Finite element modeling and defect judgment method for variable-angle fiber placement component

Abstract

The invention discloses a finite element modeling and defect judgment method for a variable-angle fiber placement component. The method comprises the following steps: acquiring layer design data of the variable-angle fiber placement component and embedded defect geometric description data generated through an independent defect positioning algorithm; according to the layering design data, generating a parameterized shell element finite element grid in a component layering plane area; based on the embedded defect geometric description data and the finite element grids, executing a multi-level judgment process, and endowing each unit grid with a defect state attribute; and integrating the geometric information of the finite element mesh and the defect state attribute, and generating a composite material laminate finite element model file which can be directly used for finite element structure analysis and comprises defect characterization. According to the method, the authenticity of a defect judgment result and the engineering credibility can be remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of computer-aided engineering technology, and in particular to a finite element modeling and defect determination method for variable angle wire-laying components. Background Technology

[0002] Due to the continuous variation of fiber layup angle along the plane of the component, variable-angle fiber-layup composite components inevitably produce manufacturing defects such as fiber overlap and gaps, while achieving excellent structural performance. To assess the impact of these defects on the structural mechanical properties, it is usually necessary to incorporate the geometric information of the defects into the finite element analysis model to achieve effective characterization of the defects at the numerical model level.

[0003] In existing technologies, finite element modeling and defect determination methods for variable-angle wire-lay components often involve projecting the defect geometry onto a finite element mesh and then determining whether an element is a defective element based on a fixed threshold or a rule based on the element center position. However, since the shell element finite element mesh itself is discrete, when the defect geometry does not match the element size, the above methods often rely on manual experience to set the determination threshold. This makes it difficult to ensure the consistency of the discrete defect results at the macroscopic geometric scale, which can easily lead to a systematic overestimation or underestimation of the defect area, thereby affecting the accuracy of the finite element analysis model in representing the true defect geometry.

[0004] Furthermore, existing methods typically rely solely on the geometric coverage ratio of defects within an element, failing to adequately consider the influence of fiber orientation variations and local stress characteristics in variable-angle fiber-lay components. When the geometric proportion of a defect is small but it is located in a region of abrupt fiber angle change or stress concentration, traditional methods are prone to misclassifying it as a non-defect element. This results in a lack of reliability in defect assessment at the local mechanical response level, reducing the engineering credibility of the finite element analysis results. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a finite element modeling and defect determination method for variable angle wire-laying components, comprising the following steps:

[0007] Acquire the layup design data of the variable angle wire layup component, as well as the embedded defect geometric description data generated by the independent defect location algorithm;

[0008] Based on the ply design data, a parameterized shell element finite element mesh is generated in the ply planar region of the component;

[0009] Based on the embedded defect geometric description data and the finite element mesh, a multi-level judgment process is executed to assign defect state attributes to each element mesh; the multi-level judgment process includes:

[0010] Calculate the percentage of the area of ​​each cell grid that is covered by the defect geometry.

[0011] Using the area ratio as input, a dynamic judgment threshold is applied for preliminary screening to obtain preliminary defect units; the dynamic judgment threshold is determined by an iterative optimization algorithm, and the optimization objective of the iterative optimization algorithm is to minimize the error between the total area of ​​all meshes judged as defect units and the total defect area represented by the embedded defect geometric description data, and to meet the preset finite element modeling accuracy requirements.

[0012] By integrating the geometric information of the finite element mesh with the defect state attributes, a finite element model file of the composite laminate containing defect characterization is generated that can be directly used for finite element structural analysis.

[0013] As a preferred embodiment of the finite element modeling and defect determination method for the variable-angle wire-laying component described in this invention, the iterative optimization algorithm employs a step-by-step accuracy approximation strategy, including:

[0014] Set an initial threshold and a sequence of precision levels from coarse to fine;

[0015] Starting from the lowest accuracy level, adjust the threshold within the allowable step size of that level, and calculate the error after each adjustment;

[0016] If a threshold that meets the modeling accuracy requirements is found at the current level, the search stops; otherwise, the search continues at the next higher accuracy level until a threshold that meets the requirements is found or the preset highest accuracy level is reached.

[0017] As a preferred embodiment of the finite element modeling and defect determination method for the variable-angle wire-laying component described in this invention, the multi-level determination process further includes a performance-oriented correction step:

[0018] For the initially identified defective units, obtain the fiber angle distribution or preliminary stress field distribution of the local area where they are located, calculated based on the fiber layup path;

[0019] If the area ratio of the defective cell is lower than a higher-order correction threshold, but its local fiber angle abrupt change rate or stress concentration factor exceeds the corresponding threshold, then its final state attribute is corrected to a defective cell.

[0020] The correction threshold is higher than the dynamic determination threshold.

[0021] As a preferred embodiment of the finite element modeling and defect determination method for the variable-angle wire-laying component described in this invention, it further includes assigning material properties different from normal elements to elements endowed with defect state attributes, specifically including:

[0022] For a cell characterizing fiber overlap defects, its layup thickness attribute is set to N times the reference thickness, where N>1;

[0023] For the unit cell characterizing fiber gap defects, its layup thickness is kept as the baseline thickness, but its material constitutive model properties are modified to simulate the material properties of the resin-rich region.

[0024] As a preferred embodiment of the finite element modeling and defect determination method for the variable-angle fiber-laying component described in this invention, the local fiber angle distribution is obtained and used to calculate the mutation rate in the following manner:

[0025] Within the cell grid and its adjacent cells, calculate the fiber angle at the geometric center of each subdivided filament path;

[0026] The distribution of all angles is statistically analyzed, and their standard deviation or the difference between the maximum and minimum values ​​is calculated as the fiber angle mutation rate.

[0027] The subdivided filament path is obtained by adaptively refining the original filament path according to the size of the finite element mesh.

[0028] As a preferred embodiment of the finite element modeling and defect determination method for the variable-angle wire-laying component described in this invention, the preliminary stress field distribution is used to calculate the local stress concentration factor, which is obtained through the following method:

[0029] Based on the parameterized shell element finite element mesh, and ignoring the differences in defect states, all elements are assigned preset reference material properties;

[0030] Under typical loading conditions of the component, a simplified finite element calculation with linear elasticity and small deformation is performed to obtain preliminary element stress data;

[0031] Based on the preliminary element stress data, the stress concentration factor of each element mesh is calculated.

[0032] As a preferred embodiment of the finite element modeling and defect determination method for the variable-angle wire-laying component described in this invention, wherein: when generating the parameterized shell element finite element mesh, an adaptive strategy including a feedback mechanism is adopted, including:

[0033] The initial mesh size distribution is determined based on the fiber curvature variation characteristics in the ply design data and the minimum defect size in the embedded defect geometric description data.

[0034] After performing the multi-level judgment process, record the unit positions where the judgment result is unstable and the unit is near the dynamic judgment threshold or the correction threshold.

[0035] Based on the recorded element location information, the finite element mesh is locally adaptively refined.

[0036] As a preferred embodiment of the finite element modeling and defect determination method for the variable-angle fiber-laying component described in this invention, when generating the finite element model file, the determination method for the main fiber direction angle attribute of the mesh identified as a defective element differs from that of a normal element:

[0037] For a normal cell, the angle of its main fiber direction is taken as the arithmetic mean of the angles of all fiber path segments within the cell.

[0038] For a defective cell, the angle of its main fiber direction is taken as the weighted average of the angles of the normal fiber path segments within the cell, excluding the defect-covered area, with the weight being the effective length of the path segment within the cell.

[0039] As a preferred embodiment of the finite element modeling and defect determination method for the variable-angle wire-laying component described in this invention, the method is executed by an integrated preprocessing system with model self-verification function, and the system further performs the following steps:

[0040] Based on the generated finite element model file, perform a fast, linear, or simplified finite element calculation to extract the global or local response;

[0041] The response is compared with the expected response based on an ideal defect-free model or a historical benchmark model; if the comparison deviation exceeds the preset tolerance, the preset accuracy requirement or the correction threshold is automatically adjusted, and a partial re-judgment and modeling process is triggered.

[0042] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0043] The beneficial effects of this invention are:

[0044] 1. This invention uses the percentage of unit defect area As a unified geometric quantitative indicator for defect judgment, and with dynamic judgment thresholds The determination process is constructed as a total discrete area corresponding to the set of defective elements. Compared with the actual total defect area calculated from the embedded defect geometric description data The objective is to minimize the area deviation between them, while being subject to the tolerance of modeling accuracy. The iterative optimization process of constraints achieves high consistency and high fidelity of the discrete mapping results of embedded defect geometry at the level of parameterized shell element finite element mesh at the macroscopic defect geometry scale without the need for manual experience threshold setting, thus avoiding the systematic distortion problem of defect area caused by traditional fixed threshold or element center criteria.

[0045] 2. This invention is based on the proportion of unit defect area. With dynamic judgment threshold Based on the preliminary geometric determination results, the fiber angle abrupt change rate calculated from the local fiber orientation data of the unit is further introduced. and the stress concentration factor obtained through simplified finite element verification analysis As a performance-oriented decision correction parameter, it makes the unit's Below the dynamic judgment threshold However, when it exhibits significant performance sensitivity in local mechanical response, it can still be identified and corrected as a defective element, thus breaking through the limitation of traditional methods that only rely on geometric coverage ratio for defect judgment. It realizes the effective identification of hidden defect areas with "small geometric proportion but significant mechanical impact", significantly improving the authenticity and engineering credibility of defect judgment results at the level of local mechanical response.

[0046] 3. This invention introduces an adaptive local mesh refinement strategy during the finite element mesh generation stage of the parametric shell element, using unstable elements based on defect determination results as feedback objects. Furthermore, after model generation, it combines the mechanical response comparison results based on rapid finite element verification analysis to adjust the defect determination threshold. The mesh density and modeling accuracy parameters are adjusted in reverse to form a closed-loop modeling process of "defect judgment - result feedback - local densification - re-judgment - response verification - parameter re-optimization". This allows the shell element mesh density, defect judgment parameters and modeling accuracy control quantities to automatically converge based on the mechanical response deviation. While ensuring the accuracy of defect geometric representation and the reliability of local mechanical response, it avoids global over-densification or repeated modeling, thereby achieving an effective balance between computational efficiency and model fidelity (a finite element analysis model with parameterized shell element finite element mesh as the carrier and embedded defect geometry and material property information). Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments 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. Wherein:

[0048] Figure 1 This is a flowchart illustrating the overall process of a finite element modeling and defect determination method for a variable-angle wire-laying component according to the present invention.

[0049] Figure 2 This is a flowchart illustrating the determination of the dynamic judgment threshold in a finite element modeling and defect judgment method for a variable-angle wire-laying component according to the present invention.

[0050] Figure 3 This is a flowchart illustrating the performance-oriented correction process of the finite element modeling and defect determination method for a variable-angle wire-laying component according to the present invention.

[0051] Figure 4 This is a flowchart illustrating the self-verification process of a finite element modeling and defect determination method for a variable-angle wire-laying component according to the present invention. Detailed Implementation

[0052] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0053] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0054] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0055] Secondly, the present invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of the present invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not according to the usual scale. Furthermore, the schematic diagrams are merely examples and should not limit the scope of protection of the present invention. In addition, actual fabrication should include three-dimensional spatial dimensions of length, width, and depth.

[0056] Example 1

[0057] Reference Figure 1-3This is the first embodiment of the present invention, which provides a finite element modeling and defect determination method for a variable-angle wire-laying component, including the following steps:

[0058] S1. Obtain the layup design data of the variable angle wire layup component and the embedded defect geometric description data generated by the independent defect location algorithm.

[0059] The layup design data defines the ideal fiber placement structure of the component, typically derived from computer-aided design (CAD) systems or dedicated fiber placement path planning software. Its core components include:

[0060] Ply planar geometry: the boundary dimensions and shape of a component, such as the length and width of a rectangular ply.

[0061] Reference path parameters: Key parameters describing the variable angle fiber path, typically including: starting point coordinates, starting angle, ending angle, path projection length, shape control parameters (such as the tension coefficient of the Bézier curve), number of curve segments, etc.

[0062] Production parameters include the width of a single filament bundle and the number of filament bundles contained in the ribbon.

[0063] Basic material properties include the nominal thickness of a single layer (without defects), the engineering elastic constant in the fiber direction, and the material properties of the resin matrix.

[0064] It should be noted that each parameter in the selected ply design data directly drives and determines the execution of subsequent key steps: ply plane geometry and reference path parameters are used to generate parametric finite element meshes and calculate element fiber angles; production parameters guide the setting of mesh size to match manufacturing features; and material fundamental properties provide accurate physical property assignment basis for the final model, thereby ensuring data consistency and model fidelity throughout the entire chain from design to analysis.

[0065] The embedded defect geometric description data originates from a separate pre-defect localization process. This process analyzes a given ply design using a dedicated algorithm or module (e.g., kinematic simulation and interference analysis based on ribbon geometry and layup rules), accurately outputting the possible defect locations, types, and planar contour information. This data describes the geometric characteristics of each embedded defect in a structured manner, typically including: defect identifier (ID), defect type (e.g., fiber overlap, fiber gap), defect boundary geometric description, and ply information (identifying which ply the defect belongs to). Among these, the defect boundary geometric description is the most critical information. It is usually represented by ordered polygon vertices, indicating the closed boundary contour of each defect within the ply plane. For example, a rectangular overlap defect might be represented by the coordinates of its four vertices. , , , Describe in order.

[0066] It should be noted that the embedded defect geometric description data provides a precise geometric benchmark for the location, outline and type of defects in the subsequent multi-level judgment process. It is the only quantitative basis for calculating the defect area ratio of the unit and realizing intelligent mapping between defects and finite element meshes.

[0067] It should also be emphasized that the combination of the layup design data and the embedded defect geometric description data together constitutes the data foundation for the subsequent steps to achieve accurate alignment and integration of the "geometric ideal model" and the "physical defect reality" within a unified finite element framework. It is the integrated input source that drives intelligent mapping, optimization judgment, and ultimately generates a high-fidelity simulation model.

[0068] S2. Based on the ply design data, generate parametric shell element finite element meshes in the ply planar region of the component. Specifically, when generating the parametric shell element finite element meshes, an adaptive strategy including a feedback mechanism is adopted, including:

[0069] The initial mesh size distribution is determined based on the fiber curvature variation characteristics in the layup design data and the minimum defect size in the embedded defect geometric description data.

[0070] After performing a multi-level judgment process, record the cell positions where the judgment result is unstable and the cell is near the dynamic judgment threshold or the correction threshold.

[0071] Based on the recorded element location information, the finite element mesh is locally adaptively refined.

[0072] In a preferred embodiment, determining the initial mesh size includes analyzing the polygon boundaries in all embedded defect geometric description data. For each defect, its own feature length is calculated. Based on this, the theoretical target mesh size of the area surrounding the defect is calculated. Offset outward by a predetermined distance from the boundary of the defect. Define the defect-affected area as the defect area, and apply the size within this area. For the portion of the ply plane covered by multiple defect-affected areas, the target mesh size is taken as the size of all areas covering that region. The minimum value in the range; for areas not covered by any defect-affected regions, a preset, relatively large global base size is used. Simultaneously, the radius of curvature in the fiber path is identified. Less than the threshold The high curvature variation region is identified, and for each high curvature variation region, the curve is extended to both sides along its path normal to generate a corresponding curvature influence region. Within this region, a target size is specified. satisfy: Ultimately, the overall target size will be determined. Input a parametric mesh generator (such as the forward propulsion method) to generate an initial mesh with appropriate density.

[0073] Among them, the feature length Let be the length of the diagonal of the smallest bounding rectangle of the defect polygon.

[0074] in, To control the scaling factor of the grid resolution, it is usually set between 0.2 and 0.5 based on engineering experience.

[0075] in, This formula establishes a direct proportional relationship between the physical size of the defect and the size of the computational mesh, ensuring that the sampling density of the mesh for the defect geometry is high enough, thus laying a solid geometric foundation for the subsequent accurate calculation of the defect area ratio of the unit.

[0076] Among them, the preset distance It can be set based on experience, for example = Or 0.5× .

[0077] The defect influence region is a polygonal buffer that completely encloses the original defect and is slightly larger than it. Creating the defect influence region not only ensures the fineness of the defect mesh itself, but also takes into account the region where the stress field or property field around the defect may change gradient, avoiding abrupt changes in mesh size at the defect boundary, which is conducive to generating meshes with better quality and numerical stability.

[0078] in, This is an empirical coefficient, set to 0.1.

[0079] Among them, radius of curvature The threshold is calculated based on the mathematical description of the reference path in the ply design data. These are empirical values ​​pre-set based on process constraints or material properties.

[0080] in, The strategy of minimizing the value enables a single mesh to simultaneously meet multiple fidelity requirements. The mesh generator automatically generates dense elements within the defect-affected region and at high fiber curvature areas, and sparser elements in other regions, based on the spatially varying size field. This results in an initial finite element mesh with a balanced density and reasonable allocation of computational resources. This provides a high-quality discretized computational foundation for subsequent multi-level intelligent judgment processes, clearly characterizing defects and accurately reflecting fiber orientation.

[0081] In another preferred embodiment, the unit with unstable determination results includes: a threshold region unit, referring to the proportion of its defect area. The value, and the currently used dynamic judgment threshold. Or the absolute difference of the performance-oriented correction threshold is less than a preset small tolerance. The term "state-reversed unit" refers to a unit whose final defect status attribute, determined in subsequent performance-oriented correction steps, differs from the initial screening result based on a dynamic judgment threshold. Specifically, it includes units initially classified as "normal" but reclassified as "defective" due to sudden changes in local fiber angles or stress concentration during performance-oriented correction; conversely, it also includes units initially classified as "defective" but reclassified as "normal" due to failing to meet the performance sensitivity threshold.

[0082] For each identified unstable element, the system records at least its unique element identifier (ID), center coordinates, defect area percentage, current defect status, and its "unstable category." This information is stored in a temporary list or log file as input for the next step of mesh encryption.

[0083] In one specific implementation, the local adaptive encryption execution steps are as follows: based on several consecutive potentially inaccurate regions generated by clustering the center coordinates of unstable units, calculate the current feature size of the region. Expand each region outwards by 1-2 times from its boundary. Form a local region to be encrypted; for each local region to be encrypted, specify a new target cell size. The system calls the mesh generator to perform refinement within the region (e.g., quartering quadrilateral elements); finally, the relevant wire-laying path data and defect geometry description data on the original mesh are mapped onto the newly generated local fine mesh using an interpolation method. After completion, the system immediately starts a new round of multi-level judgment process based on the updated mesh, forming a closed-loop optimization of "judgment-feedback-encryption-rejudgment".

[0084] Wherein, the current feature size This represents the average side length of existing grid cells within the potentially inaccurate region.

[0085] Among them, the new target cell size Set as Half of that, which ensures a significant improvement in local resolution.

[0086] It should be noted that the adaptive mesh generation strategy with feedback mechanism proposed in step S2 above dynamically determines and continuously optimizes the mesh size distribution by integrating the dual features of ply fiber curvature and embedded defect size. This aims to provide an initial computational foundation for the subsequent multi-level judgment process that can both accurately characterize defect geometry and accurately capture fiber orientation. Subsequently, based on the "unstable elements" identified in the judgment results, this strategy intelligently drives the adaptive densification and data migration of the local mesh, thereby forming a closed-loop optimization of "judgment-feedback-densification-rejudgment". Ultimately, this ensures that the generated parametric shell element finite element mesh achieves the optimal balance between computational efficiency and model fidelity (a finite element analysis model with parametric shell element finite element mesh as the carrier and embedded defect geometry and material property information). This lays the core foundation for generating composite laminate model files with defect characterization that can be directly used for high-precision finite element analysis.

[0087] S3. Based on the embedded defect geometric description data and finite element mesh, a multi-level decision-making process is executed, assigning defect state attributes to each element mesh. The multi-level decision-making process includes the following core steps:

[0088] S31: Proportion of defect area in the computational cell mesh

[0089] For each element mesh, calculate the intersection area of ​​its geometric region with all embedded defect polygon regions, denoted as . The area of ​​this cell grid is Then its defect area accounts for... Calculated according to the following formula: .

[0090] The calculation of the intersection area can be achieved through a polygon Boolean operation library or a pixelation method.

[0091] This step calculates the defect area ratio, which provides a unified, quantified geometric input benchmark for the entire judgment process. This precisely correlates the geometry of continuous defects with discrete mesh elements.

[0092] S32: Preliminary screening based on dynamically optimized thresholds yields preliminary defective units.

[0093] Set a dynamic judgment threshold The value is determined using an iterative optimization algorithm. The optimization objective is to satisfy all... > The total area of ​​the unit The actual total defect area represented by the embedded defect geometric description data. Error between Minimize the error, and the final error must be less than the preset modeling accuracy tolerance. .

[0094] Among them, modeling accuracy tolerance It can be set according to the analysis requirements, for example This means that the area error must be less than 1%.

[0095] This step involves dynamically determining the threshold. The determination is transformed into an optimization problem with the goal of macroscopic area matching accuracy. This can fundamentally avoid the distortion of the initial defect element marking results in the total area caused by subjective threshold setting, thereby ensuring the high fidelity of the final generated finite element analysis model in the overall defect geometric representation scale.

[0096] S33: Perform performance-oriented correction on the defective units initially screened out in step S32:

[0097] Fiber angle correction: Obtain the fiber angle distribution within the defective cell and its adjacent cells. Calculate the standard deviation of this distribution. If the defect area of ​​this unit accounts for Below a higher correction threshold (For example ), but its Exceeding the angular mutation threshold (For example If the final state attribute is "defective unit", then its final state attribute is corrected to "defective unit".

[0098] Stress concentration correction: Based on a fast, defect-neglecting linear elastic finite element analysis, the equivalent stress field of the local region containing the element is obtained. The stress concentration factor of the element is then calculated. If the defect area of ​​this unit accounts for... < However, its Exceeding the stress concentration threshold (For example If the value is 1.5, then its final state attribute is corrected to "defective unit".

[0099] Among them, the correction threshold Higher than the dynamic judgment threshold This ensures that the corrections only apply to areas with a low geometric percentage but high performance sensitivity.

[0100] In this step, two mechanical performance sensitivity indicators, fiber angle mutation rate and stress concentration factor, are introduced, which elevates defect judgment from a simple "geometric coverage" mindset to a "performance impact" mindset. This enables the identification of "hidden defects" that have a small geometric proportion but a key impact on local performance, greatly improving the local mechanical fidelity of the model.

[0101] Furthermore, an iterative optimization algorithm is used to determine the dynamic judgment threshold. The values ​​include:

[0102] Set initial value and precision level sequences, such as .

[0103] Starting with the first level of precision (0.1), let Increment by step size (e.g., 0.1) using the current precision as the increment. (Right now Take values ​​of 0, 0.1, 0.2, etc. in sequence.

[0104] For each test The value is used as the judgment threshold, and the above multi-level judgment process (i.e., S31-S33) is executed once to obtain a set of defective units, and the total area of ​​all units in the set is calculated. Then, we can calculate its difference from the actual total area of ​​defects. error .

[0105] Within the current accuracy level, find a way to make the error first fall below the modeling accuracy tolerance. of If found, stop optimization and output the value. .

[0106] If no matching requirement is found within the current level. Then, the search with the smallest error at the end of the current level will be selected. The value is used as the initial value for the next level. Adjust the precision step by step, using the new precision (0.01) as the increment. Value; for each adjusted The value is then re-executed through the multi-level judgment process described above, and the corresponding area error is calculated; this continues until a value is found at the new level that satisfies the accuracy requirements. of This process proceeds sequentially to higher precision levels until a value is found that satisfies the requirement. Required The value is determined until the preset highest precision level (such as 0.0001 ten-thousandths) is reached.

[0107] The above-mentioned stepwise precision approximation strategy achieves an intelligent balance between efficiency and accuracy in the optimization process, avoids blind global search, and can quickly converge to a solution that meets engineering requirements.

[0108] Furthermore, the local fiber angle distribution was obtained and used to calculate the mutation rate in the following manner:

[0109] Within the cell grid and its adjacent cells, calculate the fiber angle at the geometric center of each subdivided filament path;

[0110] The distribution of all angles is statistically analyzed, and the standard deviation or the difference between the maximum and minimum values ​​is calculated as the fiber angle mutation rate.

[0111] The subdivided filament path is obtained by adaptively refining the original filament path according to the size of the finite element mesh.

[0112] The above-mentioned adaptive path encryption based on grid size ensures that reliable fiber angle change quantification data can be obtained under any grid density, thus solving the problem of angle information distortion in variable angle layup on discrete grids.

[0113] In one specific implementation, the fiber angle calculation process is as follows:

[0114] To overcome the problem of insufficient path information sampling caused by the finite element mesh element size being larger than the original filament bundle width, the original filament bundle paths in the layup design data are first adaptively densified.

[0115] For the target cell mesh requiring performance correction, the target cell mesh itself and all its directly adjacent cell meshes are defined as the "local analysis region." All subdivided fiber paths are traversed to identify all path segments that traverse this "local analysis region." For each such path segment, the coordinates of its geometric center point are calculated, and its local fiber angle is determined by the tangent direction of the path at that point. The set of fiber angles for all identified path segments constitutes the angle sample set characterizing the local fiber orientation distribution of the target cell.

[0116] In one specific implementation, the calculation of the fiber angle abrupt change rate specifically includes:

[0117] Based on the obtained angle sample set, statistical methods are used to quantify its dispersion, i.e., to calculate the fiber angle mutation rate. In this embodiment, two equivalent calculation methods are provided:

[0118] Standard deviation method: Calculate the standard deviation of the angle sample set. Standard deviation The larger the value, the more drastic the change in local fiber orientation and the higher the mutation rate.

[0119] Range method: Calculate the maximum value in the angle sample set. and minimum value The difference Difference The larger the value, the greater the range of fiber orientation variation and the more obvious the abrupt change.

[0120] The fiber angle mutation rate refers to the aforementioned standard deviation. or range This quantified value will be directly compared with a preset mutation rate threshold, serving as the basis for performance-oriented correction.

[0121] Furthermore, the preliminary stress field distribution is used to calculate the local stress concentration factor, which is obtained in the following way:

[0122] Based on the parameterized shell element finite element mesh generated in step S2, and ignoring the differences in defect states, all elements are assigned preset reference material properties.

[0123] Under typical loading conditions of the component, a simplified finite element calculation with linear elasticity and small deformation is performed to obtain preliminary element stress data;

[0124] Based on the preliminary element stress data, the stress concentration factor of each element mesh is calculated.

[0125] Among them, the reference material property is the orthogonal anisotropic elastic constant, which represents the defect-free layup and is the nominal thickness of a single layer, as defined in the layup design data in step S1.

[0126] The typical loading conditions of a component refer to the load and boundary condition settings that reflect the most important mechanical state of the component in actual operation or testing. For example, for testing its tensile properties, this can be simplified to applying a uniform tensile displacement or load to one end of the component and a fixed constraint to the other end. Those skilled in the art can define these conditions according to their specific analytical objectives.

[0127] In one specific implementation, performing a simplified finite element calculation for linear elasticity and small deformation includes: under typical loading conditions of the component, calling a general finite element solver to obtain a linear stress distribution field of the component in an ideal defect-free state, as preliminary stress data for subsequent calculations.

[0128] In one specific implementation, calculating the stress concentration factor for each element mesh includes: after completing the finite element calculation, extracting the stress data for each element mesh from the solution results, and based on this, calculating the stress concentration factor for each element in the following manner. :

[0129] Choose an equivalent stress (such as Von Mises stress or maximum principal stress) that can comprehensively reflect the stress state of the element, and denote it as the local stress of the element. .

[0130] Determine a reference stress The reference stress The element average stress can be selected as the stress of a uniform stress region far away from any boundary or defect.

[0131] According to the formula Calculate the stress concentration factor of this element.

[0132] It should be noted that the multi-level judgment process proposed in step S3 is based on the proportion of unit defect area. As a unified geometric input benchmark, and to dynamically determine the threshold The determination of the target area is transformed into an optimization problem with the goal of macroscopic area matching accuracy, which fundamentally avoids model distortion caused by subjective threshold setting and ensures the overall geometric fidelity of defect characterization. More importantly, this method introduces fiber angle abrupt change rate and stress concentration factor. As a criterion for performance-oriented correction, defect identification has been elevated from the traditional "geometric coverage" mindset to a "performance impact" mindset, thereby enabling precise identification of the percentage of defect areas. Below the dynamic judgment threshold However, in the "hidden defect" region, which has a significant impact on local mechanical properties, high-fidelity results were achieved for embedded defects in variable-angle wire-lay components in both geometric and mechanical dimensions. This provides a crucial defect state attribute foundation for the subsequent generation of composite laminate model files that can be directly used for high-precision finite element analysis.

[0133] S4. Integrate the geometric information and defect state attributes of the finite element mesh to generate a finite element model file of the composite laminate containing defect characterization that can be directly used for finite element structural analysis.

[0134] Specifically, assigning material properties that differ from normal elements to elements that are given defect state attributes includes:

[0135] For a unit characterizing fiber overlap defects, its layup thickness attribute is set to N times the reference thickness, where N>1, and the reference thickness is the nominal thickness of a single layer in step S1.

[0136] For the unit cell characterizing fiber gap defects, its layup thickness is kept as the baseline thickness, but its material constitutive model properties are modified to simulate the material properties of the resin-rich region.

[0137] The material parameters (such as elastic modulus and Poisson's ratio) of the resin enrichment region are predefined in the basic material properties of step S1.

[0138] In a preferred embodiment, N is set to 2, i.e., the thickness is set to twice the nominal thickness of a single layer, to simulate the physical fact of double filament overlap. Its material constitutive model still uses a normal composite material layup model.

[0139] Specifically, when generating finite element model files, the determination method for the principal fiber direction angle attribute of meshes identified as defective elements differs from that of normal elements:

[0140] For a normal cell, the main fiber direction angle is taken as the arithmetic mean of the angles of all fiber path segments within the cell. Specifically, all fiber path segments passing through the cell are extracted, the average angle of each path segment within its cell is calculated, and then the arithmetic mean of the average angles of all path segments is taken as the main fiber direction angle of the normal cell.

[0141] For defective cells, the main fiber direction angle is taken as the weighted average of the angles of normal fiber path segments within the cell, excluding the defect-covered area. The weight is the effective length of the path segment within the cell. Specifically, for cells identified as defective, the geometric portion inside that is covered by the defective area (overlap or gap) is identified. Excluding the interference of the covered area, angle calculation is performed only on the remaining normal fiber-covered area within the cell: the average angle of each path segment passing through this normal area is calculated, and a weighted average is performed using the effective length of that path segment within the normal area of ​​the cell as the weight, thus obtaining a main fiber direction angle that more closely reflects physical reality.

[0142] In one specific implementation, the integration and output of the finite element model file includes: after assigning values ​​to the aforementioned material and angle attributes, the system creates or points to corresponding material attribute cards in the model data according to the different defect state attributes of the elements (normal, fiber overlap, fiber gap), and generates a unique material attribute card number. The node coordinates, element connection relationships, material attribute card numbers corresponding to the elements, and the main fiber direction angle of each element are encoded according to the input formats of current CAE software (such as Abaqus, ANSYS, Nastran) (such as .inp, .cdb, .bdf) and written into a structured text file. This file is the aforementioned finite element model file of composite laminate containing defect characterization that can be directly used for finite element structural analysis.

[0143] Example 2, as Figure 4 This is the second embodiment of the present invention. Unlike embodiment 1, the above method is executed by an integrated preprocessing system with model self-verification capabilities. This system also performs the following steps:

[0144] Based on the generated finite element model file, perform a fast, linear, or simplified finite element calculation to extract the global or local response;

[0145] Compare this response with the expected response based on an ideal defect-free model or a historical benchmark model;

[0146] If the comparison deviation exceeds the preset tolerance, the preset accuracy requirement will be automatically adjusted or the threshold will be corrected, and a partial re-judgment and modeling process will be triggered.

[0147] In one specific implementation, performing a fast, linear, or simplified finite element calculation includes: the system reading the finite element model file containing defect characterization generated in step S4, loading it into the finite element solver, and constructing a computational model instance to be verified. The same typical load conditions and boundary conditions defined in step S3 when calculating the stress concentration factor are applied to this instance. Subsequently, a static analysis is performed under the assumptions of linear elasticity and small deformation. This analysis uses the material parameters already assigned in the computational model instance, which include differentiated defect properties.

[0148] In one specific implementation, the response is compared with the expected response based on an ideal defect-free model or a historical benchmark model, including:

[0149] The first step is to obtain the verification response: Based on the results obtained from the aforementioned rapid finite element calculation, one or more quantifiable mechanical response values ​​are extracted and denoted as the verification response. Extractable responses include: the total displacement or average strain of the computational model instance to be verified in the load direction (global response), or the maximum equivalent stress in the critical defect element region identified by the performance-oriented correction step (local response).

[0150] The second step is to determine the expected response: Determine the expected response for comparison using one of the following methods. :

[0151] Ideal Defect-Free Model Response: Construct a new ideal defect-free computational model instance with the exact same mesh, loads, and boundary conditions as the aforementioned computational model instance to be verified. The difference is that all elements in the new model instance are assigned the baseline material properties defined in step S1; perform a linear elastic static analysis on this new model that is exactly the same as the aforementioned verification calculation, and use the resulting corresponding response value as... .

[0152] Historical baseline model response: Verified historical simulation or experimental results that match the current component characteristics and loading conditions are retrieved from a pre-built baseline database and used as the reference. .

[0153] When comparing, calculate the relative deviation between the verification response and the expected response: .

[0154] In one specific implementation, the triggering of a partial re-judgment and modeling process includes:

[0155] The calculated relative deviation With respect to the preset response tolerance Compare them.

[0156] like ≤ If the process ends, the final finite element model file will be output.

[0157] like > If the deviation direction is not met, the system will automatically adjust the judgment parameters and trigger a recalculation.

[0158] like Significantly greater than The system automatically lowers the modeling accuracy tolerance. (e.g. from 0.01) Adjusted to 0.005 Or adjust the performance-oriented correction threshold. This is to make the defect judgment criteria more stringent and reduce the number of defective units.

[0159] like Significantly smaller than The system automatically increases the modeling accuracy tolerance. Or lower the performance-oriented correction threshold This is to make the defect judgment criteria more lenient and increase the number of defective units.

[0160] After parameter adjustment, the system automatically re-executes from step S3 (multi-level judgment process) and reuses the mesh generated in step S2 until it runs to the current verification step again. This closed-loop process can be set to a maximum number of iterations (e.g., 3 times), when the conditions are met... ≤ Or, when the maximum number of iterations is reached, the optimization is terminated and the final model file is output.

[0161] It should be noted that, by introducing an integrated self-verification and feedback closed loop, Example 2 enables the system to automatically perform rapid verification analysis based on the finite element model file containing defect characterization after generating the file, thereby obtaining the verification response. And compare it with the expected response derived from an ideal, defect-free model or historical benchmark. When making a comparison, when the relative deviation Exceeding the preset response tolerance At that time, the system can intelligently reverse the adjustment of modeling accuracy tolerance. Or performance-oriented correction threshold This triggers partial re-judgment and modeling from the multi-level judgment process, thereby upgrading the entire finite element modeling method from an open-loop process to an intelligent closed-loop system with self-diagnosis, parameter self-optimization and iterative convergence capabilities, ultimately ensuring that the output finite element model file meets engineering expectations within the preset response reliability tolerance range.

[0162] In summary, this invention addresses the issue by adjusting the proportion of unit defect area. As a unified geometric quantitative indicator for defect judgment, and with dynamic judgment thresholds The determination process is constructed as a total discrete area corresponding to the set of defective elements. Compared with the actual total defect area calculated from the embedded defect geometric description data The objective is to minimize the area deviation between them, while being subject to the tolerance of modeling accuracy. The iterative optimization process of constraints achieves high consistency and high fidelity at the macroscopic defect geometric scale by realizing the discrete mapping results of embedded defect geometry at the level of parameterized shell element finite element mesh without the need for manual empirical threshold setting. This avoids the systematic distortion problem of defect area caused by traditional methods based on fixed thresholds or element center criteria. This invention, based on the element defect area ratio... With dynamic judgment threshold Based on the preliminary geometric determination results, the fiber angle abrupt change rate calculated from the local fiber orientation data of the unit is further introduced. and the stress concentration factor obtained through simplified finite element verification analysis As a performance-oriented decision correction parameter, it makes the unit's Below the dynamic judgment threshold However, when these elements exhibit significant performance sensitivity in local mechanical responses, they can still be identified and corrected as defective elements. This overcomes the limitations of traditional methods that rely solely on geometric coverage ratios for defect determination, enabling effective identification of latent defect regions with small geometric proportions but significant mechanical impacts. This significantly improves the authenticity and engineering reliability of defect determination results at the local mechanical response level. This invention introduces an adaptive local mesh refinement strategy with unstable elements from defect determination results as feedback objects during the parametric shell element finite element mesh generation stage. Furthermore, after model generation, it combines the mechanical response comparison results based on rapid finite element verification analysis to adjust the defect determination threshold. The mesh density and modeling accuracy parameters are adjusted in reverse to form a closed-loop modeling process of "defect judgment - result feedback - local densification - re-judgment - response verification - parameter re-optimization". This allows the shell element mesh density, defect judgment parameters and modeling accuracy control quantities to automatically converge based on the mechanical response deviation. While ensuring the accuracy of defect geometric representation and the reliability of local mechanical response, it avoids global over-densification or repeated modeling, thereby achieving an effective balance between computational efficiency and model fidelity.

[0163] Example 3 is the third embodiment of the present invention. This embodiment provides a computer-readable storage medium on which a computer program is stored, which, when executed by a processor, implements the steps of the above method.

[0164] It should be noted that the above 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A finite element modeling and defect determination method for variable angle wire-laying components, characterized in that, Includes the following steps: Acquire the layup design data of the variable angle wire layup component, as well as the embedded defect geometric description data generated by the independent defect location algorithm; Based on the ply design data, a parameterized shell element finite element mesh is generated in the ply planar region of the component; Based on the embedded defect geometric description data and the finite element mesh, a multi-level judgment process is executed to assign defect status attributes to each element mesh. The multi-level determination process includes: Calculate the percentage of the area of ​​each cell grid that is covered by the defect geometry. Using the area ratio as input, a dynamic judgment threshold is applied for preliminary screening to obtain preliminary defect units; the dynamic judgment threshold is determined by an iterative optimization algorithm, and the optimization objective of the iterative optimization algorithm is to minimize the error between the total area of ​​all meshes judged as defect units and the total defect area represented by the embedded defect geometric description data, and to meet the preset finite element modeling accuracy requirements. By integrating the geometric information of the finite element mesh with the defect state attributes, a finite element model file of the composite laminate containing defect characterization is generated that can be directly used for finite element structural analysis.

2. The finite element modeling and defect determination method for variable angle wire-laying components as described in claim 1, characterized in that: The iterative optimization algorithm employs a step-by-step accuracy approximation strategy, including: Set an initial threshold and a sequence of precision levels from coarse to fine; Starting from the lowest accuracy level, adjust the threshold within the allowable step size of that level, and calculate the error after each adjustment; If a threshold that meets the modeling accuracy requirements is found at the current level, the search stops; otherwise, the search continues at the next higher accuracy level until a threshold that meets the requirements is found or the preset highest accuracy level is reached.

3. The finite element modeling and defect determination method for variable angle wire-laying components as described in claim 2, characterized in that: The multi-level decision-making process also includes a performance-oriented correction step: For the initially identified defective units, obtain the fiber angle distribution or preliminary stress field distribution of the local area where they are located, calculated based on the fiber layup path; If the area ratio of the defective cell is lower than a higher-order correction threshold, but its local fiber angle abrupt change rate or stress concentration factor exceeds the corresponding threshold, then its final state attribute is corrected to a defective cell. The correction threshold is higher than the dynamic determination threshold.

4. The finite element modeling and defect determination method for variable angle wire-laying components as described in claim 1, characterized in that: This also includes assigning material properties that differ from normal elements to elements that are given defect state attributes, specifically including: For a cell characterizing fiber overlap defects, its layup thickness attribute is set to N times the reference thickness, where N>1; For the unit cell characterizing fiber gap defects, its layup thickness is kept as the baseline thickness, but its material constitutive model properties are modified to simulate the material properties of the resin-rich region.

5. The finite element modeling and defect determination method for variable angle wire-laying components as described in claim 3, characterized in that: The local fiber angle distribution was obtained and used to calculate the mutation rate in the following manner: Within the cell grid and its adjacent cells, calculate the fiber angle at the geometric center of each subdivided filament path; The distribution of all angles is statistically analyzed, and their standard deviation or the difference between the maximum and minimum values ​​is calculated as the fiber angle mutation rate. The subdivided filament path is obtained by adaptively refining the original filament path according to the size of the finite element mesh.

6. The finite element modeling and defect determination method for variable angle wire-laying components as described in claim 3, characterized in that: The preliminary stress field distribution is used to calculate the local stress concentration factor, which is obtained through the following method: Based on the parameterized shell element finite element mesh, and ignoring the differences in defect states, all elements are assigned preset reference material properties; Under typical loading conditions of the component, a simplified finite element calculation with linear elasticity and small deformation is performed to obtain preliminary element stress data; Based on the preliminary element stress data, the stress concentration factor of each element mesh is calculated.

7. The finite element modeling and defect determination method for variable angle wire-laying components as described in claim 1, characterized in that: When generating the parameterized shell element finite element mesh, an adaptive strategy including a feedback mechanism is adopted, including: The initial mesh size distribution is determined based on the fiber curvature variation characteristics in the ply design data and the minimum defect size in the embedded defect geometric description data. After performing the multi-level judgment process, record the unit positions where the judgment result is unstable and the unit is near the dynamic judgment threshold or the correction threshold. Based on the recorded element location information, the finite element mesh is locally adaptively refined.

8. The finite element modeling and defect determination method for variable angle wire-laying components as described in claim 1, characterized in that: When generating the finite element model file, the method for determining the principal fiber direction angle attribute of meshes identified as defective elements differs from that of normal elements: For a normal cell, the angle of its main fiber direction is taken as the arithmetic mean of the angles of all fiber path segments within the cell. For a defective cell, the angle of its main fiber direction is taken as the weighted average of the angles of the normal fiber path segments within the cell, excluding the defect-covered area, with the weight being the effective length of the path segment within the cell.

9. The finite element modeling and defect determination method for variable angle wire-laying components as described in claim 1, characterized in that: The method is executed by an integrated preprocessing system with model self-verification capabilities, which also performs the following steps: Based on the generated finite element model file, perform a fast, linear, or simplified finite element calculation to extract the global or local response; The response is compared with the expected response based on an ideal defect-free model or a historical benchmark model; if the comparison deviation exceeds the preset tolerance, the preset accuracy requirement or the correction threshold is automatically adjusted, and a partial re-judgment and modeling process is triggered.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 9.

Images