Turbine blade design method based on neural network

Abstract

The invention relates to a turbine blade design method based on a neural network, and the method comprises the steps: carrying out the modeling of turbine blade structure parameters, and obtaining a parameterized model; based on the parameterized model, performance evaluation calculation is conducted on the turbine blade structure, and performance evaluation results are obtained; selecting a neural network, and learning the turbine blade structure parameters and the performance evaluation result through the neural network to obtain a mapping relation; and designing the turbine blade to be designed based on the mapping relation. According to the method, the mapping relation between the turbine blade structure parameters and the performance evaluation result is learned through the neural network, and the design efficiency and accuracy of the turbine blade can be improved.

Description

Technical Field

[0001] The present invention belongs to the field of mechanical structure design, and particularly relates to a design method for turbine blades based on neural networks. Background Art

[0002] As a core component of aero-engines and gas turbines, the performance of turbine blades directly affects the thrust, efficiency, and service life of the whole engine. With the rapid development of the aerospace and energy fields, the requirements for the thrust-to-weight ratio, fuel efficiency, and environmental performance of engines are increasing day by day, and the complexity of turbine blade design has also increased accordingly. The design of traditional turbine blades usually needs to consider material selection, complex aerodynamic design, thermodynamics performance analysis, structural strength evaluation, and multidisciplinary coupling problems, resulting in large computational amounts and long design cycles.

[0003] In recent years, the simulation accuracy of turbine blades has been greatly improved. However, the computational time cost of high-precision simulation is huge, and optimization design requires a balance between computational accuracy and efficiency. In addition, the complex double-layer wall structure makes the number of optimization parameters increase sharply, and traditional design methods are unable to cope with multi-objective optimization and multidisciplinary coupling problems.

[0004] Traditional design methods usually separate aerodynamic design, heat transfer calculation, and structural strength analysis, ignoring the strong coupling between disciplines, resulting in self-contradictory optimization results. High-precision simulation calculations require a large amount of computing resources. Especially for the optimization design of complex double-layer wall structures, during multiple iterative optimization processes, the computational time increases exponentially, making it difficult to meet engineering requirements. There is usually a highly non-linear and complex relationship between the structural parameters of turbine blades and their performance, and the empirical formulas relied on by traditional methods are difficult to accurately describe, resulting in untrustworthy optimization results. Summary of the Invention

[0005] In view of the above deficiencies of the prior art, the object of the invention is to provide a design method for turbine blades based on neural networks.

[0006] A design method for turbine blades based on neural networks provided by the present invention includes:

[0007] S1: Modeling the structural parameters of the turbine blade to obtain a parametric model;

[0008] S2: Based on the parametric model, respectively performing performance evaluation calculations on the turbine blade structure to obtain performance evaluation results;

[0009] S3: Selecting a neural network, and learning the structural parameters of the turbine blade and the performance evaluation results through the neural network to obtain a mapping relationship;

[0010] S4: Designing the turbine blade to be designed based on the mapping relationship.

[0011] A turbine blade design method based on a neural network provided by the present invention, the turbine blade structure corresponding to the turbine blade structure parameters in step S1 includes a wall thickness structure, a film hole structure, a film hole arrangement, a connection structure, and a connection arrangement.

[0012] A turbine blade design method based on a neural network provided by the present invention, step S2 further includes:

[0013] S21: Perform a heat transfer calculation on the turbine blade structure to obtain a heat transfer calculation result;

[0014] S22: Perform a strength calculation on the turbine blade structure to obtain a strength calculation result;

[0015] S23: Evaluate the life of the turbine blade structure according to the strength calculation result and the finite element simulation result of the turbine blade structure to obtain a life evaluation result.

[0016] A turbine blade design method based on a neural network provided by the present invention, step S21 further includes:

[0017] S211: Set the position to be mesh-refined for the turbine blade structure, divide it by the finite volume method to obtain a finite volume network;

[0018] S212: Collect fluid mechanics parameters;

[0019] S213: Calculate a heat transfer calculation result by a CFD algorithm according to the fluid mechanics parameters.

[0020] A turbine blade design method based on a neural network provided by the present invention, the heat transfer calculation result in step S213 includes a temperature field and a pressure field.

[0021] A turbine blade design method based on a neural network provided by the present invention, step S22 further includes:

[0022] S221: Set the position to be mesh-refined for the turbine blade structure, divide it by the finite element method to obtain a finite element network;

[0023] S222: Collect blade rotation parameters;

[0024] S223: Calculate a simulation calculation result by an FEA algorithm according to the blade rotation parameters and the heat transfer calculation result.

[0025] A turbine blade design method based on a neural network provided by the present invention, the life evaluation result in step S23 includes a creep life and a fatigue life.

[0026] A turbine blade design method based on a neural network provided by the present invention, step S3 further includes:

[0027] S31: Based on the turbine blade structure parameters and the performance evaluation results, establish a data set;

[0028] S32: Evaluate multiple candidate neural networks through the data set to obtain the optimal neural network;

[0029] S33: Through the optimal neural network, learn the turbine blade structure parameters and the performance evaluation results to obtain the mapping relationship between the turbine blade structure parameters and the turbine blade performance.

[0030] A turbine blade design method based on a neural network provided by the present invention, the candidate neural networks in step S32 include a BP neural network, an RBF neural network, and a GRNN neural network.

[0031] A turbine blade design method based on a neural network provided by the present invention, step S4 further includes:

[0032] S41: According to the mapping relationship, obtain the value range of the structure parameters of the turbine blade to be designed;

[0033] S42: Within the value range of the structure parameters, optimize the structure parameters of the turbine blade to be designed through a metaheuristic algorithm to obtain the optimal design parameters;

[0034] S43: Design the turbine blade to be designed through the optimal design parameters.

[0035] The beneficial effects of the present invention are as follows:

[0036] A turbine blade design method based on a neural network provided by the present invention can significantly reduce the complex calculations and experimental verification processes required by traditional design methods by learning the mapping relationship between the structural parameters of turbine blades and the performance evaluation results through the neural network, thereby greatly improving the design efficiency. The trained neural network can quickly generate design schemes, which is crucial for rapid iteration and optimization of the design process. Secondly, the neural network can extract complex non-linear relationships from a large amount of data, so it can more accurately reflect the actual connection between the structural parameters and performance of turbine blades, helping to design turbine blades with better performance and more in line with actual requirements. The method of the present invention allows considering various turbine blade structural parameters, such as wall thickness structure, film hole structure, connection structure, etc., as well as their combinations and variations, making the design process highly flexible and capable of adapting to different design requirements and application scenarios. In the performance evaluation stage, the present invention not only considers the heat transfer performance and strength performance of turbine blades, but also further evaluates their service life, including creep life and fatigue life. The comprehensive performance evaluation helps to ensure that the designed turbine blades have a reliable service life in actual applications. In addition, the present invention optimizes the structural parameters of turbine blades through a meta-heuristic algorithm, which can further improve the intelligence level of the design. The meta-heuristic algorithm can automatically search for the optimal solution within the given range of structural parameter values, thereby helping designers quickly find the design scheme with the best performance.

[0037] The neural network-based design method of the present invention has good scalability. As new data accumulates continuously, the neural network can continuously learn and improve to adapt to a wider range of design requirements and technological advancements. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The drawings are only for the purpose of showing specific embodiments and are not considered to be a limitation of the present invention. Throughout the drawings, the same reference signs represent the same components. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings.

[0039] Figure 1 It is a schematic flow chart of a turbine blade design method based on a neural network provided by an embodiment of the present invention;

[0040] Figure 2 It is a schematic flow chart of a method for calculating the performance evaluation of the structure of a turbine blade provided by an embodiment of the present invention;

[0041] Figure 3 It is a schematic flow chart of a method for obtaining the mapping relationship between the structural parameters of a turbine blade and the performance of the turbine blade provided by an embodiment of the present invention;

[0042] Figure 4Schematic flow chart of the method for designing a turbine blade to be designed based on a mapping relationship provided by an embodiment of the present invention. Detailed implementation manners

[0043] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0044] In addition, in the following description, the descriptions of well-known structures and technologies are omitted to avoid unnecessarily obscuring the concepts disclosed in the present invention.

[0045] In the description of the present invention, it should be noted that, unless otherwise clearly defined and limited, the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation to the present invention. In addition, the terms "first", "second", and "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. The terms "installation", "connection", and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0046] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all the implementation manners consistent with the present invention. On the contrary, they are merely examples of methods and systems consistent with some aspects of the present invention as detailed in the appended claims.

[0047] The embodiments of the present invention will be described below with reference to the drawings.

[0048] As Figure 1 shown, a method for designing a turbine blade based on a neural network provided by the present invention includes:

[0049] S1: Model the structural parameters of the turbine blade to obtain a parametric model.

[0050] Furthermore, in step S1, first parameterize the wall thickness structure, film hole structure and arrangement parameters, and connection structure and arrangement parameters of the turbine blade, so as to achieve the effect that the model can be automatically modified by adjusting the parameter values of the turbine blade structure. After the parametric modeling is completed, it is also necessary to check whether the model has unexpected errors. If there are no errors, output the parametric model.

[0051] Among them, the turbine blade structure corresponding to the turbine blade structural parameters in step S1 includes a wall thickness structure, a film hole structure, a film hole arrangement, a connection structure, and a connection arrangement.

[0052] Furthermore, for the wall thickness structure, the wall thickness of the turbine blade has an important impact on its heat conduction performance, mechanical strength and durability. By accurately modeling the wall thickness structure, the thermal efficiency and structural strength of the blade can be optimized; for the film hole structure, the film holes are small holes on the turbine blade for cooling. Cooling gas is transported to the blade surface through them to reduce the thermal stress of the blade in a high-temperature environment. The structure of the film holes (such as diameter, shape, distribution, etc.) is crucial for the cooling effect; for the film hole arrangement, the arrangement method of the film holes determines the flow path and coverage range of the cooling gas on the blade surface. A reasonable arrangement can significantly improve the cooling efficiency and reduce the concentration of thermal stress; for the connection structure, the turbine blade is usually connected to other parts of the turbine through a complex connection structure. The design of the connection structure has an important impact on the overall stability and durability of the blade; for the connection arrangement, the connection arrangement refers to the layout and connection method of the connection structure on the turbine blade. A reasonable connection arrangement can ensure the stability of the blade in high-speed rotation and high-temperature environment.

[0053] S2: Based on the parametric model, perform performance evaluation calculations on the turbine blade structure respectively to obtain performance evaluation results.

[0054] As Figure 2 shown, step S2 further includes:

[0055] S21: Perform heat transfer calculations on the turbine blade structure to obtain heat transfer calculation results.

[0056] Step S21 mainly performs heat transfer calculations on the turbine blade. First, the model established by parametric modeling needs to be imported, the parts that require mesh encryption are determined, the mesh size and boundary layer parameters are input, and a finite volume mesh is divided. Then, it is checked whether the mesh meets the division requirements, mainly checking the mesh encryption area and the boundary layer area. After that, information such as the rotational speed, mainstream inlet parameters, and cooling flow inlet parameters is input into CFD (Computational Fluid Dynamics) for heat transfer calculations. Finally, a mesh independence test is carried out to ensure that the calculation results are not affected by the mesh, and high-precision heat transfer calculation results are output. The results mainly include the temperature field and the pressure field.

[0057] Among them, step S21 further includes:

[0058] S211: Set the positions to be mesh encrypted for the turbine blade structure, and divide them by the finite volume method to obtain a finite volume network.

[0059] In step S211, first, the structure of the turbine blade needs to be meshed to form the basis for numerical calculations. The purpose of meshing is to discretize the continuous turbine blade structure into a series of small, interconnected volume units (i.e., finite volumes) so that they can be processed in subsequent calculations.

[0060] Furthermore, in order to obtain more accurate calculation results, the positions to be mesh encrypted are usually set in the key areas of the turbine blade, such as areas with high heat loads, areas with complex flows, etc. These key positions will be divided using denser meshes to improve the resolution and accuracy of the calculations.

[0061] The finite volume method (FVM) is a commonly used numerical calculation method. It constructs discrete equations based on the integral form of the conservation equations. By dividing the turbine blade structure into a finite volume network, it can be ensured that the conservation principle is satisfied within each volume unit, thus obtaining more accurate calculation results.

[0062] S212: Collect fluid mechanics parameters.

[0063] Before performing heat transfer calculations, relevant fluid mechanics parameters need to be collected. These parameters include but are not limited to: the inlet flow velocity and temperature, which determine the state of the fluid entering the turbine blade; the outlet pressure and temperature, which reflect the state change of the fluid after passing through the turbine blade; physical property parameters such as fluid density, specific heat capacity, and thermal conductivity, which have an important impact on heat transfer calculations.

[0064] S213: Through the CFD algorithm, calculate and obtain the heat transfer calculation results according to the fluid mechanics parameters.

[0065] Among them, the heat transfer calculation results in step S213 include the temperature field and the pressure field.

[0066] In step S213, the present invention uses the Computational Fluid Dynamics (CFD) algorithm to calculate the heat transfer performance of the turbine blade. The CFD algorithm is a fluid dynamics simulation technology based on numerical methods, which can simulate the flow and heat transfer processes of fluids in complex geometries.

[0067] Specifically, the CFD algorithm calculates the flow state and temperature distribution of the fluid inside the blade based on the collected fluid mechanics parameters (such as inlet flow velocity, temperature, etc.), as well as the geometric shape and finite volume network of the turbine blade. The obtained calculation results include the temperature field and pressure field, etc., reflecting the heat transfer performance of the turbine blade under specific working conditions.

[0068] S22: Calculate the strength of the turbine blade structure to obtain the strength calculation result.

[0069] Step S22 aims to calculate the strength of the turbine blade. First, the model established by parametric modeling needs to be imported, the parts that need to be mesh-refined are determined, and the finite element mesh is divided by inputting the mesh size. Secondly, check whether the mesh meets the division requirements, mainly checking the mesh-refined area. Then, input information such as rotational speed, boundary constraints, temperature field, and pressure field into the finite element simulation for strength calculation. After that, perform mesh independence verification to ensure that the calculation results are not affected by the mesh, and output high-precision strength calculation results.

[0070] Among them, step S22 further includes:

[0071] S221: Set the positions of the turbine blade structure to be mesh-refined, and divide them by the finite element method to obtain a finite element network.

[0072] In this step, the structure of the turbine blade is further refined for accurate strength calculation. The Finite Element Analysis (FEA) is a commonly used numerical analysis method that divides a complex structure into a series of small, interconnected elements (i.e., finite elements), and equations are established on these elements to solve the stress, strain, etc. of the structure.

[0073] To obtain more accurate calculation results, the positions to be mesh-refined are usually set in the key areas of the turbine blade, such as the stress concentration area, connection parts, etc. These positions will be divided using a denser mesh to improve the resolution and accuracy of the calculation.

[0074] After division by the finite element method, the turbine blade structure is transformed into a network composed of finite elements, and this network will be used for subsequent calculations.

[0075] S222: Collect the blade rotation parameters.

[0076] In step S222, that is, before performing the strength calculation, it is necessary to collect the rotation parameters of the turbine blade. The parameters include the blade's rotation speed, rotation direction, rotation radius, etc., which have important effects on the centrifugal force and gyroscopic moment acting on the blade.

[0077] The purpose of collecting these parameters is to provide the necessary input data for subsequent finite element calculations, so as to accurately simulate the stress conditions of the blade during rotation.

[0078] S223: Use the FEA algorithm to calculate the simulation calculation results based on the blade rotation parameters and the heat transfer calculation results.

[0079] In step S223, the present invention uses the finite element analysis (FEA) algorithm to calculate the structural strength of the turbine blade to solve the stress, strain, etc. of the structure.

[0080] During the calculation process, in addition to considering the blade's geometry and finite element network, it is also necessary to consider the blade rotation parameters and the heat transfer calculation results. The rotation parameters will be used to calculate the centrifugal force, gyroscopic moment, etc. acting on the blade, while the heat transfer calculation results provide the temperature distribution inside the blade, which has an important impact on the thermal stress and thermal expansion of the material.

[0081] Through the calculation of the FEA algorithm, the simulation calculation results such as the stress field and strain field of the turbine blade under specific working conditions can be obtained. These results can be used to evaluate the structural strength of the blade and optimize the blade design accordingly.

[0082] S23: Evaluate the life of the turbine blade structure based on the strength calculation results and the finite element simulation results of the turbine blade structure, and obtain the life evaluation results.

[0083] Among them, the life evaluation results in step S23 include creep life and fatigue life.

[0084] Furthermore, creep is a slow plastic deformation phenomenon that occurs when a material is stressed at high temperatures for a long time. For turbine blades, the creep life mainly depends on the working time of the blade at high temperatures and the stress level it bears.

[0085] Specifically, the creep performance data of the material, such as creep rate and creep limit, can be obtained through creep tests. Then, based on these data, a creep life prediction model can be established. In the finite element simulation results, the stress distribution and temperature distribution of the blade at high temperatures can be extracted. By inputting these data into the creep life prediction model, the creep life of the blade can be obtained.

[0086] For fatigue life assessment, fatigue is a failure phenomenon that occurs in materials under repeated loading. For turbine blades, the fatigue life mainly depends on the number of cyclic loads the blade withstands and the amplitude of each load.

[0087] Specifically, fatigue performance data of materials, such as fatigue limit and fatigue life, can be obtained through fatigue tests. Then, based on these data, a fatigue life prediction model can be established. From the finite element simulation results, the stress-strain history of the blade under cyclic loading can be extracted. By inputting these data into the fatigue life prediction model, the fatigue life of the blade can be obtained.

[0088] S3: Select a neural network and learn the structural parameters of the turbine blade and the performance evaluation results through the neural network to obtain a mapping relationship.

[0089] In the overall step S3, first, extract the structural parameters of the turbine blade and the corresponding temperature, stress, strain, creep life, and fatigue life to establish a turbine blade data set. Divide the data set into a training group and a test group. The training group is responsible for training the neural network, and the test group is responsible for testing the neural network. Evaluate the fitting performance and generalization performance of the neural network from multiple aspects such as variance, standard deviation, mean absolute error, mean absolute percentage error, mean square error, root mean square error, and coefficient of determination. Select the neural network with the best performance based on the performance gap to establish a mapping relationship between the structural parameters of the turbine blade and the performance of the turbine blade.

[0090] As Figure 3 shown, step S3 further includes:

[0091] S31: Based on the structural parameters of the turbine blade and the performance evaluation results, establish a data set.

[0092] In step S31, first, organize the structural parameters, including the geometric dimensions of the blade (such as length, width, thickness, etc.), material properties (such as density, elastic modulus, thermal conductivity, etc.), working environment conditions (such as temperature, pressure, rotational speed, etc.), and the performance evaluation results obtained in the previous steps, which may include strength, life, efficiency, etc., into a data set for subsequent training and evaluation of the neural network.

[0093] S32: Evaluate multiple candidate neural networks through the data set to obtain the optimal neural network.

[0094] In step S32, the data set needs to be preprocessed, such as normalization, standardization, etc., to improve the training efficiency and performance of the neural network. Specifically, the candidate neural network is trained using the data set so that it can learn the mapping relationship between the turbine blade structure parameters and the performance evaluation results; the performance of the neural network is evaluated by methods such as cross-validation and test set validation, and common evaluation metrics such as accuracy, recall rate, F1 score, etc.; finally, the neural network with the best performance is selected as the final model for use according to the evaluation results.

[0095] Among them, the candidate neural networks in step S32 include BP neural network, RBF neural network, and GRNN neural network.

[0096] S33: Learn the turbine blade structure parameters and the performance evaluation results through the optimal neural network to obtain the mapping relationship between the turbine blade structure parameters and the turbine blade performance.

[0097] Finally, step S33 can obtain a neural network model that can accurately predict the performance of turbine blades with different structural parameters. This model can be used to guide the design and optimization of turbine blades, improving the performance and reliability of products.

[0098] S4: Design the turbine blade to be designed based on the mapping relationship.

[0099] In step S4, based on the mapping relationship between the turbine blade structure parameters and the turbine blade performance, the value range of the turbine blade structure parameters is limited, and the meta-heuristic algorithm is used to achieve exploration and development to optimize the turbine blade structure and obtain the optimal turbine blade structure parameters.

[0100] As Figure 4 shown, step S4 further includes:

[0101] S41: Obtain the value range of the structure parameters of the turbine blade to be designed according to the mapping relationship.

[0102] S42: Optimize the structure parameters of the turbine blade to be designed through the meta-heuristic algorithm within the value range of the structure parameters to obtain the optimal design parameters.

[0103] S43: Design the turbine blade to be designed through the optimal design parameters.

[0104] In steps S41 to S43, based on the mapping relationship between the turbine blade structure parameters and the performance, the value range of the turbine blade structure parameters is limited, the optimization goal is determined, and the design space is explored in combination with the meta-heuristic algorithm to optimize the turbine blade structure parameters and obtain the design scheme with the best performance.

[0105] The meta-heuristic algorithms therein include, but are not limited to, genetic algorithms, particle swarm algorithms, sparrow optimization algorithms, and dung beetle optimization algorithms. The optimization objectives therein include performance indicators such as the temperature, stress, creep life, and fatigue life of turbine blades, and multi-objective optimization design can be achieved.

[0106] A turbine blade design method based on a neural network provided by the present invention provides a high-precision and high-efficiency optimization design method for turbine blades based on a neural network, and is used to solve the problems of insufficient multidisciplinary coupling, high computational cost, and unsatisfactory optimization results in existing design methods. First, the present invention performs parametric modeling on the turbine blade structure to quickly generate a design model by adjusting parameter values; secondly, uses the CFD method to perform heat transfer calculations on the turbine blade to obtain the temperature field and pressure field; then, uses the finite element method to perform strength calculations on the turbine blade to obtain the stress field and life evaluation results; after that, based on the structural parameters and performance data of the turbine blade, selects the neural network with the highest accuracy to establish the mapping relationship between the structural parameters and the performance; finally, combines the meta-heuristic algorithm to perform global optimization design to explore the optimal solution of the structural parameters to meet the multi-objective optimization requirements. The present invention can significantly reduce the time cost of the optimization design and improve the optimization efficiency of the turbine blade.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or replacements that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for designing turbine blades based on neural network, characterized in that: include: S1: Model the structural parameters of the turbine blades to obtain a parameterized model; S2: Based on the parameterized model, respectively perform performance evaluation calculations on the turbine blade structure to obtain performance evaluation results; S3: selecting a neural network, learning the turbine blade structural parameters and the performance evaluation results through the neural network, and obtaining a mapping relationship; S4: Designing the turbine blade to be designed based on the mapping relationship.

2. A neural network-based turbine blade design method according to claim 1, characterized in that: The turbine blade structure corresponding to the turbine blade structural parameters in step S1 includes a wall thickness structure, an air film hole structure, an air film hole arrangement, a connection structure, and a connection arrangement.

3. The neural network-based turbine blade design method according to claim 1, characterized in that: Step S2 further comprises: S21: performing heat transfer calculation on the turbine blade structure to obtain heat transfer calculation results; S22: Perform strength calculation on the turbine blade structure to obtain a strength calculation result; S23: Evaluate the life of the turbine blade structure according to the strength calculation result and the finite element simulation result of the turbine blade structure to obtain a life evaluation result.

4. The method for designing turbine blades based on neural network according to claim 3, characterized in that: Step S21 further includes: S211: setting the position of the turbine blade structure to be meshed, dividing it by the finite volume method, and obtaining a finite volume network; S212: Collect fluid mechanics parameters; S213: Obtain heat transfer calculation results according to the fluid mechanics parameters through CFD algorithm.

5. The neural network-based turbine blade design method according to claim 4, characterized in that: The heat transfer calculation results in step S213 include temperature field and pressure field.

6. The neural network-based turbine blade design method according to claim 3, characterized in that: Step S22 further includes: S221: setting the position of the turbine blade structure to be meshed, dividing it by the finite element method, and obtaining a finite element network; S222: Collecting blade rotation parameters; S223: obtaining simulation calculation results according to the blade rotation parameters and the heat transfer calculation results by using an FEA algorithm.

7. A neural network-based turbine blade design method according to claim 3, characterized in that the life assessment result in step S23 includes creep life and fatigue life.

8. The method for designing turbine blades based on neural network according to claim 1, characterized in that step S3 further comprises: S31: establishing a data set based on the turbine blade structural parameters and the performance evaluation results; S32: Evaluate multiple candidate neural networks using the data set to obtain an optimal neural network; S33: The turbine blade structural parameters and the performance evaluation results are learned through the optimal neural network to obtain a mapping relationship between the turbine blade structural parameters and the turbine blade performance.

9. A neural network-based turbine blade design method according to claim 8, characterized in that the candidate neural networks in step S32 include BP neural networks, RBF neural networks, and GRNN neural networks.

10. The neural network-based turbine blade design method according to claim 1, characterized in that: Step S4 further comprises: S41: Obtaining a value range of a structural parameter of a turbine blade to be designed according to the mapping relationship; S42: optimizing the structural parameters of the turbine blade to be designed by using a meta-inspiration algorithm within the range of the structural parameter values ​​to obtain optimal design parameters; S43: Designing the turbine blade to be designed according to the optimal design parameters.

Images