Turbine blade bending, twisting and tilting structure gas-heat coupling optimization method and related device

Abstract

The invention relates to the technical field of structural optimization of turbine blades of aero-engines and gas turbines, in particular to a gas-thermal coupling optimization method and related device for a bending, twisting and tilting structure of a turbine blade, and the method comprises the following steps: parameterizing the bending, twisting and tilting structure of the turbine blade to obtain geometric parameters of the bending, twisting and tilting structure of the turbine blade; a turbine blade objective function is constructed, and a design objective function response value is obtained based on the geometric feature parameters of the turbine blade bending, twisting and tilting structure and the turbine blade objective function; constructing a prediction model based on the design objective function response value, inputting the geometric parameters of the bending, twisting and tilting structure of the turbine blade into the prediction model, and obtaining a cloud picture of the temperature and the total pressure loss of the turbine blade; and based on the cloud atlas of the turbine blade temperature and the total pressure loss, the design objective function response value is optimized, an optimization result is obtained, and the problem that in the prior art, traditional agent model construction cannot meet the simulation requirement of the turbine blade internal flowing and heat exchange process is solved.

Description

Technical Field

[0001] This invention relates to the field of turbine blade structure optimization technology for aero-engines and gas turbines, specifically to a method and related apparatus for optimizing the gas-thermal coupling of the bending, twisting, and tilting structure of turbine blades. Background Technology

[0002] In the modern aviation field, the performance requirements for aircraft are becoming increasingly stringent, with the pursuit of higher speeds, longer ranges, and lower energy consumption becoming core goals of industry development. As the "heart" of an aircraft, the performance of the aero-engine directly determines its overall performance. Meanwhile, gas turbines also occupy a crucial position in numerous fields such as energy, electricity, and ship propulsion, and their efficient and stable operation is essential for ensuring the normal operation of various industries. With continuous technological advancements, the thrust-to-weight ratio and overall performance improvement have become key indicators for the development of aero-engines and gas turbines. This has driven a continuous increase in turbine blade inlet temperatures to achieve more powerful output and higher thermal efficiency. Currently, the inlet temperature of aero-engines under development has exceeded 2200K, far surpassing the temperature resistance limit of high-temperature metals. This high-temperature environment poses unprecedented challenges to the design and manufacturing of turbine blades. Traditional turbine blade designs often suffer from thermal degradation and ablation of blade materials over time due to temperature gradients and uneven heat distribution. This presents a significant challenge to considering the stringent requirements for the safe, stable, and efficient operation of turbine blades during the design process.

[0003] In modern gas turbine engines, the efficiency and durability of turbine blades are key factors determining the overall system performance and reliability. Turbine blade design plays a crucial role in controlling thermal loads and ensuring optimal operation under extreme conditions. Among numerous design aspects, the shape and arrangement of the blades, especially their bending, twisting, and tilting structures, significantly impact the turbine's aerodynamic and heat transfer performance. Therefore, the design of the true three-dimensional blade shape must be given priority. In turbine blade structural optimization design, traditional optimization methods typically employ surrogate models to construct nonlinear input-output relationships between design variables and the objective function, thereby reducing data acquisition costs. However, the internal flow and heat transfer processes of turbine blades are extremely complex, involving high-dimensional nonlinear problems. Traditional surrogate models suffer significant limitations in their applicability and effectiveness when handling such complex problems, failing to accurately capture the internal physical phenomena and performance variations of turbine blades, thus restricting further improvements and optimizations in turbine blade performance. Summary of the Invention

[0004] To address the problem that traditional proxy models in existing technologies cannot meet the simulation requirements of internal flow and heat transfer processes in turbine blades, this invention provides a method and related apparatus for optimizing the gas-thermal coupling of turbine blade bending, twisting, and tilting structures.

[0005] To achieve the above objectives, the present invention employs the following technical solution:

[0006] This invention provides a method for optimizing the aero-thermal coupling of a turbine blade bending-torsional tilting structure, comprising:

[0007] The bending, twisting, and tilting structure of the turbine blade is parameterized to obtain its geometric parameters.

[0008] Construct the objective function for the turbine blade, and obtain the design objective function response value based on the turbine blade bending, torsional and tilting structure geometric parameters and the turbine blade objective function;

[0009] A prediction model is constructed based on the response value of the design objective function. The geometric parameters of the turbine blade bending, twisting and tilting structure are input into the prediction model to obtain the cloud map of turbine blade temperature and total pressure loss.

[0010] The optimization results are obtained by optimizing the response value of the design objective function based on the contour map of turbine blade temperature and total pressure loss.

[0011] The prediction model is a Pix2pix model, and the generator in the Pix2pix model adopts a Res-UNet architecture based on residual networks, while the discriminator adopts a convolutional PatchGAN classifier.

[0012] The optimization results are aimed at obtaining the lowest temperature and the minimum total pressure loss.

[0013] Optionally, the method for parameterizing the turbine blade bending-torsional tilt structure and obtaining the geometric parameters of the turbine blade bending-torsional tilt structure is as follows:

[0014] The blade bends by axially translating the parameter x from the mid-section plane of the blade. mid And the circumferential translation parameter y of the blade mid control;

[0015] The blade twist is caused by the twist angle parameter α of the mid-section plane of the blade. mid and the twist angle parameter α at the blade tip top control;

[0016] The blade tilts due to the axial translation distance parameter x. top And the circumferential translation distance y of the blade top control;

[0017] The axial translation parameter x of the mid-section plane of the blade mid Circumferential translation parameter y of the blade mid The torsional angle parameter α of the blade mid-section plane mid The twist angle parameter α at the tip of the blade top Parameter x of the blade's axial translation distancetop And the circumferential translation distance y of the blade top These together constitute the control parameters of the turbine blades.

[0018] Optionally, the method for constructing the turbine blade objective function and obtaining the design objective function response value based on the turbine blade bending, torsional, and tilting structure geometric parameters and the turbine blade objective function is as follows:

[0019] Obtain the maximum surface temperature of the turbine blades;

[0020] Obtain the total pressure loss coefficient at the turbine blade outlet;

[0021] The design objective function response value is obtained based on the turbine blade control parameters, the maximum surface temperature of the turbine blade, and the total pressure loss coefficient at the turbine blade outlet.

[0022] Optionally, the method for obtaining the total pressure loss coefficient at the turbine blade outlet is as follows:

[0023]

[0024] Among them, C ploss P is the total pressure loss coefficient at the turbine blade outlet. t P is the total pressure at the turbine blade inlet. t,1 P is the total pressure at the turbine blade inlet. t,2 P is the total pressure at the turbine blade outlet. s,2 κ is the static pressure at the turbine blade outlet and κ is the specific heat of the ideal gas.

[0025] Optionally, the method for obtaining the design objective function response value based on the turbine blade control parameters, the maximum surface temperature of the turbine blade, and the total pressure loss coefficient at the turbine blade outlet is as follows:

[0026]

[0027] The constraints of the objective function are:

[0028]

[0029] Among them, T max This represents the maximum temperature on the blade surface.

[0030] Optionally, the objective function of the prediction model is:

[0031]

[0032] Where G is the generator and D is the discriminator. The adversarial loss function is defined by λ, which is the weight hyperparameter of the L1 loss. The loss is the traditional L1 loss, where L1 is the Manhattan norm.

[0033] Optionally, the NSGA-II optimization algorithm is used to optimize the response value of the design objective function based on the contour map of turbine blade temperature and total pressure loss, and the optimization results are obtained.

[0034] A deep learning-based optimization system for the aero-thermal coupling of turbine blade bending-torsional tilting structures, comprising:

[0035] Blade geometry parameterization module: used to parameterize the bending, twisting and tilting structure of turbine blades and obtain the geometric parameters of the bending, twisting and tilting structure of turbine blades;

[0036] Objective function calculation module: used to construct the objective function of the turbine blade, and obtain the design objective function response value based on the bending, twisting and tilting structure geometric parameters of the turbine blade and the objective function of the turbine blade;

[0037] Prediction model building module: used to build a prediction model based on the response value of the design objective function, input the geometric parameters of the turbine blade bending, twisting and tilting structure into the prediction model, and obtain the contour maps of turbine blade temperature and total pressure loss;

[0038] Multi-objective optimization module: used to optimize the response value of the design objective function based on the contour map of turbine blade temperature and total pressure loss, and obtain the optimization results;

[0039] The prediction model is a Pix2pix model, and the generator in the Pix2pix model adopts a Res-UNet architecture based on residual networks, while the discriminator adopts a convolutional PatchGAN classifier.

[0040] The optimization results are aimed at obtaining the lowest temperature and the minimum total pressure loss.

[0041] A terminal device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the steps of the method described above.

[0042] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described above.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] This invention provides a method for optimizing the gas-thermal coupling of the turbine blade's bending-torsional tilt structure. The method involves: obtaining the geometric parameters of the turbine blade's bending-torsional tilt structure; constructing a turbine blade objective function; obtaining the design objective function response value based on the turbine blade's bending-torsional tilt structure geometric parameters and the objective function; constructing a prediction model based on the design objective function response value; inputting the turbine blade's bending-torsional tilt structure geometric parameters into the prediction model to obtain contour maps of the turbine blade's temperature and total pressure loss; and optimizing the design objective function response value based on the turbine blade's temperature and total pressure loss contour maps to obtain the optimization result. The prediction model is a Pix2pix model, and the generator in the Pix2pix model adopts a Res-UNet architecture based on a residual network, while the discriminator uses a convolutional PatchGAN classifier. The Pix2pix model is an image transformation model based on a conditional generative adversarial network (cGAN). It doesn't simply perform numerical mapping; instead, it learns the complex relationships between geometric parameters and corresponding temperature and pressure loss contour maps from a large amount of sample data. In handling high-dimensional nonlinear problems in complex flow and heat transfer processes inside turbine blades, it can accurately fit complex nonlinear input-output relationships and has strong generalization capabilities. Combined with deep generative models (such as variational autoencoders and generative adversarial networks), it can directly generate blade structures and two-dimensional contour maps that meet the target performance conditions. This facilitates joint integration and optimization with multiple disciplines to directly generate intuitive contour map results. The contour maps clearly show the temperature distribution and pressure loss at different locations inside the turbine blade. The generator in the Pix2pix model adopts a Res-UNet architecture based on residual networks, a network with an encoder-decoder structure and skip connections. It can directly pass low-level features from different levels in the encoder to the corresponding levels in the decoder. In the optimization of turbine blade aero-thermal coupling, the temperature and pressure loss distribution of turbine blades are influenced by both their microstructure and macroscopic shape. Skip connections allow the model to retain more detailed information from the encoder when generating contour maps, helping the model to more accurately simulate the internal physical laws of turbine blades and avoiding simulation bias caused by information loss in traditional surrogate models. Simultaneously, the discriminator employs a convolutional PatchGAN classifier. Instead of judging the entire generated contour map as true or false, it divides the contour map into multiple local regions (patches) and independently classifies each region. This local discrimination approach allows the discriminator to focus more on the detailed features of the contour map, enabling the generator to simulate the internal physical phenomena of turbine blades more precisely, thus improving the accuracy and reliability of the contour map.The optimization results aim to obtain the lowest temperature and the minimum total pressure loss. The optimization is carried out based on temperature and total pressure loss cloud maps. The optimization process can fully consider the physical phenomena and performance change laws inside the turbine blades, ensuring that the optimization results not only meet the requirements numerically, but also achieve the best aero-thermal performance of the turbine blades in a physical sense. This optimization method overcomes the problem of the optimization target being disconnected from the physical performance in traditional methods, and improves the practicality and reliability of the optimization results.

[0045] This invention also provides a deep learning-based aero-thermal coupling optimization system for turbine blade bending-torsional tilting structures. This system achieves a high degree of integration through modules for geometric parameter acquisition, objective function calculation, prediction model establishment, and multi-objective optimization. These modules collaborate closely to form an organic whole, realizing the process from acquiring the geometric parameters of the turbine blade bending-torsional tilting structure, obtaining the design objective function response value, acquiring the temperature and total pressure loss contour maps of the turbine blade, and optimizing the design objective function response value based on the temperature and total pressure loss contour maps to obtain the optimization result. This integrated design allows for efficient data flow and sharing within the system, reducing errors and delays in data transmission and improving the coherence and stability of the entire optimization process. Simultaneously, the precise optimization capability enables the system to obtain turbine blade design results that better meet actual needs. Compared with traditional single-objective optimization or simple multi-objective optimization methods, this system can more comprehensively consider the aero-thermal performance indicators of the turbine blade, effectively improving turbine efficiency and reliability while reducing operating costs.

[0046] The present invention also provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the above-described method. The processor can quickly execute the optimization process described above to ensure optimization efficiency and accuracy. The computer program in the memory can be modified and optimized according to actual needs to adapt to the optimization requirements of different turbine blades.

[0047] A computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the method described above. The computer-readable storage medium (such as a solid-state drive (SSD) or flash memory) has high-speed read capabilities, enabling rapid loading of the computer program into the processor for execution. This ensures the efficiency and reliability of the optimization process. It features flexibility and portability, high reliability and stability, support for large-scale data storage, ease of integration and expansion, reduced development and maintenance costs, high security, energy saving and environmental friendliness, support for various application scenarios, and promotion of standardization and normalization. It provides strong support for the safe, stable, and efficient operation of turbine blades and has broad application prospects. Attached Figure Description

[0048] Figure 1 This is a schematic flowchart of an optimization method for the gas-thermal coupling of a turbine blade bending-torsional tilting structure according to the present invention.

[0049] Figure 2 This is a schematic diagram of the blade shape under different geometric parameters according to the present invention, where a is the axial parameter x of the mid-section plane of the blade. mid =10mm control blade bending shape schematic diagram, d is the translation parameter y in the circumferential direction of the blade mid-section plane. mid =10mm control blade bending shape schematic diagram, b is the torsion angle parameter α of the blade mid-section plane. mid =10° control blade twist shape diagram, where e is the twist angle parameter α at the blade tip. top =10° control blade torsion shape schematic diagram, where c is the blade's axial translational distance parameter x top =10mm control blade tilt shape schematic diagram, f is the blade's circumferential translation distance parameter y top =10mm control blade tilt shape diagram.

[0050] Figure 3 This is a schematic diagram of the turbine blade model and boundary conditions, where a is the blade model and b is a schematic diagram of the boundary conditions.

[0051] Figure 4 Schematic diagram of the fluid and solid domain meshes of a turbine blade, where a is the fluid domain mesh, b is the front view of the solid domain mesh, and c is the top view of the solid domain mesh.

[0052] Figure 5 This is a schematic diagram of the Pix2pix model.

[0053] Figure 6 This is a schematic diagram of the Res-Unet architecture based on residual networks.

[0054] Figure 7 This is a schematic diagram of the multi-objective optimization principle based on the Pix2pix model.

[0055] Figure 8 A flowchart for optimizing the design method.

[0056] Figure 9 The image shows a comparison of the temperature distribution of the turbine blades before and after optimization. In the image, a represents the temperature distribution of the unoptimized prototype, and b represents the temperature distribution after optimization.

[0057] Figure 10 The diagram shows a comparison of the total pressure loss coefficient distribution at the turbine blade outlet before and after optimization. In the diagram, a represents the distribution of the total pressure loss coefficient at the blade outlet before optimization, and b represents the distribution of the total pressure loss coefficient at the blade outlet after optimization.

[0058] Figure 11 This is a schematic diagram of a deep learning-based turbine blade bending-torsional-tilting structure aero-thermal coupling optimization system according to the present invention.

[0059] Among them, 1-first cold air inlet, 2-second cold air inlet, 3-tail edge split, 4-mainstream inlet, 5-insulating wall surface, 6-fourth cold air inlet, 7-periodic surface, 8-cold air inlet, 9-third cold air inlet. Detailed Implementation

[0060] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0061] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0062] The present invention will be further described in detail below with reference to specific embodiments. These descriptions are for explanation purposes only and are not intended to limit the scope of the invention.

[0063] This invention discloses an optimization method for the aero-thermal coupling of the bending-torsional tilting structure of turbine blades, referring to... Figure 1 ,include:

[0064] S1: Parameterize the bending, torsion, and tilting structure of the turbine blade to obtain its geometric parameters; the geometric parameters of the turbine blade's bending, torsion, and tilting characteristics include the bending, torsion, and tilting feature geometric parameters of the turbine blade, specifically:

[0065] The control parameters of the turbine blade are obtained by parameterizing the geometric parameters of its bending, torsional, and tilting structures. The method is as follows:

[0066] The blade bends by translating the parameter x along the axial direction (x-axis) of the plane at the mid-section of the blade. mid The translation parameter y in the circumferential direction (y-axis direction) of the blade mid control;

[0067] The blade twist is caused by the twist angle parameter α of the mid-section plane of the blade. mid and the twist angle parameter α at the blade tip top control;

[0068] The blade tilting is caused by the blade's translation distance in the axial direction (x-axis direction) parameter x. top And the translational distance y of the blade in the circumferential direction (y-axis direction) top control;

[0069] The axial translation parameter x of the mid-section plane of the blade mid Circumferential translation parameter y of the blade mid The torsional angle parameter α of the blade mid-section plane mid The twist angle parameter α at the tip of the blade top Parameter x of the blade's axial translation distance top And the circumferential translation distance y of the blade top The control parameters that together constitute the turbine blades are shown in the figure. For blade shapes under different geometric parameters, please refer to [reference needed]. Figure 2 .

[0070] S2: Construct the objective function for the turbine blade, and based on the geometric parameters of the turbine blade's bending, torsional, and tilting structures and the objective function, obtain the design objective function response value, specifically:

[0071] Obtain the maximum surface temperature of the turbine blades; the heat transfer performance of the turbine blades is determined by the maximum surface temperature T. max Conduct an assessment;

[0072] The total pressure loss coefficient at the turbine blade exit is obtained, and the aerodynamic performance of the turbine blade is obtained through the total pressure loss coefficient C at the exit. ploss An evaluation was conducted, and the total pressure loss coefficient C was determined.ploss Defined as:

[0073]

[0074] Among them, C ploss P is the total pressure loss coefficient at the turbine blade outlet. t P is the total pressure at the turbine blade inlet. t,1 P is the total pressure at the turbine blade inlet. t,2 P is the total pressure at the turbine blade outlet. s,2 Let κ be the static pressure at the turbine blade outlet, and κ be the specific heat of the ideal gas, κ = 1.4.

[0075] Based on the turbine blade control parameters, the maximum surface temperature of the turbine blade, and the total pressure loss coefficient at the turbine blade outlet, the design objective function response value is obtained, and its expression is:

[0076]

[0077] The constraints of the objective function are:

[0078]

[0079] S3: Construct a prediction model based on the response value of the design objective function. Input the geometric parameters of the turbine blade's bending, torsional, and tilting structure into the prediction model to obtain contour maps of the turbine blade's temperature and total pressure loss. Specifically:

[0080] Within the design variable constraint range, a sample space is constructed using the Latin hypercube sampling method. Numerical calculations are performed on different sample points to obtain the corresponding design objective function response values ​​within the sample space. After completing the parametric design of the turbine blade's bending-torsional tilting structure, blade and flow channel modeling is performed using UG software. The blade model and boundary condition schematic diagram are shown below. Figure 3 As shown, the blades include a first cold air inlet 1, a second cold air inlet 2, a trailing edge slit 3, a main flow inlet 4, an adiabatic wall 5, a fourth cold air inlet 6, a periodic surface 7, a cold air inlet 8, and a third cold air inlet 9. Fluid and solid domain meshes were created using ANSYS Meshing, and a mesh diagram is shown below. Figure 4 As shown, the three-dimensional flow field distribution of the turbine blade was obtained by performing fluid-structure interaction calculations using ANSYS CFX software.

[0081] A pix2pix prediction model is established between the geometric feature parameters of the sample space and the objective function. The pix2pix model is a conditional generative adversarial network (CGAN) model, consisting of a pair of convolutional networks: a generator G and a discriminator D. Unlike other CGANs, the pix2pix model's generator uses a Res-UNet architecture based on residual networks, and the discriminator uses a convolutional PatchGAN classifier, which penalizes the image by dividing it into several scale blocks. Figure 5 As shown, the output of the trained G cannot be distinguished from the real target image by the adversarially trained D, while the trained D can detect the predicted image of G as well as possible. G takes geometric variables of different bending, torsion and inclination as input and outputs contour maps of temperature and total pressure loss.

[0082] Figure 6 The structure of the Res-UNet generator G, based on a residual network, designed in this patent, is shown for feature extraction and mapping. The first layer of Res-UNet (green square on the left in the figure) is a convolutional layer containing n 3×3 kernels. Following this are three downsampling modules, whose basic unit is a residual network (ResBlock). Downsampling is interpolation and needs to be adjusted according to the actual situation to improve prediction accuracy. Corresponding to the downsampling modules, the upsampling modules contain the same number of ResBlocks. There is a transposed convolutional layer with a 2×2 kernel and a stride of 2. As for the discriminator D, it has four direct convolutional layers, outputting a 32×32 feature map.

[0083] The objective function of the Pix2pix model can be expressed as:

[0084]

[0085] Where G is the generator, which attempts to minimize this objective; and D is the discriminator, which attempts to maximize this objective. Represents the adversarial loss function; It is the traditional L1 distance, which can force low-frequency accuracy of the image, λ is the weight hyperparameter of L1 loss, and L1 is the Manhattan norm.

[0086] The trained prediction model needs to be validated using test samples. The pix2pix model is validated every 10 training steps to monitor its predictive ability on the test samples. The model parameters for each step are saved.

[0087] If the prediction accuracy of the test samples meets the accuracy requirements, the saved pix2pix model is selected for optimization. If the prediction accuracy of the test samples does not meet the requirements, the number of training samples is increased and training continues to ensure that the prediction accuracy of the test samples meets the design requirements.

[0088] S4: Based on the contour plots of turbine blade temperature and total pressure loss, the response value of the design objective function is optimized to obtain the optimization results, specifically:

[0089] After generating the population using the NSGA-II optimization algorithm, the population needs to be sorted by Pareto order, and samples with lower Pareto orders are selected for the next generation. The principle diagram of multi-objective optimization based on the pix2pix model is shown below. Figure 7 As shown.

[0090] The optimization process involved 20 iterations, with a population size of 50 per generation. The crossover factor and mutation factor were 0.8 and 0.05, respectively, and the convergence criterion was the maximum number of iterations.

[0091] The flowchart of the deep learning-based gas-thermal coupling optimization design method for turbine blade bending-torsional tilting structures in this patent is as follows: Figure 8 As shown, it consists of three sub-loops: the UG-ANSYS-Python-UG proxy model training loop, the Python-MATLAB-Python proxy model calling loop, and the UG-ANSYS-Python-MATLAB-UG optimization loop.

[0092] The optimized turbine blades were subjected to strength and vibration verification to ensure that their strength and vibration performance met the relevant standards.

[0093] The fluid load is transferred to the structural solver ANSYS Mechanical.

[0094] The Harmonic Response module is used to verify the yield strength of the optimized turbine blade material and the blade vibration amplitude, thereby ensuring that the blade strength and vibration performance meet the requirements.

[0095] Based on the optimization results, a corresponding UG model was constructed, and numerical verification was performed using ANSYS CFX. Taking a turbine blade as an example, the results are as follows: Figure 9 and Figure 10 As shown, by comparing with the prototype blade, it can be found that blade bending helps reduce the high temperature in the middle of the leading edge of the blade. The average blade temperature decreased from 1070.16K in the prototype to 1061.47K, a reduction of 8.69K (0.81%). Blade bending also helps reduce the total pressure loss at the outlet, with the total pressure loss coefficient decreasing from 1.117‰ in the prototype to 0.790‰, a reduction of 29.28%. The average cooling efficiency of the blade increased from 0.567 to 0.577, an increase of 1.76%. The total cooling air consumption decreased from 0.248 kg·s. -1 Reduced to 0.233 kg·s -1The reduction was 6.05%, which shows that the present invention focuses on the optimized design of the turbine blade bending and twisting structure, fully studies the influence of different bending and twisting structures on the aerodynamic heat transfer performance of the turbine blade, and achieves the optimal bending and twisting blade structure under multiple design objectives.

[0096] See Figure 11 This invention provides a deep learning-based aero-thermal coupling optimization system for the bending-torsional-tilting structure of turbine blades, comprising:

[0097] Blade geometry parameterization module: Parameterizes the bending, twisting and tilting structure of turbine blades and obtains the geometric parameters of the bending, twisting and tilting structure of turbine blades;

[0098] Objective function calculation module: used to construct the objective function of the turbine blade, and obtain the design objective function response value based on the bending, twisting and tilting structure geometric parameters of the turbine blade and the objective function of the turbine blade;

[0099] Prediction model building module: used to build a prediction model based on the response value of the design objective function, input the geometric parameters of the turbine blade bending, twisting and tilting structure into the prediction model, and obtain the contour maps of turbine blade temperature and total pressure loss;

[0100] Multi-objective optimization module: used to optimize the response value of the design objective function based on the contour map of turbine blade temperature and total pressure loss, and obtain the optimization results;

[0101] The prediction model is a Pix2pix model, and the generator in the Pix2pix model adopts a Res-UNet architecture based on residual networks, while the discriminator adopts a convolutional PatchGAN classifier.

[0102] This system achieves a high degree of integration through modules for geometric parameter acquisition, objective function calculation, prediction model establishment, and multi-objective optimization. These modules collaborate closely to form an organic whole, enabling the entire process from acquiring the geometric parameters of the turbine blade's bending, torsional, and tilting structures, obtaining the design objective function response value, acquiring contour maps of turbine blade temperature and total pressure loss, and optimizing the design objective function response value based on these contour maps to obtain the optimization result. This integrated design allows for efficient data flow and sharing within the system, reducing errors and delays during data transmission and improving the coherence and stability of the entire optimization process. Simultaneously, the precise optimization capabilities enable the system to obtain turbine blade design results that better meet actual needs. Compared to traditional single-objective optimization or simple multi-objective optimization methods, this system can more comprehensively consider the aero-thermal performance indicators of the turbine blade, effectively improving turbine efficiency and reliability while reducing operating costs.

[0103] This invention provides a terminal device comprising: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the various method embodiments described above. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the various device embodiments described above.

[0104] The computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention.

[0105] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0106] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0107] The memory can be used to store the computer program and / or module. The processor implements various functions of the terminal device by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0108] If the modules / units integrated into the terminal device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0109] The above description is merely a preferred embodiment of the present invention and is not intended to limit the technical solution of the present invention in any way. Those skilled in the art should understand that, without departing from the spirit and principles of the present invention, the technical solution can be modified and replaced in several simple ways, and these modifications and replacements are all within the scope of protection covered by the claims.

Claims

1. A method for optimizing the aero-thermal coupling of a turbine blade bending-torsional tilting structure, characterized in that, include: The bending, twisting, and tilting structure of the turbine blade is parameterized to obtain its geometric parameters. Construct the objective function for the turbine blade, and obtain the design objective function response value based on the turbine blade bending, torsional and tilting structure geometric parameters and the turbine blade objective function; A prediction model is constructed based on the response value of the design objective function. The geometric parameters of the turbine blade bending, twisting and tilting structure are input into the prediction model to obtain the cloud map of turbine blade temperature and total pressure loss. The optimization results are obtained by optimizing the response value of the design objective function based on the contour map of turbine blade temperature and total pressure loss. The prediction model is a Pix2pix model, and the generator in the Pix2pix model adopts a Res-UNet architecture based on residual networks, while the discriminator adopts a convolutional PatchGAN classifier. The optimization results are aimed at obtaining the lowest temperature and the minimum total pressure loss.

2. The method for optimizing the aero-thermal coupling of the turbine blade bending-torsional tilting structure according to claim 1, characterized in that, The method for parameterizing the turbine blade bending-torsional tilt structure and obtaining its geometric parameters is as follows: The blade bends by axially translating the parameter x from the mid-section plane of the blade. mid And the circumferential translation parameter y of the blade mid control; The blade twist is caused by the twist angle parameter α of the mid-section plane of the blade. mid and the twist angle parameter α at the blade tip top control; The blade tilts due to the axial translation distance parameter x. top And the circumferential translation distance y of the blade top control; The axial translation parameter x of the mid-section plane of the blade mid Circumferential translation parameter y of the blade mid The torsional angle parameter α of the blade mid-section plane mid The twist angle parameter α at the tip of the blade top Parameter x of the blade's axial translation distance top And the circumferential translation distance y of the blade top These together constitute the control parameters of the turbine blades.

3. The method for optimizing the aero-thermal coupling of the turbine blade bending-torsional tilting structure according to claim 2, characterized in that, The method for constructing the objective function of the turbine blade and obtaining the design objective function response value based on the turbine blade bending, torsional, and tilting structure geometric parameters and the turbine blade objective function is as follows: Obtain the maximum surface temperature of the turbine blades; Obtain the total pressure loss coefficient at the turbine blade outlet; The design objective function response value is obtained based on the turbine blade control parameters, the maximum surface temperature of the turbine blade, and the total pressure loss coefficient at the turbine blade outlet.

4. The method for optimizing the aero-thermal coupling of the turbine blade bending-torsional tilting structure according to claim 3, characterized in that, The method for obtaining the total pressure loss coefficient at the turbine blade outlet is as follows: Among them, C ploss P is the total pressure loss coefficient at the turbine blade outlet. t P is the total pressure at the turbine blade inlet. t,1 P is the total pressure at the turbine blade inlet. t,2 P is the total pressure at the turbine blade outlet. s,2 κ is the static pressure at the turbine blade outlet and κ is the specific heat of the ideal gas.

5. The method for optimizing the aero-thermal coupling of the turbine blade bending-torsional tilting structure according to claim 4, characterized in that, The method for obtaining the design objective function response value based on the turbine blade control parameters, the maximum surface temperature of the turbine blade, and the total pressure loss coefficient at the turbine blade outlet is as follows: The constraints of the objective function are: Among them, T max This represents the maximum temperature on the blade surface.

6. The method for optimizing the aero-thermal coupling of the turbine blade bending-torsional tilting structure according to claim 1, characterized in that, The objective function of the prediction model is: Where G is the generator and D is the discriminator. The adversarial loss function is defined by λ, which is the weight hyperparameter of the L1 loss. The loss is the traditional L1 loss, where L1 is the Manhattan norm.

7. The method for optimizing the aero-thermal coupling of the turbine blade bending-torsional tilting structure according to claim 1, characterized in that, The NSGA-II optimization algorithm is used to optimize the response value of the design objective function based on the contour map of turbine blade temperature and total pressure loss, and the optimization results are obtained.

8. A deep learning-based optimization system for the aero-thermal coupling of a turbine blade bending-torsional tilting structure, characterized in that, include: Blade geometry parameterization module: used to parameterize the bending, twisting and tilting structure of turbine blades and obtain the geometric parameters of the bending, twisting and tilting structure of turbine blades; Objective function calculation module: used to construct the objective function of the turbine blade, and obtain the design objective function response value based on the bending, twisting and tilting structure geometric parameters of the turbine blade and the objective function of the turbine blade; Prediction model building module: used to build a prediction model based on the response value of the design objective function, input the geometric parameters of the turbine blade bending, twisting and tilting structure into the prediction model, and obtain the contour maps of turbine blade temperature and total pressure loss; Multi-objective optimization module: used to optimize the response value of the design objective function based on the contour map of turbine blade temperature and total pressure loss, and obtain the optimization results; The prediction model is a Pix2pix model, and the generator in the Pix2pix model adopts a Res-UNet architecture based on residual networks, while the discriminator adopts a convolutional PatchGAN classifier. The optimization results are aimed at obtaining the lowest temperature and the minimum total pressure loss.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-7.

Images