Fan blade progressive fatigue damage analysis method and related device

Abstract

The invention discloses a fan blade progressive fatigue damage cracking analysis method and a related device. The method comprises the following steps: establishing a data structure framework of a hierarchical structure of a fan blade finite element cohesion global model; the hierarchical relationship of the data structure framework is node group-node pair-node; applying a preset load, and traversing the finite element cohesion global model of the fan blade based on the hierarchical relationship; performing stress calculation on the traversed node pairs by adopting a Lagrange-cohesion fatigue model, and determining whether the node pairs are in a cracking state or not; if yes, performing fatigue damage process analysis on the node pair; if not, predicting node displacement of the node pair by adopting an agent model or a Lagrange algorithm, judging whether the node pair enters a crack initiation state or not, generating a fatigue damage cracking analysis result of the node pair, and further generating a comprehensive fatigue damage cracking analysis result of the fan blade; therefore, the calculation efficiency of fan blade fatigue damage cracking analysis is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of wind turbine power generation technology, and in particular to a method and related apparatus for analyzing progressive fatigue damage and cracking of wind turbine blades. Background Technology

[0002] Driven by the global energy transition, wind energy, as one of the clean and renewable energy sources with the greatest potential for large-scale development, has become an important guarantee for national energy security and sustainable development. In wind turbines, blades, as key load-bearing and energy conversion components, play a decisive role in the overall efficiency and lifespan of the turbine. However, blades are subjected to complex marine environments and wave and current loads for extended periods during operation. As a large-scale thin-walled composite material structure, blades are highly susceptible to fatigue damage such as intralaminar cracking, fiber breakage, matrix damage, and interfacial debonding under these alternating loads. These damage modes often exhibit coupling relationships, leading to overall stiffness degradation and shortened fatigue life, severely restricting the safety and economy of wind turbine operation.

[0003] Among existing methods for progressive fatigue analysis of wind turbine blades, progressive damage models, based on the damage evolution mechanism of composite materials, can establish the relationship between fatigue crack propagation and residual mechanical properties. However, the inherent cohesion model of progressive damage models has numerical defects due to its dependence on penalty stiffness, affecting the accuracy and stability of progressive fatigue damage analysis of wind turbine blades. The Lagrange multiplier method can eliminate the dependency on penalty stiffness by accurately satisfying the displacement continuity condition through constraint equations, and accurately simulate the crack coupling phenomenon of interlaminar interface cracking, intralaminar crack propagation, and multiple damage modes such as interlaminar / interlaminar cracking. However, the Lagrange multiplier method relies on iteratively solving the constraint equation set, which significantly increases the system's degrees of freedom and computational complexity in large-scale finite element models, making convergence difficult and resulting in low computational efficiency for progressive fatigue damage analysis of wind turbine blades. Summary of the Invention

[0004] This invention provides a method and related apparatus for progressive fatigue damage cracking analysis of wind turbine blades, which solves the technical problem of low computational efficiency in existing progressive fatigue damage analysis of wind turbine blades.

[0005] This invention provides a method for analyzing progressive fatigue damage and cracking of wind turbine blades, the method comprising:

[0006] The design parameters of the wind turbine blades are obtained, and a finite element cohesive global model of the wind turbine blades is constructed. The finite element cohesive global model of the wind turbine blades includes several finite element elements, and each finite element element includes several nodes.

[0007] Establish the coordinate system structure of the finite element cohesive global model of the wind turbine blade; based on the coordinate system structure and the design parameters, establish the data structure framework of the composite structure of the wind turbine blade of the finite element cohesive global model; the hierarchical relationship of the data structure framework is node group-node pair-node;

[0008] A preset load is applied to the finite element cohesive global model of the wind turbine blade, and the finite element cohesive global model of the wind turbine blade is traversed through the hierarchical relationship of the data structure framework.

[0009] The Lagrange-cohesive fatigue model is used to calculate the stress of the traversed node pairs to determine whether the node pair is in a cracked state. If so, the fatigue damage process of the node pair is analyzed. If not, the surrogate model or Lagrange algorithm is used to predict the node displacement of the node pair, determine whether the node pair has entered the crack initiation state, and generate the fatigue damage cracking analysis results of the node pair.

[0010] By combining the fatigue damage and cracking analysis results of all traversed node pairs, a comprehensive fatigue damage and cracking analysis result for the wind turbine blade is generated.

[0011] Furthermore, the step of obtaining the design parameters of the wind turbine blades and constructing a finite element cohesive global model of the wind turbine blades includes:

[0012] Obtain the design parameters of the wind turbine blades, and construct a global finite element model of the wind turbine blades based on the design parameters;

[0013] Insert cohesive elements into the finite element global model of the wind turbine blade to construct a finite element cohesive global model of the wind turbine blade.

[0014] Furthermore, the composite structure of the wind turbine blade includes an intralayer structure and an interlayer structure; the steps of establishing the coordinate system structure of the finite element cohesive global model of the wind turbine blade; and establishing the data structure framework of the composite structure of the finite element cohesive global model of the wind turbine blade based on the coordinate system structure and the design parameters include:

[0015] The spatial structure of the finite element cohesive global model of the wind turbine blade is defined, and a coordinate system structure of the finite element cohesive global model of the wind turbine blade is established based on the spatial structure; the coordinate system structure includes several cells, and each cell has corresponding coordinates;

[0016] Based on the design parameters, the intralayer structure and interlayer structure of the finite element cohesive global model of the wind turbine blade are determined, and the interlayer structure is divided into matched meshes and unmatched meshes.

[0017] For the intralayer structure, after inserting the cohesive unit, a pair of nodes on the common boundary of two adjacent finite element units connected by the cohesive unit are taken as a node pair; based on the coordinate system structure, nodes located at the same coordinates are assigned to the same cell; corresponding node groups are created through each non-empty cell, thereby determining the data structure of the intralayer structure;

[0018] For the interlayer structure, a node group is established along the material boundaries of two adjacent layers in the finite element cohesive global model of the wind turbine blade. If the node group is associated with a matching mesh, nodes with the same coordinates are assigned to the same cell in the node group, and corresponding node pairs are created through each non-empty cell, thereby determining the data structure of the matching mesh of the interlayer structure. If the node group is associated with a non-matching mesh, the association between the master face and the slave face is established based on the node group through a global-local dual search. The nodes of the slave face are projected onto the master face to obtain virtual points. Based on the virtual points within a certain range and the nodes of the slave face, node pairs are determined, thereby determining the data structure of the non-matching mesh of the interlayer structure.

[0019] Furthermore, the step of using the Lagrange-cohesive fatigue model to calculate the stress of the traversed node pairs and determine whether the node pair is in a cracked state includes:

[0020] The Lagrange-cohesive fatigue model is used to calculate the stress of the traversed node pairs and determine the Lagrange normal stress component and Lagrange tangential stress component of the node pair.

[0021] The second stress criterion is used to determine whether the node pair is in a cracked state based on the Lagrange normal stress component and the Lagrange tangential stress component of the node pair.

[0022] The secondary stress criterion is expressed as follows:

[0023]

[0024] In the formula: Represents the Lagrangian normal stress components of a node pair. Represents the Lagrange tangential stress components of a node pair; Indicates normal polymerization strength. This indicates the tangential polymerization strength.

[0025] Furthermore, the step of performing fatigue damage process analysis on the node pair includes:

[0026] Determine whether the node has entered a fatigue state;

[0027] If so, update the quasi-static damage variable according to the opening amount of the node pair; update the total damage variable through the quasi-static damage variable, and calculate the cohesion of the quasi-static stage;

[0028] If not, extrapolate the fatigue damage variable of the node pair according to the preset number of cycles, and obtain the cohesive force of the preset number of cycles through skip cycle calculation;

[0029] If the total damage variable is less than 1, it indicates that the node pair is in a state of fatigue damage; otherwise, the node displacement of the node pair is updated according to the cohesion.

[0030] Furthermore, the step of extrapolating the fatigue damage variable of the node pair based on a preset number of cycles and obtaining the cohesive stress value of the preset number of cycles through skip-cycle calculation includes:

[0031] If not, update the fatigue damage variable for that node pair according to the preset number of cycles;

[0032] The jump interval is calculated based on the preset number of cycles and the fatigue damage variable.

[0033] The extrapolated fatigue damage variable is calculated using the cyclic jump method based on the fatigue damage variable and the jump interval.

[0034] The total damage variable is updated based on the extrapolated fatigue damage variable and the quasi-static damage variable, and the cohesion for the preset number of cycles is calculated.

[0035] Furthermore, the step of predicting the nodal displacement of the node pair using a surrogate model or Lagrange algorithm, and determining whether the node pair has entered the crack initiation state, includes:

[0036] If the surrogate model has not been trained, the Lagrange algorithm is started. The Lagrange constraint force of the node pair is predicted iteratively through a prediction-correction iterative strategy, and the node displacement of the node pair is predicted.

[0037] If the surrogate model has been trained, the surrogate model is used to predict the displacement of the node pair, and the node displacement of the node pair is predicted.

[0038] Furthermore, it also includes:

[0039] The displacement difference, load parameters, topology information and damage state of the uncracked node pairs in the finite element cohesive global model of the wind turbine blade are collected as input features, and the node displacements of the corresponding uncracked node pairs are obtained as output labels.

[0040] The proxy model is trained using the input features and the output labels to learn the nonlinear mapping relationship between the input features and the output labels until a preset iteration condition is met, thus completing the training of the proxy model.

[0041] The present invention also provides a computer-readable storage medium having a computer program or instructions stored thereon, wherein the computer program or instructions, when executed by a processor, implement the steps of the progressive fatigue damage cracking analysis method for wind turbine blades as described above.

[0042] The present invention also provides a computer program product, including a computer program or instructions, which, when executed by a processor, implement the steps of the progressive fatigue damage cracking analysis method for wind turbine blades as described above.

[0043] As can be seen from the above technical solutions, the present invention has the following advantages:

[0044] This invention provides a method and related apparatus for analyzing progressive fatigue damage and cracking of wind turbine blades. The method includes: obtaining the design parameters of the wind turbine blade and constructing a finite element cohesive global model of the wind turbine blade; the finite element cohesive global model of the wind turbine blade includes several finite element elements, and each finite element element includes several nodes; establishing the coordinate system structure of the finite element cohesive global model of the wind turbine blade; based on the coordinate system structure and design parameters, establishing a data structure framework for the composite structure of the wind turbine blade of the finite element cohesive global model of the wind turbine blade; the hierarchical relationship of the data structure framework is node group-node pair-node; applying pressure to the finite element cohesive global model of the wind turbine blade. A pre-loaded, hierarchical model of the wind turbine blade's finite element cohesive global model is traversed based on the data structure framework. A Lagrange-cohesive fatigue model is used to calculate the stress of each traversed node pair to determine if it is in a cracked state. If so, a fatigue damage process analysis is performed on the node pair. If not, a surrogate model or Lagrange algorithm is used to predict the node displacement of the node pair to determine if it has entered a crack initiation state, generating the fatigue damage and cracking analysis results for that node pair. Finally, the fatigue damage and cracking analysis results of all traversed node pairs are combined to generate the comprehensive fatigue damage and cracking analysis results for the wind turbine blade.

[0045] In this invention, a data structure framework for the composite structure of wind turbine blades is constructed based on the structural characteristics of wind turbine blades. The hierarchical relationship of "node group-node pair-node" established by the hierarchical data framework supports the complex crack evolution process in the simulation of progressive fatigue damage and cracking of wind turbine blades, which can significantly improve the topology retrieval efficiency of the simulation evolution. At the same time, the Lagrange-cohesive fatigue model of wind turbine blades with fusion surrogate model is used for progressive fatigue damage and cracking analysis, which significantly improves the computational efficiency while ensuring the simulation accuracy, thereby solving the technical problem of low computational efficiency in the existing progressive fatigue damage analysis of wind turbine blades. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 A flowchart illustrating the steps of a method for analyzing progressive fatigue damage and cracking of wind turbine blades, as provided in an embodiment of the present invention.

[0048] Figure 2 (a) is a schematic diagram of a tetrahedral finite element model with cohesive units inserted at potential crack locations provided in an embodiment of the present invention; Figure 2 (b) is a schematic diagram of two finite element units containing one cohesive unit provided in an embodiment of the present invention; Figure 2 (c) is a multi-dimensional display diagram provided by an embodiment of the present invention, which groups nodes with the same coordinates into the same node group at each potential crack location;

[0049] Figure 3 (a) is a front view of the partitioned region of the partitioned computational domain node group provided in an embodiment of the present invention; Figure 3 (b) is a schematic diagram of dividing the computational domain into 9×9×9 cells according to an embodiment of the present invention; Figure 3 (c) is a schematic diagram of a cell-indexed system provided in an embodiment of the present invention; Figure 3 (d) is a schematic diagram of a node group generated by non-empty cell 201 provided in an embodiment of the present invention;

[0050] Figure 4 This is a schematic diagram of a bilinear stress-separation curve provided in an embodiment of the present invention;

[0051] Figure 5 This is a schematic diagram of fatigue cracking degradation provided in an embodiment of the present invention;

[0052] Figure 6 A schematic diagram of a bilinear stress-separation curve considering fatigue cracking is provided for an embodiment of the present invention.

[0053] Figure 7 This is a schematic diagram of the cyclic skipping method provided in an embodiment of the present invention;

[0054] Figure 8 This is a flowchart illustrating the simulation of progressive fatigue damage cracking analysis of wind turbine blades based on the Langron-cohesion model, provided in an embodiment of the present invention. Detailed Implementation

[0055] This invention provides a method and related apparatus for progressive fatigue damage cracking analysis of wind turbine blades, which solves the technical problem of low computational efficiency in existing progressive fatigue damage analysis of wind turbine blades.

[0056] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0057] Please see Figure 1 This invention provides a method for analyzing progressive fatigue damage and cracking of wind turbine blades, the method comprising:

[0058] Step 101: Obtain the design parameters of the wind turbine blades and construct a finite element cohesive global model of the wind turbine blades.

[0059] Understandably, the design parameters of wind turbine blades can include geometric parameters and structural parameters. Geometric parameters include spanwise length distribution, chord length variation, torsion angle distribution characteristics, and airfoil cross-sectional shape, while structural parameters include layered material parameters and material thickness. Geometric parameters can be obtained through analysis using specialized design software (such as Profili).

[0060] This step specifically includes steps S11 to S12:

[0061] S11. Obtain the design parameters of the wind turbine blades and construct a global finite element model of the wind turbine blades based on the design parameters.

[0062] In this embodiment, after obtaining the discrete coordinate data of each section airfoil of the wind turbine blade by professional design software, the data is imported into a three-dimensional modeling platform (such as CAD software) for geometric reconstruction to construct a finite element global model of the wind turbine blade.

[0063] In the modeling process, precise alignment of the reference points of each section is first achieved through coordinate translation and rotation. Then, a three-dimensional surface model of the blade with a continuous and smooth shape is generated using surface lofting technology. For the surface geometry of the generated three-dimensional surface model of the blade, shell elements are used for finite element discretization: structured quadrilateral elements are arranged in the middle region with gentle geometric changes to improve computational efficiency; unstructured triangular elements are used in the tip and complex curvature transition region to ensure local geometric fitting accuracy and numerical stability.

[0064] Next, the composite material layup of the blades was set according to the structural material parameters. The layup angle, thickness, and material system of each region were configured differently according to functional requirements. Different parts of the wind turbine blade were designed according to their respective main functions: the main beam region focused on load-bearing performance, the leading edge region focused on impact resistance, the trailing edge region considered vibration suppression and aerodynamic stability, and the web region considered structural connection and overall stiffness. Through the above modeling, meshing, and material definition process, a high-fidelity finite element global model that can accurately characterize the real structure and mechanical properties of the wind turbine blade was finally established.

[0065] S12. Insert cohesive elements into the finite element global model of the wind turbine blade to construct the finite element cohesive global model of the wind turbine blade.

[0066] It should be noted that the finite element cohesive global model of the wind turbine blade consists of several finite element elements, and each finite element element consists of several nodes. To accurately capture the arbitrary propagation of potential cracks within the wind turbine blade layer, zero-thickness cohesive elements need to be inserted on the common surface of all finite element elements to replicate the finite element nodes, such as... Figure 2 As shown in (a).

[0067] In practical applications, after generating the finite element global model of the wind turbine blade in step S11, a finite element model containing cohesive elements can be established using LS-PrePost software (version 4.10-x64) to achieve numerical simulation of fatigue damage during the entire cracking process of the wind turbine blade. Simultaneously, LS-PrePost software can export the finite element elements, nodes, and cohesive element information of the finite element global model of the wind turbine blade, facilitating unified processing of the subsequently constructed data structure framework.

[0068] Step 102: Establish the coordinate system structure of the finite element cohesive global model of the wind turbine blade; based on the coordinate system structure and design parameters, establish the data structure framework of the composite structure of the wind turbine blade in the finite element cohesive global model of the wind turbine blade.

[0069] It is understandable that the fatigue damage and cracking problem of wind turbine blades includes intralayer fatigue crack propagation and interlayer cracks. To this end, this embodiment establishes a corresponding hierarchical data structure based on the intralayer and interlayer structural characteristics of the composite structure of wind turbine blades using a finite element cohesive global model. The hierarchical relationship of the data structure framework is node group-node pair-node. This structural relationship not only ensures the integrity of the topological relationship but also significantly improves the retrieval efficiency of subsequent data simulation.

[0070] This step specifically includes steps S21 to S25:

[0071] S21. Divide the spatial structure of the finite element cohesive global model of the wind turbine blade, and establish the coordinate system structure of the finite element cohesive global model of the wind turbine blade based on the spatial structure; the coordinate system structure includes several cells, each cell having corresponding coordinates;

[0072] For example, such as Figure 3 As shown, assuming the spatial structure of the finite element cohesive global model of a wind turbine blade is a cube with sides parallel to the coordinate axes, its boundaries are determined by the maximum and minimum values ​​of the node coordinates, it is divided into multiple small cells. The side length of each cell depends on the element feature length (i.e., less than the minimum element length). Assuming nx, ny, and nz are the number of elements along the x, y, and z directions, respectively, the total number of cells in the entire computational domain (the finite element cohesive global model of the wind turbine blade) is nx × ny × nz. Each small cell is assigned an integer identifier, and all nodes with the same coordinates are assigned to the same index cell, thus forming a set of nodes at potential crack locations.

[0073] S22. Based on the design parameters, determine the intra-layer structure and inter-layer structure of the finite element cohesive global model of the wind turbine blade, and divide the inter-layer structure into matched meshes and unmatched meshes.

[0074] It should be noted that the wind turbine blade is a composite material structure, and there are differences in the structural materials within and between the composite material layers. In this embodiment, the data structure within each material layer of the finite element cohesive global model of the wind turbine blade is established as a hierarchical data structure of "node group > node pair > node". For the hierarchical data structure between material layers, it is necessary to determine the matching mesh and the non-matching mesh according to the differences in the layer materials in the design parameters between the layers. That is, for the interlayer structure, the interlayer interface of the same part using the same material is divided into a matching mesh, while the interlayer interface between different materials using different materials is divided into a non-matching mesh.

[0075] S23. For the intra-layer structure, after inserting the cohesive element, a pair of nodes on the common boundary of two adjacent finite element elements connected by the cohesive element is taken as a node pair; based on the coordinate system structure, nodes located at the same coordinates are assigned to the same cell; corresponding node groups are created through each non-empty cell to determine the data structure of the intra-layer structure.

[0076] Specifically, in the intralayer structure, when a cohesive element is inserted on the common boundary between two finite element elements, the cohesive element shares a surface with each of the two adjacent finite element elements; at this time, the three nodes of one surface are connected to the three nodes of the other surface, forming three node pairs, such as... Figure 2 As shown in (b); the intra-layer data structure groups nodes with the same coordinates at potential crack locations into the same node group, such as... Figure 2 As shown in (c).

[0077] For example, Figure 3 (a) shows a front view of the computational domain, where nodes located at the same potential crack location can be grouped into the same cell; the computational domain is divided into 9×9×9 cells, as shown below. Figure 3 As shown in (b); to improve storage and retrieval efficiency, a linked list method is used to store the node identifiers in each cell after domain decomposition. This method uses a one-dimensional array to store data, which helps save memory and improve processing speed; as Figure 3 As shown in (c), the 3D index of cell 201 (highlighted in blue) is (3, 5, 3), corresponding to the 1D index 201. The node group generated from the non-empty cell 201 contains 4 nodes and 4 node pairs, as follows: Figure 3 As shown in (d). Subsequently, all non-empty cells in the computation domain are traversed to create node groups, and these groups are identified using a one-dimensional linked list, ensuring that nodes with the same coordinates belong to only one group. This storage strategy ensures that nodes with the same coordinates belong to only one unique node group, effectively maintaining topological consistency.

[0078] For the intra-layer structure, nodes located at the same coordinates are grouped into the same cell. The corresponding node groups are created by traversing each non-empty cell. The topology information of each node group includes: (a) a list of cells to which the node belongs; (b) a list of node groups to which the node belongs; (c) a list of node pairs to which the node belongs; and (d) a list of nodes that the node has paired with it.

[0079] S24. For interlayer structures, establish node groups along the material boundaries of the two adjacent layers of the finite element cohesive global model of the wind turbine blade, and store all node pairs and nodes involved in the material boundaries in the same node group.

[0080] S25. If the node group is associated with a matching mesh, nodes located at the same coordinates are assigned to the same cell in the node group, and corresponding node pairs are created through each non-empty cell, thereby determining the data structure of the matching mesh of the interlayer structure; if the node group is associated with a non-matching mesh, the association between the master face and the slave face is established based on the node group through a global-local dual search, and the nodes of the slave face are projected onto the master face to obtain virtual points; based on the virtual points within a certain range and the nodes of the slave face, node pairs are determined, thereby determining the data structure of the non-matching mesh of the interlayer structure.

[0081] In this embodiment, for the data structure between layers, a node group is first defined along the material boundaries of the two adjacent layers of the finite element cohesive global model of the wind turbine blade; node pairs are formed in the matched mesh through node-node mapping; in the unmatched mesh, each node group establishes the association between the master / slave surface through global-local dual search, and then establishes node-virtual point type node pairs on the master / slave surface through virtual point projection technology.

[0082] Specifically, a node group is defined along the material boundary of two adjacent layers. This node group establishes the association between the master and slave surfaces through a global-local dual search. The slave surface nodes are projected onto the master surface to obtain virtual points. The virtual points within a certain range are paired with the nodes of the slave surface to determine node pairs. Thus, the master and slave surface numbers and the master and slave node numbers can be obtained.

[0083] For interlayer structures, nodes with the same coordinates are grouped into the same node group. The topology information of each node group includes: (a) master / slave node and master / slave face number index list; (b) geometric association information of nodes and master faces; (c) node pair number index list and node binding relationship.

[0084] The data structure framework provided in this embodiment achieves efficient access to multi-level related information through hierarchical collaborative management, supports fast model reading and efficient indexing, optimizes memory usage and computing resource allocation, provides efficient data support for large-scale node constraint calculation under matched / unmatched meshes, and has excellent access speed and scalability, ensuring data processing efficiency in large-scale numerical simulation of wind turbine blades.

[0085] Step 103: Apply a pre-set load to the finite element cohesive global model of the wind turbine blade, and traverse the finite element cohesive global model of the wind turbine blade based on the hierarchical relationship of the data structure framework.

[0086] It should be noted that applying cyclic preloads to the finite element cohesive global model of wind turbine blades can enable progressive fatigue damage and cracking analysis of wind turbine blades.

[0087] In the node pair looping and retrieval method of this embodiment, the node information stored in the data structure framework can quickly obtain node group information within and between layers; among them, for the intra-layer structure, combined with the stored information of nodes and node pairs, the priority search algorithm can conveniently retrieve adjacent nodes, realizing efficient topology traversal. Figure 3 As shown in (c), in a nonlinear graph data structure, node adjacency relationships are formed within each cell.

[0088] Because node connections are spatially arbitrary, any two nodes may be connected. Therefore, this invention employs a breadth-first search algorithm to perform ordered traversal and numbering of node pairs within the same group. The specific process is as follows: starting with a node pair, it is numbered and recorded as the first node pair in the node group according to the search order. This node pair is then numbered and recorded in the node group list. Subsequently, all unvisited adjacent nodes of the current node are visited sequentially, and these are used as new starting points to continue traversing until all nodes are covered. This method effectively avoids redundant access and repeated calculations, fully traversing all adjacent node pairs of the current node before entering the next level, ensuring complete acquisition of information on all node pairs within the group.

[0089] Step 104: Using the Lagrange-cohesive fatigue model, stress calculations are performed on the traversed node pairs to determine whether the node pair is in a cracked state. If so, fatigue damage process analysis is performed on the node pair. If not, a surrogate model or Lagrange algorithm is used to predict the node displacement of the node pair, determine whether the node pair has entered the crack initiation state, and generate the fatigue damage cracking analysis results of the node pair.

[0090] This invention employs a Lagrange-cohesive fatigue model to simulate interlaminar damage such as delamination and degumming in wind turbine blades, as well as the propagation of interlaminar fatigue cracks and the coupling behavior of intralaminar / interlaminar fatigue damage. The corresponding weak-form governing equations are as follows:

[0091]

[0092] In the formula: In this invention, the research object specifically refers to the finite element cohesive global model of wind turbine blades; For stress tensor; For crack opening displacement; For strain tensor; For volume; The interface is cracked; This represents a binary parameter. A value of 0 indicates that no damage has occurred at the interface, and the displacement continuity is controlled by Lagrange multipliers. A value of 1 indicates that damage has occurred, and the subsequent interface mechanical behavior is controlled by the cohesion model. and These represent the Lagrange multiplier and cohesive stress, respectively. For area; and Represent the material's density and volume force, respectively; acceleration vector. , For partial derivatives, For displacement vectors, For time; To act on the boundary External loads on the surface.

[0093] When simulating progressive fatigue damage and cracking of wind turbine blades, a pre-set load is applied to the cohesive global model of the finite element method of the wind turbine blade. Before cracks occur, Lagrangian constraints are applied to ensure the multi-displacement continuity condition at each potential cracking point for the node group of the intra-layer structure; and to ensure the displacement continuity condition between the node and the main surface node for each potential cracking interface for the node group of the inter-layer structure.

[0094] Furthermore, the process of using the Lagrange-cohesive fatigue model to calculate the stress of the traversed node pairs and determine whether the node pair is in a cracked state includes steps S31~S32:

[0095] S31. Using the Lagrange-cohesive fatigue model, stress calculations are performed on the traversed node pairs to determine the Lagrange normal stress component and Lagrange tangential stress component of the node pair.

[0096] S32. The second stress criterion is used to determine whether the node pair is in a cracked state based on the Lagrange normal stress component and the Lagrange tangential stress component of the node pair.

[0097] In this embodiment, the Lagrangian force applied to the node pair determines which node pair transitions from Lagrangian constraint to cohesive constraint, i.e., from an uncracked state to the initial cracking state. This invention uses a secondary stress criterion to determine crack initiation, thereby determining the state of the node pair.

[0098]

[0099] In the formula: Represents the Lagrangian normal stress components of a node pair. Represents the Lagrange tangential stress components of a node pair; Indicates normal polymerization strength. This indicates the tangential polymerization strength.

[0100] Specifically, if If the calculated result is greater than or equal to 1, it indicates that the stress state of the node pair has reached or exceeded the material's bearing capacity limit, and the node pair is judged to be in a cracked state; otherwise, the node pair is judged to be in an uncracked state.

[0101] Furthermore, for uncracked node pairs, a surrogate model or Lagrange algorithm is used to predict the node displacement of the node pair, and the process of determining whether the node pair has entered the crack initiation state includes steps S41~S44:

[0102] S41. If the surrogate model has not been trained, the Lagrange algorithm is started. The Lagrange constraint force of the node pair is predicted iteratively through the prediction-correction iterative strategy to obtain the node displacement of the node pair.

[0103] In this embodiment, the Lagrangian algorithm employs a "prediction-correction iterative correction" strategy, using the Gauss-Seidel algorithm to iteratively solve the equation system. The iteration continues until two convergence criteria are met: the total error L1 of all node pairs within the node group and the incremental error L2 between iteration steps are both less than their respective tolerance thresholds ε1 and ε2. In the... At time step, the expression for the interface force at the slave node contributed by the Lagrange force is:

[0104]

[0105] In the formula: For the first Lagrange interfacial forces applied to the nodes at time steps; For the shape function of the Lagrange field; Given the total number of nodes on the master face corresponding to a given slave node; and The first The time step is given by the first node on the principal face corresponding to the slave node. The node and its slave node at the 1st The displacement predicted by the time step is:

[0106]

[0107] In the formula: and The first The external and internal forces applied to the nodes at each time step; In the first The Lagrangian force contributed by adjacent nodes at each time step; and They are the first on the main surface The quality of the node and its corresponding slave node; This refers to the time step. It's worth noting that in the mesh matching interface, when a node is directly projected onto a master node on the master face, at this time... Furthermore, for intra-layer node groups, node pairs consist of two matched nodes, and the corresponding Lagrangian force calculation formula also corresponds to the formula in [equation missing]. The situation.

[0108] Furthermore, after predicting the displacement of the node pair using the Lagrange algorithm or surrogate model, the secondary stress criterion can be used to determine whether the node pair has developed cracks.

[0109] S42. If the surrogate model has been trained, the surrogate model is used to predict the displacement of the node pair and the node displacement of the node pair is predicted.

[0110] In this embodiment, considering the potential computational inefficiency of the Lagrange multiplier algorithm during iterative solutions, this invention proposes a machine learning-integrated framework for efficient prediction of crack initiation and propagation. Through deep coupling of the physical model and the data-driven model, computational efficiency is significantly improved while maintaining numerical accuracy. This framework constructs a machine learning proxy model based on accurate data provided by the Lagrange multiplier method in the pre-crack stage, replacing some traditional numerical iterations and significantly reducing computational costs.

[0111] The surrogate model is constructed from a machine learning model consisting of a temporal convolutional network and a random forest classifier. The temporal convolutional network is used to learn the temporal damage evolution characteristics under cyclic loading, and the random forest classifier is used to determine the cracking state of node pairs. The surrogate model provided in this embodiment can replace the iterative solution process of the traditional Lagrange multiplier method, determine whether the node meets the fatigue cracking condition in real time, and determine the crack initiation direction and propagation trend, thereby enhancing the robustness and generalization ability of the node in cracking discrimination.

[0112] Furthermore, if the proxy model has not been fully trained, steps S43-S44 are also included:

[0113] S43. Collect the displacement difference, load parameters, topology information and damage state of the uncracked node pairs of the finite element cohesive global model of the wind turbine blade and use them as input features to obtain the node displacement of the corresponding uncracked node pairs as output labels.

[0114] It should be noted that, in step S41, numerical simulations of uncracked node pairs are performed based on the Lagrange algorithm. Systematic sampling is performed on the numerical calculation results under various load paths and boundary conditions. The displacement difference (opening amount), load parameters, topological information, and damage state (uncracked state) of the node pairs are collected as input features, and the corresponding node displacements are obtained as output labels, thereby constructing the training dataset for the surrogate model. The topological information reflects the spatially related topological structural characteristics.

[0115] S44. Train the surrogate model by input features and output labels to learn the nonlinear mapping relationship between input features and output labels until the preset iteration conditions are met, and complete the training of the surrogate model.

[0116] Understandably, the surrogate model is trained using input features and output labels to learn the nonlinear mapping relationship between the input features and output labels, capture potential damage evolution patterns, and predict the initiation time, propagation direction, and rate of cracks. The preset iteration condition is that the loss function of the surrogate model reaches a preset threshold.

[0117] Furthermore, after the surrogate model training meets the preset iteration conditions, the iteration can be further optimized. That is, after predicting and updating the node pairs of cracks in step S42, the updated results can be fed back to the surrogate model for iterative optimization. Through this closed-loop training-prediction cycle that combines "physics-driven" and "data-driven" approaches, the numerical stability and physical consistency of the Lagrange multiplier method are preserved in the pre-cracking stage, while machine learning significantly improves the simulation efficiency of fatigue crack initiation and propagation stages. This provides a high-precision and high-efficiency solution for fatigue life assessment of multi-crack coupled systems such as wind turbine blades.

[0118] For the nodal pairs at the crack initiation location, the constraint state changes from Lagrangian force to cohesion. At this point, the cohesive stress is in a non-zero region. Directly applying the conventional cohesive stress-separation criterion would lead to jumps in cohesive stress, causing numerical oscillations. To avoid such non-physical phenomena and ensure the continuity of time integration, this embodiment uses an offset bilinear stress-separation curve to characterize the progressive interlaminar damage mechanical behavior of the composite material. Figure 4 As shown.

[0119] Furthermore, the fatigue damage process analysis and simulation process for the cracked node pair includes steps S51~S54:

[0120] S51. Determine whether the node pair has entered a fatigue state;

[0121] In this embodiment, for each node pair in the quasi-static traction-separation curve softening stage, firstly, it is determined whether it is in the calculation step of the fatigue cycle; if so, its maximum energy release rate is calculated. ,in Used to determine whether it satisfies If the condition is met, the node pair is determined to have entered a fatigue state; where, and These are the pre-set energy release thresholds.

[0122] S52. If so, update the quasi-static damage variable according to the opening amount of the node pair; update the total damage variable through the quasi-static damage variable, and calculate the cohesion of the quasi-static stage.

[0123] It should be noted that the Lagrangian stress during the transition phase can be used to... Calculate the initial normal opening when a nodal crack initiates. and initial tangential opening And based on this, the cohesive stress is determined (as in equations (8) and (9)). Among them, It is elastic stiffness. The initial Lagrangian stress is in the normal direction. The initial tangential Lagrangian normal stress is given.

[0124] By establishing a link between the transition criterion and the traction-separation curve, the mechanical contribution of the nodal pairs can smoothly transition from Lagrange multiplier forces to cohesion, thus maintaining the continuity and physical consistency of the mechanical response during the crack initiation stage. The nodal pairs in the crack are constrained by cohesion, and the mechanical behavior of the nodules is described by the stress softening stage of the bilinear stress-separation curve.

[0125] For the quasi-static stage, the effective crack opening at the initiation point Represented as:

[0126]

[0127] In the formula: It is the crack mode mixing coefficient. This represents the current normal opening of the interface.

[0128] The maximum effective crack opening in the quasi-static stage is determined using the power criterion. The formula is as follows; exceeding this value may indicate crack propagation:

[0129]

[0130] In the formula: To stretch the critical energy release rate; The critical shear energy release rate; This represents the damage mode coefficient.

[0131] Update the quasi-static damage variable based on the opening amount of the node pair. :

[0132]

[0133] In the formula: This is the effective opening amount currently being calculated. This represents the largest effective opening in history.

[0134] Determining quasi-static damage variables Then, the total fatigue damage variable in the quasi-static phase can be further determined. ;at this time, .

[0135] The cohesive force in the quasi-static stage includes the normal cohesive force and the tangential cohesive force in the quasi-static stage;

[0136] Quasi-static stage normal cohesion Represented as:

[0137]

[0138] Quasi-static stage tangential cohesion Represented as:

[0139]

[0140] In the formula: normal opening amount Tangential opening .

[0141] S53. If not, extrapolate the fatigue damage variable of the node pair according to the preset number of cycles, and obtain the cohesive force of the preset number of cycles through jump cycle calculation.

[0142] In this step, the fatigue damage variable of the node pair is updated according to the preset number of cycles N. Based on the preset number of cycles N and fatigue damage variables Calculate the jump interval ΔN; use the cyclic jump method based on fatigue damage variables. The extrapolated fatigue damage variable is calculated based on the jump interval ΔN; the total damage variable is updated based on the extrapolated fatigue damage variable and the quasi-static damage variable, and the cohesion for the preset number of cycles is calculated.

[0143] This embodiment employs the following method to perform numerical analysis of fatigue damage in wind turbine blades: Figure 5 The fatigue cracking degradation model is shown. Based on the quasi-static damage model, this model constructs a fatigue damage evolution equation and characterizes the fatigue degradation effect by reducing the tensile and shear strengths of the stress-separation curve, expressed as:

[0144]

[0145]

[0146] In the formula, Indicates total cohesive strength, superscript and These represent time t1 and the initial time, respectively, with the subscripts t and s indicating tensile and shear deformation, respectively. In this embodiment, the total fatigue damage variable at time t1 is represented. From quasi-static damage variables and fatigue damage variables Composition, that is .

[0147] Based on Paris's law, the fatigue damage variable at time t1 This can be expressed as:

[0148]

[0149] in, Indicates the area of ​​the cohesive region; and These represent fatigue cracking and the increase in the number of cycles, respectively. and These represent the initial opening amount of damage and the failure opening amount, respectively; This represents the increment of energy release rate; C and m are parameters of Paris's law.

[0150] To address the issue of excessively long computation times in fatigue numerical simulations, this embodiment employs the cycle jump method to reduce computational costs. By performing calculations on a finite number of load cycles at selected intervals and extrapolating the effects of these cycles on material stiffness degradation, the number of full-cycle calculations required for the simulation is reduced (e.g., ...). Figure 6 As shown), the expression for the jump interval ΔN is:

[0151]

[0152] In the formula, Let J be the fatigue damage variable at the integration point; This represents the maximum permissible damage increment.

[0153] Using a cyclic jumping method, based on Paris's law and the jump interval ΔN, a certain number of jumps are made. The detailed calculation loop directly updates the fatigue damage variable through extrapolation.

[0154] The total damage variable is updated based on the extrapolated fatigue damage variable and quasi-static damage variable. Then, the cohesive strength (including quasi-static) is calculated according to equations (18) and (19). And after fatigue The cohesive force at a preset number of cycles can be calculated, i.e. as well as In the formula, A is the equivalent area of ​​the cohesive unit surface.

[0155] S54. When the total damage variable is greater than or equal to 1, it indicates that the node pair is in a state of fatigue damage; otherwise, the node displacement of the node pair is updated according to the cohesion.

[0156] Understandably, this is achieved through the total damage variable. It can be determined whether the node has completely failed in relation to the material; if If the result is positive, it indicates that the fatigue failure is not complete. In this case, the displacement of the node pair can be updated using the cohesion calculated above. Conversely, if the result is negative, it indicates that the fatigue failure has occurred, and the node pair will no longer participate in the mechanical calculations involved in the fatigue damage cracking analysis method.

[0157] In the process of simulating fatigue damage of wind turbine blades, the fatigue damage cracking analysis results for each node pair (potential crack interface) include node pair status, crack information, and fatigue damage information; among which, the node pair status includes uncracked state, cracked but undamaged state, and fatigued damaged state.

[0158] In the uncracked state, the crack information includes nodal pair displacement and whether cracks have started. The nodal pair displacement records the propagation trend of the initiated crack. In the cracked but undamaged state and the fatigued damaged state, the fatigue damage information includes total damage variables, cohesive stress, cohesion, and nodal pair displacement. The total damage is used to record the degree of damage to the nodal pair, while the cohesive stress and cohesion are used to record the load-bearing capacity of the nodal pair. For the cracked but undamaged state, the nodal pair displacement records the crack trend.

[0159] Step 105: Combine the fatigue damage and cracking analysis results of all traversed node pairs to generate the comprehensive fatigue damage and cracking analysis results of the wind turbine blades.

[0160] In this embodiment, by combining the fatigue damage and cracking analysis results of all node pairs, a comprehensive fatigue damage and cracking analysis result of the entire wind turbine blade can be generated. This achieves efficient and high-precision simulation of the multi-damage evolution process of the blade, and displays the crack distribution, damage area, structural residual strength and predicted fatigue life of the wind turbine blade from a global perspective, providing support for the structural optimization design and performance improvement of the wind turbine blade.

[0161] The present invention provides a method for progressive fatigue damage cracking analysis of wind turbine blades, which significantly improves computational efficiency while strictly ensuring computational accuracy. It realizes efficient and high-precision simulation and life prediction of the multi-damage evolution process of blades, and provides support for its structural optimization design and performance improvement.

[0162] The present invention has the following advantages:

[0163] 1. Based on the intra-layer and inter-layer structural features of wind turbine blades, this invention constructs a data structure framework for the composite structure of wind turbine blades. The hierarchical relationship of "node group-node pair-node" established by the data framework supports the complex crack evolution process in the simulation of progressive fatigue damage cracking of wind turbine blades, which can significantly improve the topology retrieval efficiency of the simulation evolution, thereby improving the computational efficiency of progressive fatigue damage cracking analysis of wind turbine blades.

[0164] 2. The Lagrange multiplier method is adopted to avoid the numerical defects of artificial compliance introduced by the penalty stiffness in the inherent cohesive force model, accurately capture the complex damage evolution behavior of wind turbine blades during service, such as interlayer interface cracking, intralayer crack propagation and multi-crack coupling, and ensure the calculation accuracy of progressive fatigue damage cracking analysis of wind turbine blades.

[0165] 3. The progressive fatigue damage and cracking analysis of wind turbine blades is realized using a Lagrange-cohesive fatigue model with a surrogate model. For uncracked node pairs, the surrogate model is used to replace the Lagrange multiplier iterative algorithm to predict crack initiation and cracking, thus solving the problem of low computational efficiency caused by convergence difficulties in the iterative solution of the Lagrange multiplier method. For the fatigue cracking stage, an offset bilinear stress-separation curve is used to prevent stress jumps when cracks initiate. At the same time, for the problem of excessively long fatigue numerical simulation calculation time, the present invention adopts a cyclic jump method, which can effectively reduce the computational cost.

[0166] The present invention also provides a computer-readable storage medium having a computer program or instructions stored thereon, wherein the computer program or instructions, when executed by a processor, implement the steps of the progressive fatigue damage cracking analysis method for wind turbine blades as described above.

[0167] The present invention also provides a computer program product, including a computer program or instructions, which, when executed by a processor, implement the steps of the progressive fatigue damage cracking analysis method for wind turbine blades as described above.

[0168] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0169] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0170] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0171] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0172] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0173] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for analyzing progressive fatigue damage and cracking of wind turbine blades, characterized in that, The method includes: The design parameters of the wind turbine blades are obtained, and a finite element cohesive global model of the wind turbine blades is constructed. The finite element cohesive global model of the wind turbine blades includes several finite element elements, and each finite element element includes several nodes. Establish the coordinate system structure of the finite element cohesive global model of the wind turbine blade; based on the coordinate system structure and the design parameters, establish the data structure framework of the composite structure of the wind turbine blade of the finite element cohesive global model; the hierarchical relationship of the data structure framework is node group-node pair-node; A preset load is applied to the finite element cohesive global model of the wind turbine blade, and the finite element cohesive global model of the wind turbine blade is traversed based on the hierarchical relationship of the data structure framework. The Lagrange-cohesive fatigue model is used to calculate the stress of the traversed node pairs to determine whether the node pair is in a cracked state. If so, the fatigue damage process of the node pair is analyzed. If not, the surrogate model or Lagrange algorithm is used to predict the node displacement of the node pair, determine whether the node pair has entered the crack initiation state, and generate the fatigue damage cracking analysis results of the node pair. By combining the fatigue damage and cracking analysis results of all traversed node pairs, a comprehensive fatigue damage and cracking analysis result for the wind turbine blade is generated.

2. The method for analyzing progressive fatigue damage and cracking of wind turbine blades according to claim 1, characterized in that, The steps of obtaining the design parameters of the wind turbine blades and constructing a finite element cohesive global model of the wind turbine blades include: Obtain the design parameters of the wind turbine blades, and construct a global finite element model of the wind turbine blades based on the design parameters; Insert cohesive elements into the finite element global model of the wind turbine blade to construct a finite element cohesive global model of the wind turbine blade.

3. The method for analyzing progressive fatigue damage and cracking of wind turbine blades according to claim 2, characterized in that, The wind turbine blade composite structure includes an intralayer structure and an interlayer structure; the steps of establishing the coordinate system structure of the finite element cohesive global model of the wind turbine blade; and establishing the data structure framework of the wind turbine blade composite structure based on the coordinate system structure and the design parameters include: The spatial structure of the finite element cohesive global model of the wind turbine blade is defined, and a coordinate system structure of the finite element cohesive global model of the wind turbine blade is established based on the spatial structure; the coordinate system structure includes several cells, and each cell has corresponding coordinates; Based on the design parameters, the intralayer structure and interlayer structure of the finite element cohesive global model of the wind turbine blade are determined, and the interlayer structure is divided into matched meshes and non-matched meshes. For the intralayer structure, after inserting the cohesive unit, a pair of nodes on the common boundary of two adjacent finite element units connected by the cohesive unit are taken as a node pair; based on the coordinate system structure, nodes located at the same coordinates are assigned to the same cell; corresponding node groups are created through each non-empty cell, thereby determining the data structure of the intralayer structure; For the interlayer structure, a node group is established along the material boundaries of two adjacent layers in the finite element cohesive global model of the wind turbine blade. If the node group is associated with a matching mesh, nodes with the same coordinates are assigned to the same cell in the node group, and corresponding node pairs are created through each non-empty cell, thereby determining the data structure of the matching mesh of the interlayer structure. If the node group is associated with a non-matching mesh, the association between the master face and the slave face is established based on the node group through a global-local dual search. The nodes of the slave face are projected onto the master face to obtain virtual points. Based on the virtual points within a certain range and the nodes of the slave face, node pairs are determined, thereby determining the data structure of the non-matching mesh of the interlayer structure.

4. The method for analyzing progressive fatigue damage and cracking of wind turbine blades according to claim 1, characterized in that, The step of using the Lagrange-cohesive fatigue model to calculate the stress of the traversed node pairs and determine whether the node pair is in a cracked state includes: The Lagrange-cohesive fatigue model is used to calculate the stress of the traversed node pairs and determine the Lagrange normal stress component and Lagrange tangential stress component of the node pair. The second stress criterion is used to determine whether the node pair is in a cracked state based on the Lagrange normal stress component and the Lagrange tangential stress component of the node pair. The secondary stress criterion is expressed as: In the formula: Represents the Lagrangian normal stress components of a node pair. Represents the Lagrange tangential stress components of a node pair; Indicates normal polymerization strength. This indicates the tangential polymerization strength.

5. The method for analyzing progressive fatigue damage and cracking of wind turbine blades according to claim 1, characterized in that, The steps for analyzing the fatigue damage process of the node pair include: Determine whether the node has entered a fatigue state; If so, update the quasi-static damage variable according to the opening amount of the node pair; update the total damage variable through the quasi-static damage variable, and calculate the cohesion of the quasi-static stage; If not, extrapolate the fatigue damage variable of the node pair according to the preset number of cycles, and obtain the cohesive force of the preset number of cycles through skip cycle calculation; If the total damage variable is greater than or equal to the total damage variable, it indicates that the node pair is in a state of fatigue damage; otherwise, the node displacement of the node pair is updated according to the cohesion.

6. The method for analyzing progressive fatigue damage and cracking of wind turbine blades according to claim 5, characterized in that, If not, the step of extrapolating the fatigue damage variable of the node pair based on a preset number of cycles and obtaining the cohesive stress value of the preset number of cycles through skip-cycle calculation includes: If not, update the fatigue damage variable for that node pair according to the preset number of cycles; The jump interval is calculated based on the preset number of cycles and the fatigue damage variable. The extrapolated fatigue damage variable is calculated using the cyclic jump method based on the fatigue damage variable and the jump interval. The total damage variable is updated based on the extrapolated fatigue damage variable and the quasi-static damage variable, and the cohesion for the preset number of cycles is calculated.

7. The method for analyzing progressive fatigue damage and cracking of wind turbine blades according to claim 1, characterized in that, The step of predicting the nodal displacement of the node pair using a surrogate model or Lagrange algorithm and determining whether the node pair has entered the crack initiation state includes: If the surrogate model has not been trained, the Lagrange algorithm is started. The Lagrange constraint force of the node pair is predicted iteratively through a prediction-correction iterative strategy, and the node displacement of the node pair is predicted. If the surrogate model has been trained, the surrogate model is used to predict the displacement of the node pair, and the node displacement of the node pair is predicted.

8. The method for analyzing progressive fatigue damage and cracking of wind turbine blades according to claim 7, characterized in that, Also includes: The displacement difference, load parameters, topology information and damage state of the uncracked node pairs in the finite element cohesive global model of the wind turbine blade are collected as input features, and the node displacements of the corresponding uncracked node pairs are obtained as output labels. The proxy model is trained using the input features and the output labels to learn the nonlinear mapping relationship between the input features and the output labels until a preset iteration condition is met, thus completing the training of the proxy model.

9. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by the processor, they implement the steps of the progressive fatigue damage cracking analysis method for wind turbine blades as described in any one of claims 1-8.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by the processor, they implement the steps of the progressive fatigue damage cracking analysis method for wind turbine blades as described in any one of claims 1-8.

Images