variables and the change range based on the influence of the mode cluster energy transient growth rate on the compressor that can be implemented includes:

[0064] The influence of a single design variable on the inherent mode cluster energy transient growth rate and the characteristic mode cluster energy transient growth rate in the change range is calculated, and according to the calculation result, the key design variables in the impeller, diffuser and casing processing that have an influence on the mode cluster energy transient growth rate exceeding a preset threshold are selected, and the key design variables are determined as the design space.

[0065] As a specific embodiment, a single parameter analysis is implemented to determine the design variables. Under the condition that the compressor geometry is not severely distorted and deformed, the influence of a single design variable on the energy transient growth rate of the natural and characteristic modal cluster in the variation range is calculated, that is, the energy transient growth rate of the natural and characteristic modal cluster corresponding to different values of the single variable in the variation range is calculated, the key design variables in the impeller, diffuser and casing processing which have an impact on the modal cluster energy transient growth rate exceeding 4% are selected, and the variation range of the design variable which makes the natural modal cluster transient growth rate positive and the characteristic modal cluster transient growth rate negative is taken as the design space.

[0066] The process of generating a sample set based on the compressor design variables, performing unsteady numerical simulation on each sample, and calculating the modal cluster energy transient growth rate of each sample based on the functional relationship of the modal cluster energy transient growth rate can be implemented as follows:

[0067] Based on the compressor design variables, an optimal Latin hypercube sampling method is used to generate a sample set, so that the sample points are uniformly distributed in the entire design space. For each sample in the sample set, unsteady numerical simulation method is used to obtain the unsteady flow field data of the compressor under the highest efficiency condition and the near stall condition. Based on the functional relationship of the modal cluster energy transient growth rate, the modal cluster energy obtained by the dynamic modal decomposition is combined to calculate the natural modal cluster energy transient growth rate and the characteristic modal cluster energy transient growth rate of each sample under the highest efficiency condition and the near stall condition.

[0068] As a specific embodiment, an optimal Latin hypercube sampling method is used to generate a design variable sample set. This method is an optimization based on Latin hypercube sampling, so that the sample points are uniformly distributed in the entire design space. The initial sample number in the sample set is generally related to the number of design variables, and the minimum number requirement is (n+1)×(n+2) / 2 times the number of design variables.

[0069] The process of constructing an agent model based on the calculation results using a generalized neural network can be implemented as follows:

[0070] The natural modal cluster energy transient growth rate and the characteristic modal cluster energy transient growth rate data of the compressor under different working conditions calculated based on unsteady numerical simulation are collected, a generalized neural network is used to construct an agent model, and a mapping relationship between the compressor modal cluster energy transient growth rate and the key design variables is established through the agent model.

[0071] As a specific embodiment, a mapping relationship between the compressor geometry parameters and the modal cluster energy transient growth rate is constructed. A generalized neural network (GRNN) is used to construct the mapping relationship between the compressor geometry parameters and the modal cluster energy transient growth rate, to form a surrogate model for rapid evaluation of the compressor modal cluster energy transient growth rate. Based on the surrogate model, the inherent and characteristic modal cluster energy transient growth rates at the highest efficiency operating condition and the near stall operating condition of the compressor are obtained.

[0072] The process of implementing the multi-objective optimization of the compressor based on the surrogate model using the non-dominated sorting genetic algorithm includes:

[0073] Based on the mapping relationship between the modal cluster energy transient growth rate and the design variables established by the surrogate model, the non-dominated sorting genetic algorithm is used to perform multi-objective optimization of the compressor. The objective function is to maximize the inherent modal cluster energy transient growth rate and minimize the characteristic modal cluster energy transient growth rate at the highest efficiency operating condition and the near stall operating condition of the compressor. Through iterative calculation by the algorithm, a Pareto frontier solution set is generated, and the design variable parameters corresponding to the optimal solution are selected. These parameters are numerically simulated and verified to finally determine the design scheme with the optimal transient growth rate.

[0074] As a specific embodiment, the non-dominated sorting genetic algorithm (NSGA-II) is used to perform optimization in combination with the surrogate model. The objective function is to maximize the inherent modal cluster energy transient growth rate at the highest efficiency operating condition and the near stall operating condition of the compressor, and to minimize the characteristic modal cluster energy transient growth rate at the highest efficiency operating condition and the near stall operating condition of the compressor. Finally, a Pareto frontier solution set is obtained, and the optimization scheme with a large improvement in all objective functions is selected for verification, thereby obtaining the optimal design scheme.

[0075] Embodiment Two

[0076] In this embodiment, a certain type of supercharger compressor is used, and the unsteady numerical method in Embodiment One is used for simulation to obtain the unsteady flow field of the compressor at the highest efficiency operating condition (mass flow rate of 0.78 kg / s) and the near stall operating condition (mass flow rate of 0.68 kg / s).

[0077] The unsteady flow field data of the compressor impeller at 96% blade height section at the highest efficiency operating condition and the near stall operating condition are selected for dynamic modal decomposition. The main modes and their corresponding frequencies are shown in Table 1. Cluster analysis is performed on the main modes at the two operating conditions to obtain the modal clusters at the two operating conditions as shown in Figure 1 and Figure 2 The inherent modal cluster at the highest efficiency operating condition is determined to be modal cluster 1 and 3, and the characteristic modal cluster is modal cluster 2 based on the flow field analysis. The inherent modal cluster at the near stall operating condition is determined to be modal cluster 1 and 3, and the characteristic modal cluster is modal cluster 2.

[0078] Table 1

[0079]

[0080] As Figure 3 shown, the design variables and their variation ranges are determined: casing suction groove width g1 (1mm~4mm), suction groove axial position g2 (-47.3mm~-32.7mm), suction groove angle g3 (90°~132°), backflow groove axial position g4 (-80.5mm~-59.5mm), backflow groove angle g5 (120°~160°), cavity lower inclination angle g6 (0°~3°), cavity upper inclination angle g7 (0°~3°), backflow groove cavity radial width g8 (1mm~5.7mm), backflow groove radial width g9 (0.7mm~18.7mm), cavity radial width g10 (22.5mm~32.5mm), cavity axial length g11 (77.0mm~91.9mm), diffuser solidity g12 (6~12). 10 11 12

[0081] A sample set is determined, the number of samples is 120, and the inherent and characteristic modal cluster energy transient growth rates of each sample at the highest efficiency condition and the near stall condition of the compressor are obtained;

[0082] An agent model is constructed, and multi-objective optimization is performed, and the best scheme is obtained: at the highest efficiency condition, the inherent modal cluster energy transient growth rate is 0.08%, and the characteristic modal cluster energy transient growth rate is -1.75%; at the near stall condition, the inherent modal cluster energy transient growth rate is 12.09%, and the characteristic modal cluster energy transient growth rate is -46.39%. As Figure 4 and Figure 5 shown, compared with the original model, the isentropic efficiency and total pressure ratio of the optimized compressor at the highest efficiency condition are increased by 1.2% and 1.1% respectively, and the isentropic efficiency and total pressure ratio at the near stall condition are increased by 3.5% and 4.6% respectively.

[0083] Example Three

[0084] The embodiment also discloses a computer device, which comprises a memory, a processor and a computer program stored in the memory, and the processor executes the computer program to realize the steps of the method in the embodiment one.

[0085] Example Four

[0086] The embodiment also discloses a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the steps of the method in the embodiment one. ​​​

[0087] Example Five

[0088] The embodiment also discloses a computer program product comprising a computer program which, when executed by a processor, implements the steps of the method of the first embodiment.

[0089] The above is only the preferred specific implementation of the present application, but the protection scope of the present application is not limited to this, any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for compressor stability augmentation based on cluster modal energy transient growth rate regulation, characterized in that, The method comprises the following steps: obtaining unsteady flow field data based on unsteady numerical simulation of the centrifugal compressor; performing dynamic mode decomposition based on the flow field data, and extracting spatial and time-varying characteristics of the mode; obtaining mode clusters and calculating mode cluster energy by using cluster analysis based on the spatial and time-varying characteristics of the mode; presetting a functional relationship of mode cluster energy transient growth rate; determining the design variables and variation range of the centrifugal compressor based on the influence of the mode cluster energy transient growth rate on the centrifugal compressor; generating a sample set based on the design variables of the centrifugal compressor, performing unsteady numerical simulation on each sample, and calculating the mode cluster energy transient growth rate of each sample based on the functional relationship of the mode cluster energy transient growth rate; constructing an agent model by using a generalized neural network based on the calculation results; performing multi-objective optimization on the centrifugal compressor by using a non-dominated sorting genetic algorithm based on the agent model, and obtaining a design scheme with optimal transient growth rate.

2. The method of claim 1, wherein the process of obtaining unsteady flow field data based on unsteady numerical simulation of the centrifugal compressor comprises: generating a grid for the inlet section, the impeller and the diffuser by using a structured grid, wherein the O-type grid is used for the impeller blades and the diffuser blades, and the H-type, J-type, C-type and L-type grids are used for the impeller and diffuser passages; performing numerical simulation by using an unsteady Reynolds average method based on the grid, selecting an SST k-ω model as a turbulence model, using a high-order difference format for spatial discretization, and using a second-order backward Euler format for time discretization; setting the inlet boundary condition as total pressure and total temperature, the outlet boundary condition as static pressure or mass flow, the solid wall as an adiabatic no-slip boundary condition, and the dynamic-static interface as a transient rotor-stator method, and obtaining the unsteady flow field of the centrifugal compressor.

3. The method of claim 1, wherein the process of performing dynamic mode decomposition based on the flow field data, and extracting spatial and time-varying characteristics of the mode comprises: selecting the unsteady flow field data of the centrifugal compressor impeller spanwise section as the analysis object, performing mode decomposition on the selected flow field data by using a dynamic mode decomposition method, extracting the spatial characteristics of the mode, and expressing the spatial structure of the mode by using the first five order central moments; and extracting the time-varying characteristics of the mode, and expressing the time variation characteristics of the mode by using the increase and decrease of the frequency and time coefficient.

4. The method of claim 1, wherein the functional relationship of the mode cluster energy transient growth rate is: presetting the total energy of the inherent or characteristic mode cluster before optimization as a first energy value, and the total energy of the inherent or characteristic mode cluster after optimization as a second energy value, then the mode cluster energy transient growth rate is the difference between the second energy value and the first energy value divided by the first energy value; and calculating the inherent mode cluster energy transient growth rate and the characteristic mode cluster energy transient growth rate based on the functional relationship of the mode cluster energy transient growth rate.

5. The method of claim 4, wherein the process of determining the design variables and variation range of the centrifugal compressor based on the influence of the mode cluster energy transient growth rate on the centrifugal compressor comprises: The influence of a single design variable on the energy transient growth rate of the intrinsic modal cluster and the characteristic modal cluster is calculated in a range of variation. According to the calculation results, the key design variables in the impeller, diffuser and casing treatment that have an influence on the energy transient growth rate of the modal cluster exceeding a preset threshold are selected as the design space.

6. The method of claim 1, wherein, The process of generating a sample set based on the compressor design variables, performing unsteady numerical simulation on each sample, and calculating the energy transient growth rate of the modal cluster of each sample based on the functional relationship of the energy transient growth rate of the modal cluster includes: Based on the compressor design variables, an optimal Latin hypercube sampling method is used to generate a sample set, so that the sample points are uniformly distributed in the entire design space. For each sample in the sample set, an unsteady numerical simulation method is used to obtain the unsteady flow field data of the compressor at the highest efficiency condition and the near stall condition. Based on the functional relationship of the energy transient growth rate of the modal cluster, the modal cluster energy obtained by dynamic modal decomposition is used to calculate the intrinsic modal cluster energy transient growth rate and the characteristic modal cluster energy transient growth rate of each sample at the highest efficiency condition and the near stall condition.

7. The method of claim 1, wherein, The process of constructing a surrogate model based on the calculation results using a generalized neural network includes: The intrinsic modal cluster energy transient growth rate and the characteristic modal cluster energy transient growth rate data of the compressor at different operating conditions calculated based on the unsteady numerical simulation are collected, and a generalized neural network is used to construct a surrogate model to establish the mapping relationship between the modal cluster energy transient growth rate of the compressor and the key design variables.

8. The method of claim 1, wherein, The process of implementing multi-objective optimization on the compressor based on the surrogate model using a non-dominated sorting genetic algorithm to obtain the design scheme with the optimal transient growth rate includes: Based on the mapping relationship between the modal cluster energy transient growth rate and the design variables established by the surrogate model, a non-dominated sorting genetic algorithm is used to perform multi-objective optimization on the compressor. The objective function is to maximize the intrinsic modal cluster energy transient growth rate and minimize the characteristic modal cluster energy transient growth rate at the highest efficiency condition and the near stall condition of the compressor. Through iterative calculation by the algorithm, a Pareto frontier solution set is generated, and the design variable parameters corresponding to the optimal solution are screened out. These parameters are subjected to numerical simulation calculation verification to finally determine the design scheme with the optimal transient growth rate.

9. A computer apparatus comprising a memory, a processor, and a computer program stored on the memory, wherein the computer program, when executed by the processor, causes the processor to perform the method of any one of claims 1 to 8. The processor executes the computer program to implement the steps of the method of any one of claims 1-8.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1-8.

Images