A method, apparatus, equipment and medium for coupled analysis of rotor aerodynamic structure

Abstract

This application discloses a rotor aerodynamic structure coupling analysis method, apparatus, equipment, and medium, relating to the field of rotor performance analysis. The method includes: inputting rotor operating parameters and model parameters at the current aerodynamic time step into rotor blade models of different fidelity for solution, obtaining current aerodynamic load parameters corresponding to different fidelity levels; inputting the current rotor operating parameters and aerodynamic load parameters into a structural deformation prediction model to obtain the current rotor structural deformation; the structural deformation prediction model is determined based on a reduced-order modeling method and a neural network model; updating the spatial positions of each section of the blade according to the current rotor structural deformation at different fidelity levels, changing the blade geometry and aerodynamic boundary conditions, and using the changed geometry and aerodynamic boundary conditions as model parameters for the next aerodynamic time step, until the rotor structural deformation at all aerodynamic time steps is obtained. This application can reduce overall computational costs while ensuring aerodynamic analysis accuracy.

Description

Technical Field

[0001] This application relates to the field of rotor performance analysis, and in particular to a rotor aerodynamic structure coupling analysis method, apparatus, equipment and medium. Background Technology

[0002] Predicting rotor aerodynamic performance is a crucial issue in the design and performance evaluation of rotorcraft, as the analysis results directly impact the rationality of rotor aerodynamic layout, load distribution, and structural design. The rotor flow field exhibits highly unsteady characteristics, especially under maneuvering flight conditions (such as rapid pull-up, roll maneuvers, and sudden changes in collective pitch) and complex incoming flow environments. In these conditions, rotor aerodynamic loads and structural responses display significant time-varying characteristics and strong nonlinear features, placing higher demands on the applicability and computational efficiency of aerodynamic analysis models. In engineering practice, sufficient accuracy in aerodynamic analysis results is required, while computational efficiency must meet the needs of multi-condition analysis and design iteration.

[0003] Currently, there are several main methods for simulating unsteady flow fields in rotors.

[0004] The first category is high-fidelity methods based on computational fluid dynamics (CFD). These methods can accurately describe the complex unsteady flow characteristics around the rotor by solving the Navier-Stokes equations (NS equations), but their computational cost is high, making them difficult to apply directly to large-scale parameter analysis, multi-condition evaluation, or rapid design phases at present.

[0005] The second category is based on low-fidelity or medium-fidelity aerodynamic models, including the Free Wake Model (FWM) and the Vortex Particle Method (VPM). By simplifying or approximating the evolution of the rotor wake, the computational scale is significantly reduced, but there are certain limitations in terms of the ability to resolve flow field details and the accuracy of prediction.

[0006] The third category is a combination of medium- and high-fidelity aerodynamic analysis methods, which uses aerodynamic models with different accuracies and computational costs in combination. High-fidelity CFD is used to solve the near-field region of the rotor, while VPM or FWM is used to solve the far-field wake region, so as to realize the partitioned modeling of the unsteady flow field of the rotor and thus achieve a trade-off between prediction accuracy and computational efficiency.

[0007] However, most existing studies employ rigid rotors or indirectly consider structural effects through empirical corrections or equivalent stiffness corrections. This makes it difficult to accurately reflect the impact of rotor structural deformation on aerodynamic load distribution and flow field characteristics without explicitly introducing structural solutions. Further employing strongly coupled aerodynamic-structural analysis methods, simultaneously solving the fluid control equations and structural dynamics equations in each iteration step, can improve analytical accuracy but significantly increases overall computational complexity and cost. This limits its application in rapid multi-condition analysis and iterative engineering design.

[0008] Therefore, how to reduce the overall computational cost while ensuring the accuracy of aerodynamic analysis has become an urgent problem to be solved. Summary of the Invention

[0009] The purpose of this application is to provide a rotor aerodynamic structure coupling analysis method, device, equipment and medium, which can reduce the overall calculation cost while ensuring the accuracy of aerodynamic analysis.

[0010] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a rotor aerodynamic structure coupling analysis method, including: Obtain the rotor operating parameters and model parameters of the target rotor at the current aerodynamic time step; Construct rotor blade models with different fidelity; Input the rotor operating parameters and model parameters at the current aerodynamic time step into rotor blade models with different fidelity, and use the aerodynamic solver to solve the rotor blade models with different fidelity to obtain the current aerodynamic load parameters corresponding to different fidelity. The rotor operating parameters at the current aerodynamic time step and the current aerodynamic load parameters corresponding to different fidelities are input into the structural deformation prediction model to obtain the rotor structural deformation at different fidelities at the current aerodynamic time step; wherein, the structural deformation prediction model is determined based on the order reduction modeling method and the neural network model; The spatial positions of each section of the target rotor blade are updated based on the rotor structure deformation at different fidelities under the current aerodynamic time step, thereby changing the geometric shape and aerodynamic boundary conditions of the target rotor blade. The changed geometric shape and aerodynamic boundary conditions are used as model parameters under the next aerodynamic time step, until the rotor structure deformation under all aerodynamic time steps is obtained.

[0011] In one embodiment, the method for determining the structural deformation prediction model includes: Construct structural deformation sample data; the structural deformation sample data includes: structural deformation field data of rotor blades used for training under different azimuth angles and different operating parameters; The structural deformation sample data is subjected to feature extraction and dimensionality reduction using a reduced-order modeling method to obtain structural deformation features; the structural deformation features include: rotor structural deformation amount used for training; Different operating parameters and aerodynamic load parameters are used as inputs to the neural network model, and the structural deformation characteristics are used as the outputs of the neural network model. The model is trained using supervised learning, and the trained neural network model is determined as the structural deformation prediction model.

[0012] In one embodiment, constructing structural deformation sample data specifically includes: Obtain the geometric parameters, cross-sectional properties, and constraint methods of the rotor blades used for training; Based on the geometric parameters, the cross-sectional properties, and the constraint method, a structural dynamics model of the rotor blade for training is established using geometrically precise beam theory. The structural dynamics model of the rotor blade used for training was solved under different operating parameters to obtain structural deformation field data under different azimuth angles and different operating parameters, thereby determining the structural deformation sample data.

[0013] In one embodiment, a reduced-order modeling method is used to extract and reduce the dimensionality of the structural deformation sample data to obtain structural deformation features, specifically including: The structural deformation sample data is decomposed using the intrinsic orthogonal decomposition method to obtain multiple orthogonal modes and corresponding structural deformation mode coefficients; The structural deformation modal coefficients corresponding to orthogonal modes with energy contributions greater than a set value are selected to characterize the rotor structure deformation, thus obtaining the structural deformation characteristics.

[0014] In one embodiment, constructing rotor blade models with different fidelity specifically includes: A three-dimensional blade mesh model was constructed using computational fluid dynamics to obtain a high-fidelity rotor blade model; A vortex lattice model was constructed using the free wake model or the vortex particle method to obtain a rotor blade model with medium to low fidelity. Computational fluid dynamics and eddy particle methods were used to construct the blade model, resulting in a high-fidelity rotor blade model.

[0015] In one embodiment, the neural network model is a multilayer perceptron.

[0016] In one embodiment, after obtaining the rotor structure deformation at all aerodynamic time steps, the rotor aerodynamic structure coupling analysis method further includes: The rotor structure deformation at each aerodynamic time step is processed to form time history data and output; the time history data includes the rotor disk load distribution, vortex field evolution results and structural deformation response at each aerodynamic time step.

[0017] Secondly, this application provides a rotor aerodynamic structure coupling analysis device, comprising: The parameter acquisition module is used to acquire the rotor operating parameters and model parameters of the target rotor at the current aerodynamic time step. The model building module is used to build rotor blade models with different fidelity. The model solving module is used to input the rotor operating parameters and model parameters at the current aerodynamic time step into rotor blade models with different fidelity. The aerodynamic solver is used to solve the rotor blade models with different fidelity to obtain the current aerodynamic load parameters corresponding to different fidelity. The structural deformation prediction module is used to input the rotor operating parameters at the current aerodynamic time step and the current aerodynamic load parameters corresponding to different fidelities into the structural deformation prediction model to obtain the rotor structural deformation at different fidelities at the current aerodynamic time step; wherein, the structural deformation prediction model is determined based on the order reduction modeling method and the neural network model; The aerodynamic structure coupling analysis module is used to update the spatial position of each section of the target rotor blade based on the rotor structure deformation at different fidelities under the current aerodynamic time step, thereby changing the geometric shape and aerodynamic boundary conditions of the target rotor blade. The changed geometric shape and aerodynamic boundary conditions are used as model parameters under the next aerodynamic time step, until the rotor structure deformation under all aerodynamic time steps is obtained.

[0018] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the rotor aerodynamic structure coupling analysis method described in any one of the above.

[0019] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the rotor aerodynamic structure coupling analysis method described above.

[0020] According to the specific embodiments provided in this application, this application has the following technical effects: This application provides a rotor aerodynamic structure coupling analysis method, device, equipment, and medium. Based on the characteristics of rotor blade modeling and aerodynamic load calculation with different fidelity, the structural deformation prediction model determined by the reduced-order modeling method and neural network model is embedded into the aerodynamic solution process. Without explicitly solving the structural mechanics equations, the rotor structure deformation is predicted in real time, and the rotor geometry and aerodynamic boundary conditions are updated accordingly. This enables the coupling analysis of rotor aerodynamic load and structural deformation effect, ensuring the accuracy of aerodynamic analysis while reducing the overall computational cost. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A schematic flowchart illustrating a rotor aerodynamic structure coupling analysis method provided in this application embodiment; Figure 2 A flowchart illustrating the method for determining a structural deformation prediction model provided in an embodiment of this application; Figure 3 This is a schematic diagram of the data flow coupling between rotor aerodynamic load and structural deformation provided in an embodiment of this application; Figure 4 A flowchart of the structural solution provided for an embodiment of this application; Figure 5 A high-fidelity aerodynamic solution flowchart provided for embodiments of this application; Figure 6 A flowchart of medium-to-high fidelity aerodynamic solution provided for embodiments of this application; Figure 7 A flowchart of low-to-medium fidelity aerodynamic solution provided for embodiments of this application; Figure 8 A schematic diagram of the waving deformation provided in the embodiments of this application; Figure 9 A schematic diagram of torsional deformation provided for an embodiment of this application; Figure 10 Visualization results of the first-order modes provided in the embodiments of this application; Figure 11 Visualization results of the second-order modes provided in the embodiments of this application; Figure 12 Visualization results of the third-order modes provided in the embodiments of this application; Figure 13 Visualization results of the fourth-order modes provided in the embodiments of this application; Figure 14 This is a schematic diagram of the structure of the neural network model provided in the embodiments of this application; Figure 15 This is a schematic diagram of a three-dimensional blade mesh model provided in an embodiment of this application; Figure 16 A schematic diagram of a vortex lattice model provided in an embodiment of this application; Figure 17 A schematic diagram of mesh deformation provided for an embodiment of this application; Figure 18 A schematic diagram of vortex lattice deformation provided in an embodiment of this application; Figure 19 This is a schematic diagram of the propeller disk load distribution provided in an embodiment of this application; Figure 20 This is a schematic diagram of the rotor vortex field during the pull-up motion provided in an embodiment of this application; Figure 21 This is a schematic diagram of the propeller tip torsional deformation response provided in an embodiment of this application; Figure 22 A schematic diagram of the functional modules of a rotor aerodynamic structure coupling analysis device provided in this application embodiment; Figure 23 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] In the context of reducing overall computational costs while maintaining the accuracy of aerodynamic analysis, data-driven and machine learning methods have been increasingly introduced into aerodynamic and structural analysis to construct surrogate models of complex physical processes. Among these, neural network (NN) models, with their ability to map nonlinear relationships, are used to learn the mapping relationships between operating parameters, aerodynamic loads, and structural responses, thereby reducing reliance on numerical solutions to some extent. However, directly using neural networks to predict the deformation field of high-dimensional structures often requires a large number of training samples, and the model's generalization ability, stability, and physical consistency are difficult to guarantee, thus limiting its application in engineering.

[0026] To address the challenge of efficiently incorporating structural deformations into aerodynamic analysis, a feasible approach is to combine reduced-order modeling techniques such as Proper Orthogonal Decomposition (POD) with neural networks. This reduces the order of high-dimensional structural deformation data, retaining only low-dimensional modes and their corresponding coefficients that characterize the main structural deformation features. The neural network then learns the mapping relationship between aerodynamic loads and modal coefficients, enabling rapid prediction of structural deformation. This approach offers potential advantages in reducing problem dimensionality, decreasing training sample requirements, and improving prediction efficiency.

[0027] However, in the practice of rotor aerodynamic analysis and structural modeling, such structural deformation modeling methods based on order reduction and machine learning have not yet formed an engineering integration scheme with aerodynamic analysis models. In particular, they lack a unified interface and calling mechanism that can adapt to multiple aerodynamic models in rotor multi-fidelity aerodynamic analysis. Therefore, it is difficult to use them stably and reliably to compensate for structural deformation effects in multi-fidelity aerodynamic analysis processes.

[0028] This application can significantly reduce the overall computational cost while ensuring the accuracy and numerical stability of aerodynamic analysis. It is applicable to rotor aerodynamic structure coupling analysis, rotor aerodynamic performance evaluation and engineering design analysis under multiple operating conditions.

[0029] In one exemplary embodiment, such as Figure 1 As shown, a rotor aerodynamic structure coupling analysis method is provided, including: Step 101: Obtain the rotor operating parameters and model parameters of the target rotor at the current aerodynamic time step.

[0030] Step 102: Construct rotor blade models with different fidelity.

[0031] Step 103: Input the rotor operating parameters and model parameters at the current aerodynamic time step into rotor blade models with different fidelity, and use the aerodynamic solver to solve the rotor blade models with different fidelity to obtain the current aerodynamic load parameters corresponding to different fidelity.

[0032] Step 104: Input the rotor operating parameters at the current aerodynamic time step and the current aerodynamic load parameters corresponding to different fidelities into the structural deformation prediction model to obtain the rotor structural deformation at different fidelities at the current aerodynamic time step; wherein, the structural deformation prediction model is determined based on the order reduction modeling method and the neural network model.

[0033] Step 105: Update the spatial position of each section of the target rotor blade according to the rotor structure deformation at different fidelities under the current aerodynamic time step, thereby changing the geometric shape and aerodynamic boundary conditions of the target rotor blade. Use the changed geometric shape and aerodynamic boundary conditions as model parameters under the next aerodynamic time step, until the rotor structure deformation under all aerodynamic time steps is obtained.

[0034] In one implementation, such as Figure 2 As shown, the method for determining the structural deformation prediction model includes: Step 201: Construct structural deformation sample data; the structural deformation sample data includes: structural deformation field data of the rotor blades used for training under different azimuth angles and different operating parameters. Specifically: (1) Obtain the geometric parameters (rotor radius, blade chord length, installation angle distribution, airfoil, etc.), cross-sectional properties (mass distribution, center of mass position, bending stiffness, torsional stiffness, etc.) and constraint methods (articulated, hingeless) of the rotor blades used for training.

[0035] (2) Based on the geometric parameters, the cross-sectional properties and the constraint method, a structural dynamics model of the rotor blade for training is established using the Geometrically Exact Beam Theory.

[0036] (3) Solve the structural dynamics model of the rotor blade used for training under different operating parameters to obtain structural deformation field data under different azimuth angles (at different times) and different operating parameters, thereby determining the structural deformation sample data. The operating parameters include, but are not limited to, forward speed, rotor pitch angle, incoming flow angle, rotor thrust, rotor pitch moment, and rotor roll moment. The structural deformation field data includes: flapping deformation, torsional deformation, etc.

[0037] Step 202: Use a reduced-order modeling method to extract features and reduce the dimensionality of the structural deformation sample data to obtain structural deformation features.

[0038] The structural deformation features include: the amount of rotor structural deformation used for training.

[0039] This step specifically includes: decomposing the structural deformation sample data using the intrinsic orthogonal decomposition (POD) method to obtain multiple orthogonal modes and corresponding structural deformation mode coefficients; selecting the structural deformation mode coefficients corresponding to the orthogonal modes with energy contributions greater than a set value to characterize the rotor structure deformation amount, thereby obtaining structural deformation characteristics.

[0040] Step 203: Different working condition parameters and aerodynamic load parameters are used as inputs to the neural network model, and the structural deformation characteristics are used as the outputs of the neural network model. The model is trained using supervised learning, and the trained neural network model is determined as the structural deformation prediction model.

[0041] The neural network model can be a multilayer perceptron (MLP) structure with no fewer than four hidden layers to ensure the network's ability to learn nonlinear mappings.

[0042] In another exemplary embodiment of this application, step 102 specifically includes: A three-dimensional blade mesh model was constructed using computational fluid dynamics (CFD) to obtain a high-fidelity rotor blade model; a vortex lattice model was constructed using the free wake model (FWM) or the vortex particle method (VPM) to obtain a medium-to-low fidelity rotor blade model; and a blade model was constructed using both computational fluid dynamics and the vortex particle method to obtain a medium-to-high fidelity rotor blade model.

[0043] In another exemplary embodiment of this application, after step 105, the method further includes: organizing the rotor structure deformation (i.e., coupling analysis results) at each aerodynamic time step to form time history data and outputting it; the time history data includes the rotor disk load distribution, vortex field evolution results, and structural deformation response at each aerodynamic time step.

[0044] In practical applications, a more specific implementation process of the above-mentioned rotor aerodynamic structure coupling analysis method is as follows.

[0045] This method introduces a neural network surrogate model (i.e., a structural deformation prediction model) into the multi-fidelity aerodynamic analysis process to predict rotor structural deformation. This achieves efficient coupling between rotor aerodynamic loads and structural deformation effects without explicitly solving the structural mechanics equations. The data flow is as follows: Figure 3 As shown. The structure solution is divided into two parts: offline training and online inference. The specific process is as follows: Figure 4 As shown. Three aerodynamic analysis methods with different fidelity are employed for the aerodynamic solution: high-fidelity CFD, medium-high fidelity CFD / VPM, and medium-low fidelity VPM. The high-fidelity method uses a motion-nested mesh approach. Before solving, the motion and blade mesh deformation are updated, followed by solving the mesh blocks. After convergence, the blade loads are calculated and processed, and the load results, along with the rotor operating conditions, are transferred to the structural model for inference. The medium-high fidelity method has a similar workflow to the high-fidelity method, but it uses Lagrange particles to describe the wake, thus requiring particle updates and mesh flux corrections in the workflow. The medium-low fidelity method removes the CFD solution portion from the medium-high fidelity workflow and uses VPM entirely to calculate aerodynamic loads. The specific workflow is as follows: Figures 5-7 As shown.

[0046] The specific details of the rotor aerodynamic structure coupling analysis method are as follows.

[0047] Step 1: Construction of structural deformation sample data.

[0048] Based on the geometric parameters, cross-sectional characteristics, and constraint methods of the rotor blades, a structural dynamics model of the rotor blades is established using geometrically precise beam theory. The structural response of the rotor blades is then solved under various flight conditions, including forward speed, rotor pitch angle, incoming flow angle, rotor thrust, rotor pitch moment, and rotor roll moment.

[0049] Through the above structural solution process, structural deformation field data of the rotor blade under different azimuth angles (different times) and different operating conditions are obtained, including flapping deformation and torsional deformation, which are used to construct the structural deformation sample dataset required for training the neural network model. Figure 8 and Figure 9 The deformation of a single blade at different azimuth angles is shown. Figure 8 middle It is the amount of deformation during waving. Where is the rotor radius.

[0050] Step 2: Extraction of structural deformation features and data dimensionality reduction.

[0051] Feature extraction and dimensionality reduction are performed on the high-dimensional structural deformation field sample data obtained in step 1 to reduce the data dimensionality and extract the dominant features of structural deformation. Intrinsic orthogonal decomposition (POD) is used to represent the structural deformation field as a linear combination of several orthogonal modes and their corresponding modal coefficients, as shown in the following formula.

[0052] .

[0053] In the formula, This is the original data. Here is the modal matrix. These are the modal coefficients.

[0054] Figures 10-13 The visualization results of the first four modes are presented (normalized). Figures 10-13 The horizontal and vertical axes and Based on rotor radius Normalized planar position coordinates. These modes are essentially feature vectors extracted from structural deformation field samples, sorted by their energy contribution. Figures 10-13 The upper right corner of the graph gives the energy percentage of the modes, which can characterize the main deformation modes of the rotor structure under different load conditions, thus preserving the main physical characteristics and mechanical information of the structural response during the dimensionality reduction process.

[0055] Step 3: Training the neural network structure deformation model.

[0056] After completing the structural deformation feature extraction and dimensionality reduction, the flight condition parameters and aerodynamic load parameters corresponding to step 1 are used as the input of the neural network model, and the structural deformation modal coefficients obtained in step 2 are used as the output to construct a neural network model to learn the mapping relationship between aerodynamic loads and structural deformation features.

[0057] The neural network model employs a multilayer perceptron structure with at least four hidden layers to ensure the network's ability to learn nonlinear mappings. The network structure is as follows: Figure 14 As shown, the model is trained using supervised learning, with the training set comprising 80% of the total samples, and the validation and test sets each comprising 10%. Training ends when the error on the test set is less than 5%, resulting in a neural network surrogate model for structural deformation prediction (i.e., a structural deformation prediction model).

[0058] Step 4: Blade modeling for multi-fidelity aerodynamic analysis.

[0059] Establish the rotor blade model required for multi-fidelity aerodynamic analysis. For aerodynamic models of different precision, construct corresponding geometric or discrete models, including: a three-dimensional blade mesh model for computational fluid dynamics (CFD) solutions, such as... Figure 15 As shown; a vortex lattice model used for the Free Wake Model (FWM) or the Vortex Particle Method (VPM), such as Figure 16 As shown; for the medium-to-high fidelity hybrid method, the blade model required for the corresponding composition method is also established simultaneously.

[0060] Step 5: Solve the aerodynamic structure coupling.

[0061] The structural deformation prediction model trained in step 3 is embedded into the multi-fidelity aerodynamic analysis process. During the aerodynamic solution process, the neural network predicts the rotor structure deformation in real time based on the current working conditions and aerodynamic loads, and updates the blade geometry model accordingly.

[0062] (1) Online inference of the model.

[0063] During each aerodynamic time step (tightly coupled) or each simulation cycle (loosely coupled), the current rotor operating parameters, aerodynamic loads, and motion attitude parameters are obtained from the aerodynamic solver and input into the neural network model. The neural network model then infers the rotor structure deformation online and updates the rotor blade geometry and aerodynamic boundary conditions. After the blades deform, the aerodynamic solution continues to be executed to obtain new aerodynamic loads, which are then used as input for the next neural network inference, thus completing the coupled iterative process between rotor aerodynamic loads and structural deformation.

[0064] (2) Blade deformation strategy.

[0065] Based on the deformation predictions from the neural network, the spatial positions of each section of the blade are updated to reflect flapping and torsional deformation. In the specific implementation, axial deformation can be ignored; only the displacement and rotation within the cross-sectional plane are processed, and the blade surface mesh or vortex grid distribution is updated accordingly. For the spanwise position... The deformation of each grid point on the cross section can be expressed by the following formula.

[0066] .

[0067] In the formula, Here are the coordinates of each node after deformation. Let its coordinates be before deformation. For the direction of the exhibition The rotation matrix at that point, This corresponds to in-plane displacement. Since the axial deformation of the blade is relatively small, it can be ignored. The deformation diagrams of different blade models are shown below. Figure 17 and Figure 18 As shown.

[0068] Step 6: Output the results.

[0069] Output rotor time history data for rotor aerodynamic performance evaluation, structural safety analysis, load prediction under maneuvering flight conditions, and rotor parameter design and optimization. Time history data includes: rotor disk load distribution (e.g., ... Figure 19 As shown), the evolution results of the eddy current field (such as...) Figure 20 As shown), structural deformation response (such as...) Figure 21 (As shown).

[0070] Compared with the prior art, the above embodiments have at least the following advantages.

[0071] 1) Efficiently incorporate structural deformation effects in multifidelity aerodynamic analysis.

[0072] By introducing a structural deformation proxy model based on POD-neural network, the rotor deformation state corresponding to aerodynamic load can be obtained without explicitly solving the structural dynamic equations during multi-fidelity aerodynamic calculations, thus achieving efficient coupling between aerodynamic and structural effects.

[0073] This effect comes from the POD reduced-order representation in step 2 and the neural network mapping model in step 3.

[0074] 2) Improve the accuracy and stability of aerodynamic prediction while ensuring computational efficiency.

[0075] Neural networks can provide structural deformation inputs uniformly under aerodynamic models of different fidelity (such as CFD, VPM, and FWM), enabling rotor geometry and aerodynamic boundary conditions to be updated self-consistently with loads. This avoids systematic errors caused by the rigid rotor assumption and improves the numerical stability and engineering reliability of multifidelity aerodynamic solutions.

[0076] This effect stems from the model embedding and closed-loop coupling mechanism in step 5.

[0077] 3) Preserve the structural physical characteristics to improve the interpretability and generalization ability of the model.

[0078] Since the neural network learns the POD mode coefficients rather than directly learning the high-dimensional deformation field, and the POD mode itself reflects the dominant deformation mode of the rotor structure, this method reduces the learning difficulty and improves the generalization ability to different working conditions while maintaining physical interpretability, making it more engineering-usable than a completely black-box neural network.

[0079] This effect stems from the co-design of POD reduction modeling and neural network prediction (steps 2 and 3).

[0080] Without deviating from the overall technical concept of this application, there are several alternative technical solutions for the problem of coupled analysis of unsteady flow field and structural deformation of rotor.

[0081] In terms of structural deformation modeling, the feature processing methods used to reduce the dimensionality of high-dimensional structural deformation data are not limited to the intrinsic orthogonal decomposition used in the above embodiments of this application. Autoencoders (AE), convolutional neural networks (CNN) or other feature extraction methods can also be used to represent the structural deformation field.

[0082] Based on the same inventive concept, this application also provides a rotor aerodynamic structure coupling analysis device for implementing the rotor aerodynamic structure coupling analysis method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations of one or more rotor aerodynamic structure coupling analysis device embodiments provided below can be found in the limitations of the rotor aerodynamic structure coupling analysis method described above, and will not be repeated here.

[0083] In one exemplary embodiment, such as Figure 22 As shown, a rotor aerodynamic structure coupling analysis device is provided, comprising: The parameter acquisition module is used to acquire the rotor operating parameters and model parameters of the target rotor at the current aerodynamic time step.

[0084] The model building module is used to build rotor blade models with different fidelity.

[0085] The model solving module is used to input the rotor operating parameters and model parameters at the current aerodynamic time step into rotor blade models with different fidelity. The aerodynamic solver is used to solve the rotor blade models with different fidelity to obtain the current aerodynamic load parameters corresponding to different fidelity.

[0086] The structural deformation prediction module is used to input the rotor operating parameters at the current aerodynamic time step and the current aerodynamic load parameters corresponding to different fidelities into the structural deformation prediction model to obtain the rotor structural deformation at different fidelities at the current aerodynamic time step; wherein, the structural deformation prediction model is determined based on the order reduction modeling method and the neural network model.

[0087] The aerodynamic structure coupling analysis module is used to update the spatial position of each section of the target rotor blade based on the rotor structure deformation at different fidelities under the current aerodynamic time step, thereby changing the geometric shape and aerodynamic boundary conditions of the target rotor blade. The changed geometric shape and aerodynamic boundary conditions are used as model parameters under the next aerodynamic time step, until the rotor structure deformation under all aerodynamic time steps is obtained.

[0088] This application implements multi-fidelity aerodynamic structural coupling analysis of rotors based on a neural network model. Specifically, it uses a geometrically accurate beam model to calculate the structural deformation of rotor blades under various operating conditions, constructing a structural deformation sample dataset, and training a neural network model to predict rotor structural deformation based on this dataset. During the rotor aerodynamic analysis, according to the characteristics of blade modeling and load calculation for different fidelity aerodynamic models, the neural network structural deformation model is embedded into the aerodynamic solution process. This allows for real-time prediction of rotor structural deformation without explicitly solving the structural mechanics equations, and updates the rotor geometry and aerodynamic boundary conditions accordingly, thereby achieving coupled analysis of rotor aerodynamic loads and structural deformation effects. This application can significantly reduce overall computational costs while ensuring aerodynamic analysis accuracy and numerical stability, and is suitable for rotor aerodynamic performance evaluation and engineering design analysis under multiple operating conditions.

[0089] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 23As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores rotor structure deformation values ​​at all aerodynamic time steps. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a rotor aerodynamic structure coupling analysis method.

[0090] Those skilled in the art will understand that Figure 23 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer equipment to which the present application is applied. Specific computer equipment may include, for example, [the following is a list of possible additional structures]. Figure 23 The embodiments show more or fewer components, combinations of certain components, or different component arrangements. In one exemplary embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, which the processor executes to implement the steps in the above-described method embodiments.

[0091] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0092] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0093] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0094] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0095] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, etc., and are not limited to these.

[0096] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0097] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for coupled analysis of rotor aerodynamic structures, characterized in that, The rotor aerodynamic structure coupling analysis method includes: Obtain the rotor operating parameters and model parameters of the target rotor at the current aerodynamic time step; Construct rotor blade models with different fidelity; Input the rotor operating parameters and model parameters at the current aerodynamic time step into rotor blade models with different fidelity, and use the aerodynamic solver to solve the rotor blade models with different fidelity to obtain the current aerodynamic load parameters corresponding to different fidelity. The rotor operating parameters at the current aerodynamic time step and the current aerodynamic load parameters corresponding to different fidelities are input into the structural deformation prediction model to obtain the rotor structural deformation at different fidelities at the current aerodynamic time step; wherein, the structural deformation prediction model is determined based on the order reduction modeling method and the neural network model; The spatial positions of each section of the target rotor blade are updated based on the rotor structure deformation at different fidelities under the current aerodynamic time step, thereby changing the geometric shape and aerodynamic boundary conditions of the target rotor blade. The changed geometric shape and aerodynamic boundary conditions are used as model parameters under the next aerodynamic time step, until the rotor structure deformation under all aerodynamic time steps is obtained.

2. The rotor aerodynamic structure coupling analysis method according to claim 1, characterized in that, The method for determining the structural deformation prediction model includes: Construct structural deformation sample data; the structural deformation sample data includes: structural deformation field data of rotor blades used for training under different azimuth angles and different operating parameters; The structural deformation sample data is subjected to feature extraction and dimensionality reduction using a reduced-order modeling method to obtain structural deformation features; the structural deformation features include: rotor structural deformation amount used for training; Different operating parameters and aerodynamic load parameters are used as inputs to the neural network model, and the structural deformation characteristics are used as the outputs of the neural network model. The model is trained using supervised learning, and the trained neural network model is determined as the structural deformation prediction model.

3. The rotor aerodynamic structure coupling analysis method according to claim 2, characterized in that, Constructing structural deformation sample data specifically includes: Obtain the geometric parameters, cross-sectional properties, and constraint methods of the rotor blades used for training; Based on the geometric parameters, the cross-sectional properties, and the constraint method, a structural dynamics model of the rotor blade for training is established using geometrically precise beam theory. The structural dynamics model of the rotor blade used for training was solved under different operating parameters to obtain structural deformation field data under different azimuth angles and different operating parameters, thereby determining the structural deformation sample data.

4. The rotor aerodynamic structure coupling analysis method according to claim 2, characterized in that, The structural deformation sample data are subjected to feature extraction and dimensionality reduction using a reduced-order modeling method to obtain structural deformation features, specifically including: The structural deformation sample data is decomposed using the intrinsic orthogonal decomposition method to obtain multiple orthogonal modes and corresponding structural deformation mode coefficients; The structural deformation modal coefficients corresponding to orthogonal modes with energy contributions greater than a set value are selected to characterize the rotor structure deformation, thus obtaining the structural deformation characteristics.

5. The rotor aerodynamic structure coupling analysis method according to claim 1, characterized in that, Constructing rotor blade models with different fidelity levels, specifically including: A three-dimensional blade mesh model was constructed using computational fluid dynamics to obtain a high-fidelity rotor blade model; A vortex lattice model was constructed using the free wake model or the vortex particle method to obtain a rotor blade model with medium to low fidelity. Computational fluid dynamics and eddy particle methods were used to construct the blade model, resulting in a high-fidelity rotor blade model.

6. The rotor aerodynamic structure coupling analysis method according to claim 1, characterized in that, The neural network model is a multilayer perceptron.

7. The rotor aerodynamic structure coupling analysis method according to claim 1, characterized in that, After obtaining the rotor structure deformation at all aerodynamic time steps, the rotor aerodynamic structure coupling analysis method further includes: The rotor structure deformation at each aerodynamic time step is processed to form time history data and output; the time history data includes the rotor disk load distribution, vortex field evolution results and structural deformation response at each aerodynamic time step.

8. A rotor aerodynamic structure coupling analysis device, characterized in that, The rotor aerodynamic structure coupling analysis device includes: The parameter acquisition module is used to acquire the rotor operating parameters and model parameters of the target rotor at the current aerodynamic time step. The model building module is used to build rotor blade models with different fidelity. The model solving module is used to input the rotor operating parameters and model parameters at the current aerodynamic time step into rotor blade models with different fidelity. The aerodynamic solver is used to solve the rotor blade models with different fidelity to obtain the current aerodynamic load parameters corresponding to different fidelity. The structural deformation prediction module is used to input the rotor operating parameters at the current aerodynamic time step and the current aerodynamic load parameters corresponding to different fidelities into the structural deformation prediction model to obtain the rotor structural deformation at different fidelities at the current aerodynamic time step; wherein, the structural deformation prediction model is determined based on the order reduction modeling method and the neural network model; The aerodynamic structure coupling analysis module is used to update the spatial position of each section of the target rotor blade based on the rotor structure deformation at different fidelities under the current aerodynamic time step, thereby changing the geometric shape and aerodynamic boundary conditions of the target rotor blade. The changed geometric shape and aerodynamic boundary conditions are used as model parameters under the next aerodynamic time step, until the rotor structure deformation under all aerodynamic time steps is obtained.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that the processor executes the computer program to implement the rotor aerodynamic structure coupling analysis method according to any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the rotor aerodynamic structure coupling analysis method according to any one of claims 1-8.

Images