Generation method and device of wind turbine generator part simulation model

Abstract

The invention discloses a wind turbine generator part simulation model generation method and device, and relates to the technical field of simulation, and the method comprises the steps: determining a design variable based on a pre-established basic three-dimensional model; obtaining a target sample space according to the design variable; generating a corresponding model sample based on each sample point in the target sample space; determining a finite element graph carrying a stress value label based on a finite element analysis result of the model sample; generating a corresponding training sample based on the finite element graph; and training the graph neural network model by using the training sample to obtain a target simulation model. According to the method, variables are designed, training samples are generated, it is ensured that training data needed in the model training process is reliable and sufficient, the training samples are generated in combination with a finite element graph, a target simulation model can be different from a traditional neural network model, and the stress field simulation result is directly predicted through geometric input; and the target simulation model is enabled to meet the simulation requirements of the wind turbine generator components.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of simulation, in particular to a wind turbine component simulation model generation method and device. BACKGROUND

[0002] In the development process of modern wind turbines, computer-aided engineering technology has become an important tool for design and optimization. In the structural strength checking of wind turbine components, finite element technology is widely used, forming a complete simulation process. Due to the complexity of the modeling process and the long solution time, the iterative cycle of the design process is prolonged, which becomes a bottleneck for rapid iteration of the model. In recent years, AI (Artificial Intelligence) technology has made significant progress in various fields, especially in data processing and pattern recognition. In the engineering field, artificial intelligence is mostly used in the development of various proxy models to replace conventional finite element simulation and improve iteration speed. Such proxy models are usually designed based on full connection layers to realize the prediction of regression problems from parameterized input to parameterized output.

[0003] However, the proxy model only uses engineering problems that can be parameterized with input and output, and cannot directly predict physical quantities related to spatial position (such as stress contour, strain contour, etc.) from geometric input, so it cannot replace the wind turbine finite element simulation. And for wind turbines, due to the flexible structure deformation mode of the components and the complex load, the training set for training the AI model is huge, which brings additional challenges to the calculation resources and model training. SUMMARY

[0004] The present application provides a wind turbine component simulation model generation method and device, aiming to realize the simulation of wind turbine components.

[0005] In order to achieve the above purpose, the present application provides the following technical solutions:

[0006] A wind turbine component simulation model generation method, comprising:

[0007] Based on the pre-established basic three-dimensional model, determine the design variables;

[0008] According to the design variables, obtain a target sample space; the target sample space includes sample points for representing structural changes and load changes of the basic three-dimensional model;

[0009] Based on each sample point in the target sample space, generate a corresponding model sample; the model sample is a component model after structural change and load change;

[0010] Based on the finite element analysis results of the model sample, a finite element diagram with stress value labels is determined; the stress value labels are used to represent the stress values ​​of each vertex in the finite element diagram;

[0011] Based on the finite element diagram, generating corresponding training samples;

[0012] The graph neural network model is trained using the training samples to obtain a target simulation model; the target simulation model takes the geometric model and global attributes as input, and outputs the stress field simulation results of the geometric model, wherein the stress field simulation results include the stress field distribution of the geometric model, and the stress field distribution includes a cloud map composed of the stress values ​​of each vertex.

[0013] Optionally, the design variables include geometric design variables and load design variables. The design variables are determined based on a pre-established basic three-dimensional model, including:

[0014] Establishing a basic three-dimensional model of a sample wind turbine component using a two-dimensional grid, and determining a structural change area of ​​the sample wind turbine component;

[0015] Using a control body deformation technique in the structural change region, the structural features of the basic three-dimensional model are modified to obtain a part model after the structural change;

[0016] Establishing corresponding geometric design variables based on the deformation mode of the part model after the structural change;

[0017] Based on the load of the part model after the structural change, corresponding load design variables are established.

[0018] Optionally, obtaining a target sample space according to the design variables includes:

[0019] Obtaining geometric design variables and load design variables corresponding to a plurality of part models after the structural changes;

[0020] Generate a first sample point corresponding to each of the geometric design variables using a space filling method;

[0021] For each of the first sample points, based on the stress of the first sample point as a target response, an interaction analysis is performed on each of the geometric design variables in combination with an experimental design method to obtain an interaction level of each of the geometric design variables;

[0022] Recombining the geometric design variables according to the interaction levels of the geometric design variables to obtain a third sample point after the structure is changed;

[0023] Using the space filling method, generating a second sample point corresponding to each of the load design variables;

[0024] For each of the second sample points, based on the stress of the second sample point as the target response, in combination with the experimental design method, a main effect analysis is performed on each of the load design variables to obtain the effect level of each of the load design variables;

[0025] Recombining the load design variables according to their effect levels to obtain a fourth sample point after the load is changed;

[0026] A target sample space is generated based on the third sample point and the fourth sample point.

[0027] Optionally, determining a finite element diagram with stress value labels based on the finite element analysis results of the model sample includes:

[0028] Performing finite element analysis on the model sample to obtain corresponding finite element analysis results; the finite element analysis results include a finite element model and corresponding global properties; the finite element model includes a plurality of nodes, and connection relationships and stress values ​​between each of the nodes;

[0029] constructing an initial finite element graph based on the nodes as vertices, the connection relationships as edges, and the stress values ​​of the nodes as stress value labels of the vertices;

[0030] Adding corresponding edges to vertex pairs in the initial finite element graph that meet preset conditions; the preset conditions are: both vertices in the vertex pair are in a specified mesh encryption area, and the spatial distance between the two vertices meets a preset distance threshold;

[0031] A main vertex corresponding to the global attribute is generated, and the main vertex is connected to all vertices of the initial finite element graph to obtain a finite element graph carrying stress value labels.

[0032] Optionally, generating corresponding training samples based on the finite element diagram includes:

[0033] Determine a vertex attribute tensor corresponding to each vertex in the finite element graph and a global attribute tensor corresponding to the main vertex;

[0034] Combining the vertex attribute tensors and the global attribute tensor to obtain a graph attribute tensor;

[0035] Determining an adjacency matrix based on each edge and corresponding vertex pair in the finite element graph;

[0036] Based on the graph attribute tensor and the adjacency matrix, corresponding training samples are generated.

[0037] Optionally, the graph neural network model adopts an encoder-decoder architecture, which includes multiple network layers, and the multiple network layers include two multi-layer perceptrons and three graph convolution layers.

[0038] Optionally, in the process of training the graph neural network model, the Adam optimizer is used to calculate the mean square error between the predicted node stress value and the node stress value, and solve the gradient direction in which the mean square error decreases the fastest, and update the weights of the graph convolution layer and the fully connected layer in the multi-layer perceptron according to the gradient direction.

[0039] A device for generating simulation models of wind turbine components, comprising:

[0040] A variable determination unit, used for determining design variables based on a pre-established basic three-dimensional model;

[0041] A space determination unit, configured to obtain a target sample space according to the design variables; the target sample space includes sample points for characterizing structural changes and load changes of the basic three-dimensional model;

[0042] A model generation unit, configured to generate corresponding model samples based on each sample point in the target sample space; the model samples are part models after structural changes and load changes;

[0043] an original image determining unit, configured to determine a finite element image carrying a stress value label based on the finite element analysis result of the model sample; the stress value label is used to represent the stress value of each vertex in the finite element image;

[0044] A sample generating unit, configured to generate corresponding training samples based on the finite element diagram;

[0045] A model training unit is used to train the graph neural network model using the training samples to obtain a target simulation model; the target simulation model is based on the geometric model and global attributes as input, and outputs the stress field simulation results of the geometric model, wherein the stress field simulation results include the stress field distribution of the geometric model, and the stress field distribution includes a cloud map composed of the stress values ​​of each vertex.

[0046] A storage medium includes a stored program, wherein the program is executed by a processor to execute the method for generating a simulation model of a wind turbine component.

[0047] An electronic device comprises: a processor, a memory and a bus; the processor and the memory are connected via the bus;

[0048] The memory is used to store programs, and the processor is used to run the programs, wherein the program is executed by the processor to execute the method for generating simulation models of components of a wind turbine generator set.

[0049] The technical solution provided by the present application determines the design variables based on a pre-established basic three-dimensional model. According to the design variables, a target sample space is obtained. Based on each sample point in the target sample space, a corresponding model sample is generated. Based on the finite element analysis results of the model sample, a finite element diagram carrying a stress value label is determined. Based on the finite element diagram, a corresponding training sample is generated. The graph neural network model is trained using the training samples to obtain a target simulation model. The present application uses design variables to generate training samples to ensure that the training data required for the model training process is reliable and sufficient, and combines the finite element diagram to generate training samples, so that the target simulation model can be distinguished from the traditional neural network model, and the stress field simulation results can be directly predicted by the geometric input, so that the target simulation model meets the simulation requirements of wind turbine components. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0051] Figure 1 A schematic flow chart of a method for generating a simulation model of a wind turbine component provided in an embodiment of the present application;

[0052] Figure 2 A flow chart of another method for generating a simulation model of a wind turbine component provided in an embodiment of the present application;

[0053] Figure 3 A flow chart of another method for generating a simulation model of a wind turbine component provided in an embodiment of the present application;

[0054] Figure 4 A flow chart of another method for generating a simulation model of a wind turbine component provided in an embodiment of the present application;

[0055] Figure 5 A flow chart of another method for generating a simulation model of a wind turbine component provided in an embodiment of the present application;

[0056] Figure 6 A schematic diagram of the architecture of a device for generating a simulation model of a wind turbine component provided in an embodiment of the present application;

[0057] Figure 7 A schematic diagram of an encoder-decoder architecture provided in an embodiment of the present application;

[0058] Figure 8 A schematic diagram of a model training method provided in an embodiment of the present application;

[0059] Figure 9 A schematic diagram of a control body provided in an embodiment of the present application;

[0060] Figure 10 A schematic diagram of another control body provided in an embodiment of the present application;

[0061] Figure 11 A schematic diagram of an initial finite element diagram provided in an embodiment of the present application;

[0062] Figure 12 A schematic diagram of a finite element diagram provided in an embodiment of the present application;

[0063] Figure 13 A schematic diagram of obtaining a graph attribute tensor provided in an embodiment of the present application. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0065] In this application, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. The terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or apparatus comprising the element.

[0066] like Figure 1 FIG. 1 is a flow chart of a method for generating a simulation model of a wind turbine component provided in an embodiment of the present application, which includes the following steps.

[0067] S101: Determine design variables based on a pre-established basic three-dimensional model.

[0068] The definition of design variables is the independent parameters that need to be determined in advance during the design process and used to construct model samples.

[0069] Optionally, the design variables include geometric design variables and load design variables. Based on the pre-established basic 3D model, the implementation process of the design variables can be determined. Figure 2 The steps are shown and the corresponding explanations.

[0070] S102: Obtain a target sample space based on the design variables.

[0071] The target sample space includes sample points used to characterize the structural changes and load changes of the basic three-dimensional model.

[0072] Optionally, the implementation process of obtaining the target sample space based on the design variables can be found in Figure 3 The steps are shown and the corresponding explanations.

[0073] S103: Generate corresponding model samples based on each sample point in the target sample space.

[0074] Among them, the model sample is the part model after structural change and load change.

[0075] S104: Based on the finite element analysis results of the model sample, a finite element diagram carrying a stress value label is determined.

[0076] Among them, the stress value label is used to represent the stress value of each vertex in the finite element diagram.

[0077] Optionally, based on the finite element analysis results of the model sample, determine the implementation process of the finite element diagram with stress value labels, which can be seen in Figure 4 The steps are shown and the corresponding explanations.

[0078] S105: Generate corresponding training samples based on the finite element diagram.

[0079] The finite element diagram carries stress value labels, so the generated training samples also carry stress value labels.

[0080] Optionally, based on the finite element diagram, the corresponding training sample generation process can be found in Figure 5 The steps are shown and the corresponding explanations.

[0081] S106: Using the training samples, train the graph neural network model to obtain a target simulation model.

[0082] The target simulation model takes the geometric model and global attributes as input, and outputs the stress field simulation results of the geometric model. The stress field simulation results include the stress field distribution of the geometric model, and the stress field distribution includes a cloud map composed of the stress values ​​of each vertex.

[0083] Optionally, the graph neural network model adopts an encoder-decoder architecture, which includes multiple network layers, and the multiple network layers include two multi-layer perceptrons and three graph convolutional layers.

[0084] Optionally, during the training of the graph neural network model, the Adam optimizer is used to calculate the mean square error between the predicted node stress value and the node stress value, and solve the gradient direction in which the mean square error decreases the fastest. The weights of the graph convolution layer and the fully connected layer in the multi-layer perceptron are updated according to the gradient direction.

[0085] In some examples, the working principle of the Multilayer Perceptron (MLP) mainly includes two processes: forward propagation of signals and back propagation of errors.

[0086] For some examples, see Figure 7 The encoder-decoder architecture shown in the figure consists of a multilayer perceptron, three graph convolutional layers, and a multilayer perceptron. The number of neurons E in each layer and the number of grid nodes n satisfy E > 3*n / 700. An additional dropout layer with a dropout rate of 0.1 is added after each graph convolutional layer. Data normalization is performed before the first multilayer perceptron, using different normalization strategies for different dimensions of the vertex attribute tensor. All layers except the last fully connected layer use the Reinforced Luma (ReLU) activation function.

[0087] In some examples, when starting the graph neural network model, the training and validation sets are randomly split in a ratio of 0.85 / 0.15. The order of the training set is randomly shuffled after each training round, and an early stopping strategy is used. During training, the Adam optimizer is used to calculate the mean squared error between the predicted node stress values ​​and the node stress values ​​to determine the gradient direction with the fastest decrease in this error. The weights of the fully connected layers in the graph convolutional layer and the multilayer perceptron are updated accordingly. The initial learning rate lr0 used to calculate the gradient satisfies lr0 = 0.0005 * bs, where the training batch bs is the number of samples required for each gradient calculation of the graph neural network model. The learning rate uses a fixed step size decay strategy, satisfying the relationship: lrk+1 = lrk * decay_rate. When k%step == 0, step is the decay step size, and the decay step size step satisfies epoch_num / 8 <= step <= epoch_num / 15. k is the current training epoch, the decay rate decay_rate is 0.96, and epoch_num is the total number of training epochs. After the training rounds reach 5000, the weighted model with the best performance on the validation set is selected and saved for subsequent predictions.

[0088] It should be noted that the convolution operation in the graph convolutional neural network is different from the convolution in the traditional convolutional neural network: since the neighbor information of the node is uncertain, the update of the node attributes depends on the adjacency matrix generated according to the finite element mesh, rather than the directional sweep of the convolution kernel. How to construct the adjacency matrix and update the node attributes so that it can learn the intrinsic connection of unstructured data is the difference between the graph neural network model in this embodiment and other neural network models.

[0089] It should be noted that when using a target simulation model, users only need to input the geometric model and global properties (such as boundary conditions and material properties) into the target simulation model. The target simulation model, combined with automated mesh generation technology, converts the geometric model and global properties into graph attribute tensors and adjacency matrices. These are then input into the target simulation model and subjected to stress field prediction, resulting in stress field simulation results. After the prediction, the graph label tensor of the stress field simulation results can be visualized, creating a 3D stress cloud map to facilitate structural strength assessment.

[0090] Combine Figure 2-Figure 5 The method shown in the embodiment of the present application can be summarized as follows: Figure 8As shown in the figure, the target simulation model has the following advantages compared with the conventional finite element model: (1) The stress field prediction speed is fast, and the result of the conventional finite element calculation for 4 hours can be obtained in just tens of seconds, which is convenient for designers to quickly iterate the model in the early stage of development; (2) There is no need for complex modeling, and the stress field results can be obtained by simply inputting the geometry, which reduces the technical threshold for users; (3) It has good accuracy and generalization ability, and can predict the geometry and load of different topological structures with an average error of <10%.

[0091] In some examples, Figure 3-Figure 5 The method shown can be regarded as a training sample generation method, which also has the following advantages: (1) It solves the dependence on historical data when training the model. Even when designing a new product, the training model can be trained based on the basic three-dimensional training set to quickly guide the design; (2) By performing DOE analysis on the design variables, the sample space can be minimized while reflecting the sample changes, thereby improving the training efficiency while ensuring the prediction accuracy.

[0092] The process shown in S101-S106 above utilizes design variables to generate training samples, ensuring that the training data required for the model training process is reliable and sufficient. The training samples are generated in combination with finite element diagrams, so that the target simulation model can be distinguished from the traditional neural network model, and the physical quantities related to the spatial position (i.e., the stress field simulation results) can be directly predicted from the geometric input, so that the target simulation model meets the simulation requirements of the wind turbine components.

[0093] like Figure 2 FIG. 1 is a flow chart of another method for generating a simulation model of a wind turbine component provided in an embodiment of the present application, which includes the following steps.

[0094] S201: Using a two-dimensional grid, a basic three-dimensional model of a sample wind turbine component is established, and a structural change area of ​​the sample wind turbine component is determined.

[0095] The two-dimensional mesh can be in the form of triangular facets. Specifically, the original three-dimensional model of the sample wind turbine components is converted into triangular facets (such as STL format) using structural design software to obtain a basic three-dimensional model, which is then easily deformed. Generally speaking, directly deforming the original three-dimensional model may fail due to interference from different parameters.

[0096] It should be noted that the structural change area can be understood as: clarifying which structural features in the sample wind turbine components need to be changed, such as hole size, wall thickness, chamfer size, etc. Since the deformation mode of such structural features is complex and cannot be directly parameterized, it is necessary to combine control body deformation technology for processing.

[0097] S202: using a control body deformation technique to change the structural features of the basic three-dimensional model in the structural change region to obtain a part model after structural change.

[0098] The control body deformation technique can use free-form surface deformation, which is a geometric modeling technique for deforming a three-dimensional model through a control point grid system. The volume of the part model feature change created by human beings can be regarded as a corresponding control body. The control body is any fixed and invariable volume in space with a certain scale (finite value or infinitesimal).

[0099] S203: establishing a corresponding geometric design variable based on the deformation mode of the part model after structural change.

[0100] The geometric design variable corresponding to the deformation mode of the part model after structural change can be recorded by the three-dimensional modeling software as a design variable in the design environment of the three-dimensional modeling software, as shown in Figure 9 , the shape of the control body deformation hole edge rounding (which can be understood as a deformation body) is used to control the size of the rounding by dragging the anchor point on the deformation body. The deformation can be recorded by the three-dimensional modeling software as a design variable, and the deformation amplitude can be adjusted, as shown in Figure 10 . Similarly, design variables can be created for other structural features by this method, and a plurality of design variables with different values can be generated by an external script, and the design variables with different values are referred to as geometric design variables.

[0101] S204: establishing a corresponding load design variable based on the load of the part model after structural change.

[0102] The load is an external force and other factors that cause internal force and deformation of the structure or component.

[0103] In some examples, the load is applied to the finite element model corresponding to the part model after structural change, and the load of the part model after structural change is obtained, so that the corresponding load design variable is established based on the load of the part model after structural change.

[0104] The above-mentioned S201-S204 process can use the basic three-dimensional model of the sample wind turbine part to obtain corresponding geometric design variables and load design variables, and provide effective data support for subsequent model sample establishment.

[0105] As shown in Figure 3 , another flowchart of a wind turbine part simulation model generation method provided by the embodiment of the application is shown, which includes the following steps.

[0106] S301: Acquire geometric design variables and load design variables corresponding to a plurality of part models after structural changes.

[0107] S302: Generate a first sample point corresponding to each geometric design variable using a space filling method.

[0108] Among them, the essence of the space filling method is a stratified sampling method, such as Latin hypercube sampling. This method can ensure that the sample points are evenly distributed in the multidimensional space, and at the same time set different sampling densities in different dimensions to make the sample points representative.

[0109] In some examples, the first sample points refer to samples with different values ​​of geometric design variables generated using a space filling sampling method. After finite element solution calculations, these samples obtain corresponding stress value labels and are converted into a graph as a training set to facilitate model training.

[0110] It should be noted that by using the space filling method to generate the first sample points corresponding to the geometric design variables, a sample space of geometric composition after structural features have been changed to varying degrees can be obtained, and the sample space is composed of a plurality of first sample points.

[0111] S303: For each first sample point, based on the stress of the first sample point as the target response, combined with the experimental design method, an interaction analysis is performed on each geometric design variable to obtain the interaction level of each geometric design variable.

[0112] Among them, the DOE (DESIGN OF EXPERIMENT) method is a method of systematically planning experimental plans to efficiently determine the relationship between influencing factors and response variables.

[0113] In some examples, DOE methods are used to study whether different levels of variable A affect variable B, with maximum stress as the response. If the maximum stress of the part is independent of the level of B when A is at level 1 or 2, it can be considered that there is no interaction between A and B.

[0114] In some examples, the interaction level can be used to assess the strength of the interaction to achieve classification of geometric design variables. A higher interaction level indicates a stronger interaction, and a lower interaction level indicates a weaker interaction.

[0115] S304: Recombining the geometric design variables according to their interaction levels to obtain a third sample point after the structure is changed.

[0116] Wherein, according to the interaction level of each geometric design variable, each geometric design variable can be classified, for the design variable with weak interaction, the sample set is constructed by adopting the synchronous increase and decrease mode, and for the design variable with strong interaction, the sample set is constructed by adopting the orthogonal design mode, and the sample set comprises the third sample points.

[0117] It should be noted that the interaction refers to the phenomenon that the difference between the reaction quantities of each level of a factor changes with different levels of other factors.

[0118] In some examples, the principle of synchronous increase and decrease is that if the interaction of each design variable is weak, the influence of the combination mode of different levels on the response can be ignored, and only one combination mode is considered, that is, synchronous increase and decrease.

[0119] In some examples, the principle of orthogonal design is that representative points are selected from the experimental design of all possible combination modes of the levels of the geometric design variables for experiment, and the points cover the combination modes of different levels of each set of design variables and can represent the interaction between the geometric design variables.

[0120] In possible embodiments, it is assumed that there are five geometric design variables, A, B, C, D, E, and F, the variable ABC has weak interaction, and the variable DE has strong interaction, under the premise of considering two interaction levels, the final sample points are shown in Table 1.

[0121] Table 1

[0122]

[0123] In the content shown in Table 1 above, level 1 and level 2 both represent the interaction level, and level 2 is higher than level 1, 1, 2, 3, and 4 are the third sample points.

[0124] S305: Using a space-filling method, the second sample points corresponding to each load design variable are generated.

[0125] Wherein, the second sample points refer to samples with different load design variable values generated by using the space-filling sampling method, and the samples are subjected to finite element calculation to obtain corresponding stress value labels and are converted into a graph form as a training set to facilitate model training.

[0126] S306: For each second sample point, based on the stress of the second sample point as the target response, the main effect analysis of each load design variable is performed by combining the experimental design method, to obtain the effect level of each load design variable.

[0127] Wherein, the load effect refers to the internal force or displacement of the structure under the action of a certain load.

[0128] S307: Recombining the load design variables according to their effect levels to obtain a fourth sample point after the load is changed.

[0129] Among them, based on the effect level of each load design variable, the distribution of sample points in the corresponding dimension satisfies 1.1*(N+1)*(N+2)*(C_rate*2), where N is the total number of load design variables and C_rate is the contribution of the load design variable to the response, which can be obtained by solving the eigenvalue of the normalized covariance matrix.

[0130] In some examples, a space-filling method is used to combine N load design variables and multiple geometric design variables to obtain a fourth sample point, or N+1 design variables are used to combine to obtain a fourth sample point.

[0131] S308: Generate a target sample space based on the third sample point and the fourth sample point.

[0132] The third sample point is not additionally encrypted.

[0133] In the process shown in S301 to S308 above, the geometric design variables and the load design variables are combined using the space filling method and the experimental design method to obtain the target sample space.

[0134] like Figure 4 FIG. 1 is a flow chart of another method for generating a simulation model of a wind turbine component provided in an embodiment of the present application, which includes the following steps.

[0135] S401: Perform finite element analysis on the model sample to obtain corresponding finite element analysis results.

[0136] The finite element analysis results include a finite element model and corresponding global properties. The finite element model includes multiple nodes, as well as the connection relationships and stress values ​​between each node.

[0137] In some examples, the finite element model relies on the spatial coordinates and connection relationships of the nodes to predict the stress values ​​of the nodes, and the spatial coordinates of the nodes are labeled as corresponding attributes.

[0138] S402: Constructing an initial finite element diagram based on nodes as vertices, connection relationships as edges, and stress values ​​of nodes as stress value labels of vertices.

[0139] Among them, the initial finite element diagram is a non-Euclidean data with an irregular structure, and the number of neighbors of each basic unit is uncertain. In this embodiment, due to the complexity of spatial structures such as stress cloud maps and boundary conditions, they are also converted into a diagram for processing.

[0140] S403: Adding corresponding edges to vertex pairs that meet preset conditions in the initial finite element graph.

[0141] The preset conditions are as follows: both vertices in the vertex pair are located in a specified mesh encryption area, and the spatial distance between the two vertices meets a preset distance threshold.

[0142] For some examples, see Figure 11 As shown, the designated encryption area is the area composed of 3, 5, 6, 7, and 8. The vertex pairs that meet the preset conditions include 2 and 8, and 5 and 8. Therefore, the edges between 2 and 8 and the edges between 5 and 8 are added.

[0143] It should be noted that adding additional edges to vertex pairs that meet the preset conditions can realize the transmission of node information and improve the prediction accuracy of the graph neural network model for encrypted areas.

[0144] S404: Generate a main vertex corresponding to the global attribute, and connect the main vertex with all vertices of the initial finite element graph to obtain a finite element graph carrying stress value labels.

[0145] Global attributes include, but are not limited to, boundary conditions, material properties, and mesh density. Use a primary vertex to represent global attributes, then connect the primary vertex to all vertices to pass the global attributes to all vertices.

[0146] For some examples, see Figure 12 As shown, M is the main vertex and 1-6 are vertices.

[0147] The process shown in S401 to S404 can generate a finite element diagram with stress value labels based on the finite element analysis results of the model sample.

[0148] like Figure 5 FIG. 1 is a flow chart of another method for generating a simulation model of a wind turbine component provided in an embodiment of the present application, which includes the following steps.

[0149] S501: Determine the vertex attribute tensor corresponding to each vertex in the finite element graph and the global attribute tensor corresponding to the main vertex.

[0150] Among them, the vertex attribute tensor and the global attribute tensor can be determined by filling them into the same dimension using the method of solving the minimum value of the gradient influence.

[0151] S502: Combining the attribute tensors of each vertex and the global attribute tensor to obtain a graph attribute tensor.

[0152] Among them, see Figure 13 As shown in Figure 2, the graph attribute tensor is obtained by combining the attribute tensors of each vertex and the global attribute tensor.

[0153] S503: Determine an adjacency matrix based on each edge and the corresponding vertex pair in the finite element graph.

[0154] Among them, the adjacency matrix is ​​stored in the form of edge vectors, so that the adjacency matrix can be represented in a sparse form, minimizing memory usage and improving the training efficiency of the graph neural network model.

[0155] S504: Generate corresponding training samples based on the graph attribute tensor and the adjacency matrix.

[0156] Among them, the training samples also carry stress value labels.

[0157] The process shown in S501-S504 above can obtain corresponding training samples based on the finite element diagram.

[0158] like Figure 6 , which is a schematic diagram of the architecture of a wind turbine component simulation model generation device provided in an embodiment of the present application, including the units shown below.

[0159] The variable determination unit 100 is used to determine design variables based on a pre-established basic three-dimensional model.

[0160] Optionally, the design variables include geometric design variables and load design variables, and the variable determination unit 100 is specifically used to: use a two-dimensional grid to establish a basic three-dimensional model of the sample wind turbine component, and determine the structural change area of ​​the sample wind turbine component; use the control body deformation technology in the structural change area to change the structural characteristics of the basic three-dimensional model to obtain the part model after the structural change; establish corresponding geometric design variables based on the deformation mode of the part model after the structural change; establish corresponding load design variables based on the load of the part model after the structural change.

[0161] The space determination unit 200 is used to obtain a target sample space according to the design variables; the target sample space includes sample points for characterizing the structural changes and load changes of the basic three-dimensional model.

[0162] Optionally, the space determination unit 200 is specifically used to: obtain geometric design variables and load design variables corresponding to multiple part models after structural changes; generate a first sample point corresponding to each geometric design variable using a space filling method; for each first sample point, based on the stress of the first sample point as the target response, combined with the experimental design method, perform an interaction analysis on each geometric design variable to obtain the interaction level of each geometric design variable; according to the interaction level of each geometric design variable, recombine each geometric design variable to obtain a third sample point after the structural change; generate a second sample point corresponding to each load design variable using a space filling method; for each second sample point, based on the stress of the second sample point as the target response, combined with the experimental design method, perform a main effect analysis on each load design variable to obtain the effect level of each load design variable; according to the effect level of each load design variable, recombine each load design variable to obtain a fourth sample point after the load change; generate a target sample space based on the third sample point and the fourth sample point.

[0163] The model generation unit 300 is used to generate corresponding model samples based on each sample point in the target sample space; the model samples are part models after structural changes and load changes.

[0164] The original image determining unit 400 is used to determine a finite element image with stress value labels based on the finite element analysis results of the model sample; the stress value labels are used to represent the stress values ​​of each vertex in the finite element image.

[0165] Optionally, the original image determination unit 400 is specifically used to: perform finite element analysis on the model sample to obtain corresponding finite element analysis results; the finite element analysis results include a finite element model and corresponding global attributes; the finite element model includes multiple nodes, and the connection relationship and stress value between each node; construct an initial finite element diagram based on the nodes as vertices, the connection relationship as edges, and the stress value of the node as the stress value label of the vertex; add corresponding edges for vertex pairs in the initial finite element diagram that meet preset conditions; the preset conditions are: both vertices in the vertex pair are in the specified mesh encryption area, and the spatial distance between the two vertices meets the preset distance threshold; generate a main vertex corresponding to the global attribute, and connect the main vertex with all vertices of the initial finite element diagram to obtain a finite element diagram with stress value labels.

[0166] The sample generating unit 500 is configured to generate corresponding training samples based on the finite element graph.

[0167] Optionally, the sample generation unit 500 is specifically configured to: determine a vertex attribute tensor corresponding to each vertex in the finite element graph, and a global attribute tensor corresponding to the main vertex; synthesize the vertex attribute tensor and the global attribute tensor to obtain a graph attribute tensor; determine an adjacency matrix based on each edge in the finite element graph and the corresponding vertex pair; and generate a corresponding training sample based on the graph attribute tensor and the adjacency matrix.

[0168] The model training unit 600 is configured to train the graph neural network model using the training sample to obtain a target simulation model; the target simulation model takes the geometric model and the global attribute as input and outputs a stress field simulation result of the geometric model, and the stress field simulation result includes a stress field distribution of the geometric model, and the stress field distribution includes a cloud picture formed by stress values of each vertex.

[0169] Optionally, the graph neural network model adopts an encoder-decoder architecture, and the encoder-decoder architecture includes a plurality of network layers, and the plurality of network layers include two multilayer perceptrons and three graph convolution layers.

[0170] Optionally, the model training unit 600 adopts an Adam optimizer in the process of training the graph neural network model, calculates the mean square error of the predicted node stress value and the node stress value, solves the gradient direction at which the mean square error decreases fastest, and updates the weights of the graph convolution layer and the fully connected layer in the multilayer perceptron according to the gradient direction.

[0171] The above-mentioned various units generate training samples using design variables to ensure that the training data required in the model training process is reliable and sufficient, and generate training samples in combination with finite element graphs, so that the target simulation model can be distinguished from traditional neural network models, directly predict physical quantities related to spatial positions (i.e., stress field simulation results) from geometric inputs, and make the target simulation model meet the simulation requirements of wind turbine components.

[0172] The application also provides a computer-readable storage medium, which includes a stored program, wherein the program executes the wind turbine component simulation model generation method provided by the application.

[0173] The application also provides an electronic device, which includes a processor, a memory and a bus. The processor is connected to the memory through the bus, the memory is used to store a program, and the processor is used to run the program, wherein the program runs to execute the wind turbine component simulation model generation method provided by the application.

[0174] Although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented in multiple embodiments individually or in any suitable sub-combination.

[0175] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the disclosure herein is not limited to technical solutions formed by specific combinations of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned disclosure. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for generating a simulation model of a wind turbine component, characterized in that: include: Determine design variables based on the pre-established basic 3D model; According to the design variables, a target sample space is obtained; The target sample space includes sample points for characterizing structural changes and load changes of the basic three-dimensional model; Based on each sample point in the target sample space, a corresponding model sample is generated; the model sample is a part model after the structure and load are changed; Determining a finite element diagram with stress value labels based on the finite element analysis results of the model sample; The stress value label is used to represent the stress value of each vertex in the finite element diagram; Based on the finite element diagram, generating corresponding training samples; Using the training samples, training the graph neural network model to obtain a target simulation model; The target simulation model takes the geometric model and global properties as input and outputs the stress field simulation result of the geometric model. The stress field simulation result includes the stress field distribution of the geometric model. The stress field distribution includes a cloud map composed of the stress values ​​of each vertex.

2. The method according to claim 1, characterized in that The design variables include geometric design variables and load design variables. Based on the pre-established basic three-dimensional model, the design variables are determined, including: Establishing a basic three-dimensional model of a sample wind turbine component using a two-dimensional grid, and determining a structural change area of ​​the sample wind turbine component; Using a control body deformation technique in the structural change region, the structural features of the basic three-dimensional model are modified to obtain a part model after the structural change; Establishing corresponding geometric design variables based on the deformation mode of the part model after the structural change; Based on the load of the part model after the structural change, corresponding load design variables are established.

3. The method according to claim 2, characterized in that According to the design variables, a target sample space is obtained, including: Obtaining geometric design variables and load design variables corresponding to a plurality of part models after the structural changes; Generate a first sample point corresponding to each of the geometric design variables using a space filling method; For each of the first sample points, based on the stress of the first sample point as a target response, an interaction analysis is performed on each of the geometric design variables in combination with an experimental design method to obtain an interaction level of each of the geometric design variables; Recombining the geometric design variables according to the interaction levels of the geometric design variables to obtain a third sample point after the structure is changed; Using the space filling method, generating a second sample point corresponding to each of the load design variables; For each of the second sample points, based on the stress of the second sample point as the target response, in combination with the experimental design method, a main effect analysis is performed on each of the load design variables to obtain the effect level of each of the load design variables; Recombining the load design variables according to their effect levels to obtain a fourth sample point after the load is changed; A target sample space is generated based on the third sample point and the fourth sample point.

4. The method according to claim 1, wherein Based on the finite element analysis results of the model sample, a finite element diagram with stress value labels is determined, including: Performing finite element analysis on the model sample to obtain corresponding finite element analysis results; the finite element analysis results include a finite element model and corresponding global properties; the finite element model includes a plurality of nodes, and connection relationships and stress values ​​between each of the nodes; constructing an initial finite element graph based on the nodes as vertices, the connection relationships as edges, and the stress values ​​of the nodes as stress value labels of the vertices; Adding corresponding edges to vertex pairs in the initial finite element graph that meet preset conditions; the preset conditions are: both vertices in the vertex pair are in a specified mesh refinement area, and the spatial distance between the two vertices meets a preset distance threshold; A main vertex corresponding to the global attribute is generated, and the main vertex is connected to all vertices of the initial finite element graph to obtain a finite element graph carrying stress value labels.

5. The method according to claim 4, characterized in that Based on the finite element diagram, generating corresponding training samples includes: Determine a vertex attribute tensor corresponding to each vertex in the finite element graph and a global attribute tensor corresponding to the main vertex; Combining the vertex attribute tensors and the global attribute tensor to obtain a graph attribute tensor; Determining an adjacency matrix based on each edge and corresponding vertex pair in the finite element graph; Based on the graph attribute tensor and the adjacency matrix, corresponding training samples are generated.

6. The method according to claim 1, characterized in that The graph neural network model adopts an encoder-decoder architecture, which includes multiple network layers, and the multiple network layers include two multi-layer perceptrons and three graph convolution layers.

7. The method according to claim 6, characterized in that In the process of training the graph neural network model, the Adam optimizer is used to calculate the mean square error between the predicted node stress value and the node stress value, and solve the gradient direction where the mean square error decreases the fastest. The weights of the graph convolution layer and the fully connected layer in the multi-layer perceptron are updated according to the gradient direction.

8. A device for generating simulation models of wind turbine components, characterized in that: include: A variable determination unit, used for determining design variables based on a pre-established basic three-dimensional model; A space determination unit, configured to obtain a target sample space according to the design variables; The target sample space includes sample points for characterizing structural changes and load changes of the basic three-dimensional model; A model generation unit, configured to generate corresponding model samples based on each sample point in the target sample space; the model samples are part models after structural changes and load changes; an original image determining unit, configured to determine a finite element image carrying a stress value label based on a finite element analysis result of the model sample; The stress value label is used to represent the stress value of each vertex in the finite element diagram; A sample generating unit, configured to generate corresponding training samples based on the finite element diagram; A model training unit, configured to train a graph neural network model using the training samples to obtain a target simulation model; The target simulation model takes the geometric model and global properties as input and outputs the stress field simulation result of the geometric model. The stress field simulation result includes the stress field distribution of the geometric model. The stress field distribution includes a cloud map composed of the stress values ​​of each vertex.

9. A storage medium, characterized in that: The storage medium includes a stored program, wherein the program is executed by a processor to execute the method for generating a simulation model of a wind turbine component according to any one of claims 1 to 7.

10. An electronic device, characterized in that: include: processor, memory, and bus; The processor is connected to the memory via the bus; The memory is used to store programs, and the processor is used to run the programs, wherein the program, when run by the processor, executes the method for generating a simulation model of a wind turbine component according to any one of claims 1 to 7.

Images