Target-oriented engineering constant parameter screening method and system for woven composite materials

Abstract

The present application relates to the technical field of woven composite structure design, and provides a woven composite parameter screening method and system for target engineering constants, which comprises: establishing a woven composite multi-scale equivalent performance database, wherein the equivalent performance database contains continuous material parameters, discrete woven configurations and equivalent engineering constants; for different woven configurations, an agent model of continuous material parameters to equivalent engineering constants is established, a target engineering constant constraint is input, and feasible material parameters meeting the constraint are identified in the continuous material parameter space; after the feasible material parameters are sorted, representative screening is performed, and the representative scheme is checked by a forward finite element. The target performance design efficiency is significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of braided composite material structure design technology, and particularly relates to a method and system for screening braided composite material parameters oriented towards target engineering constants. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Braided composite materials, due to their high specific strength, specific stiffness, and good anti-delamination ability, have broad application prospects in aerospace, advanced power, and lightweight load-bearing structures. Compared with traditional laminated composite materials, 2.5D (2.5-dimensional) braided composite materials form a certain interlocking effect in the thickness direction through the splicing of yarns, which can improve the interlayer constraint ability while taking into account manufacturing feasibility, thus having high application value in engineering.

[0004] Existing research on braided composite materials largely focuses on predicting positive performance under given material parameters and configurations. This involves first setting fiber parameters, resin parameters, and braiding configuration, then obtaining the overall equivalent performance through finite element method homogenization or experimental techniques. While this approach can describe the influence of material parameters and structural configuration on the response results, it often requires extensive repeated calculations when designing for target performance, resulting in high computational costs. Furthermore, it struggles to directly answer the question of which fiber parameters and braiding configuration should be chosen to achieve the target engineering constants.

[0005] With the development of surrogate models, parameter identification, and data-driven methods, composite material design has gradually expanded from simple forward prediction to parameter inversion and target performance-driven design. However, for 2.5D braided composite materials, there is a significant coupling relationship between continuous material parameters and discrete braiding configurations, and the distribution characteristics of feasible parameter regions are not consistent under different configurations. If only material parameters are screened without considering the braiding configuration at the same time, it is difficult to obtain optimal results with engineering significance.

[0006] Furthermore, existing methods often only conclude that the objective cannot be met when no fully feasible solution is available, lacking a systematic evaluation of the optimal approximation scheme. Regarding the reliability of the screening results, they also often lack closed-loop verification from objective setting and candidate screening to forward finite element verification. Therefore, it is necessary to propose a reverse screening method that can simultaneously handle the coupled effects of fiber parameters and weaving configuration, and possesses the capabilities of feasible region identification, candidate scheme ranking, approximation scheme output, and forward verification. Summary of the Invention

[0007] To address the technical problems mentioned above, this invention provides a method and system for screening parameters of braided composite materials oriented towards target engineering constants. This method conducts collaborative screening within the joint space of continuous material parameters and discrete braiding configurations, rather than separating the two. Furthermore, it transforms discrete finite element samples into a feasible domain identification problem in a continuous parameter space, thereby significantly improving the efficiency of target performance design.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of the present invention provides a method for screening parameters of braided composite materials oriented towards target engineering constants, comprising: A multi-scale equivalent property database of braided composite materials is established, which includes continuous material parameters, discrete braiding configurations, and equivalent engineering constants. For different weaving configurations, surrogate models from continuous material parameters to equivalent engineering constants are established. The target engineering constant constraints are input, and feasible material parameters that satisfy the constraints are identified in the continuous material parameter space. After sorting the feasible material parameters, representative selection is carried out, and the representative schemes are verified by positive finite element method.

[0009] Furthermore, the continuous material parameters include fiber volume fraction, fiber longitudinal elastic modulus, and fiber transverse elastic modulus; The equivalent engineering constants include the principal elastic modulus, Poisson's ratio, and shear modulus.

[0010] Furthermore, the surrogate model adopts a quadratic response surface surrogate model, for any equivalent engineering constant. The prediction expression is: in, These are the standardized design variables; β0 is a constant term, β i Standardize design variables The coefficient of the first term, β ij The coefficients are the quadratic terms of each design variable. Each coefficient is obtained by fitting discrete sample data under the corresponding weaving configuration and is used to characterize the response relationship of the target engineering constant with the change of design variables. Equivalent engineering constant The predicted values ​​are: i,j=1,2,3.

[0011] Furthermore, for feasible material parameters, weighting coefficients corresponding to each equivalent engineering constant are introduced. Relative ratio to target And a weighted scoring function is used. Sort; where, For the target engineering constant set, For each target value, These are the predicted values ​​of the surrogate model for each equivalent engineering constant.

[0012] Furthermore, when there are no feasible material parameters that fully satisfy the constraints under a certain weaving configuration, a shortage degree is defined. And an approximate scheme scoring function is adopted. The substandard solutions are evaluated to output the optimal approximate solution.

[0013] Furthermore, the representativeness screening includes: for any two candidate points and Define the normalized parameter distance as: ;in, Indicates the first The scale difference of each variable within a preset search range; when Less than the preset threshold When two candidate points are considered as similar schemes in the same local cluster, only the one with the better score is retained.

[0014] Furthermore, the basis for the positive finite element verification is as follows: The relative deviation between the surrogate prediction and the positive check value: ; The deviation of the verification value from the target value is: ; in, For the surrogate model's predicted values, These are the values ​​calculated using the forward finite element method. For the target value, This is the set of target engineering constants.

[0015] A second aspect of the present invention provides a parameter screening system for braided composite materials oriented towards target engineering constants, comprising: The database construction module is configured to: establish a multi-scale equivalent performance database of braided composite materials, wherein the equivalent performance database includes continuous material parameters, discrete braiding configurations, and equivalent engineering constants; The identification module is configured to: establish a proxy model from continuous material parameters to equivalent engineering constants for different weaving configurations, input the target engineering constant constraints, and identify feasible material parameters that satisfy the constraints in the continuous material parameter space; The screening module is configured to sort feasible material parameters, perform representative screening, and perform positive finite element verification on representative schemes.

[0016] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method for screening parameters of braided composite materials oriented towards target engineering constants.

[0017] A fourth aspect of the present invention provides a computer device including a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, wherein the processor executes the program to implement the steps in the above-described method for screening parameters of woven composite materials oriented towards target engineering constants.

[0018] Compared with the prior art, the beneficial effects of the present invention are: This invention enables collaborative screening within the joint space of continuous material parameters and discrete weaving configurations, rather than treating the two separately, and transforms discrete finite element samples into a feasible region identification problem in continuous parameter space, thereby significantly improving the efficiency of target performance design.

[0019] This invention can still output the optimal approximate solution when there is no completely feasible solution, thus improving the engineering applicability of the method.

[0020] This invention constructs a verification closed loop through forward finite element verification, thereby enhancing the reliability and reusability of the screening results. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0022] Figure 1 This is a flowchart of the method for screening parameters of braided composite materials oriented towards target engineering constants according to Embodiment 1 of the present invention; Figure 2 This is a schematic diagram illustrating the hierarchical mapping relationship between the microscopic fiber bundle unit cell, the fine-scale braided unit cell, and the target engineering constant in Embodiment 1 of the present invention. Figure 3 This is a schematic diagram illustrating the identification of the feasible region of parameters and the ranking of candidate solutions under the target engineering constant constraint in Embodiment 1 of the present invention. Figure 4 This is a schematic diagram of the positive finite element verification closed loop of a representative candidate scheme in Embodiment 1 of the present invention; Figure 5 This is a schematic diagram of the structure of a computer device according to Embodiment 4 of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0024] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0025] It should be noted that the following implementation method mainly revolves around a parameter screening method for 2.5D braided composite materials oriented towards target engineering constants. The core of the method is: within the joint space of continuous fiber parameters and discrete braiding configuration, based on a pre-established multi-scale equivalent performance database and configuration discrete surrogate model, the feasible domain identification, candidate scheme ranking, approximate scheme output, and forward finite element verification under the constraint of target engineering constant are realized, thereby obtaining fiber parameter combinations and braiding configuration schemes that meet design requirements.

[0026] Example 1 This embodiment provides a method for screening parameters of braided composite materials oriented towards target engineering constants, in order to solve the problems in the prior art, such as mainly relying on forward trial calculations, low screening efficiency, separation of configuration and material parameters, and lack of approximate schemes and closed-loop verification mechanisms.

[0027] The parameter screening method for braided composite materials oriented towards target engineering constants provided in this embodiment is applied to the target engineering constant-driven reverse screening of 2.5D braided composite materials. Specifically, it involves a collaborative design method for identifying feasible domains, ranking candidate schemes, outputting approximate schemes, and verifying forward finite element analysis within the joint space of continuous fiber parameters and discrete braiding configurations.

[0028] The parameter screening method for braided composite materials oriented towards target engineering constants provided in this embodiment is designed for 2.5D braided composite material design problems with clearly defined mechanical performance requirements. It is particularly suitable for the coordinated screening of fiber volume fraction, fiber longitudinal elastic modulus, fiber transverse elastic modulus, and braid configuration under the condition of given target values ​​of elastic modulus in three principal directions. Unlike traditional methods that rely on iterative forward trial calculations, this embodiment does not re-perform a large number of finite element calculations from scratch after each given target. Instead, it first establishes a multi-scale equivalent performance database with parameter coverage, constructs a predictive model on a continuous parameter space based on this database, and then completes the reverse parameter screening by combining target constraints and ranking criteria.

[0029] In this embodiment, the design variables consist of two parts: continuous material parameters and discrete weaving configuration. The continuous material parameters are defined as follows: ; in, This represents the fiber volume fraction. This refers to the longitudinal elastic modulus of the fiber. This refers to the transverse elastic modulus of the fiber. To ensure the parameter selection results are engineering-feasible and cover the range of commonly used high-performance fibers, this embodiment preferably sets it as follows: The range of values ​​is , The range of values ​​is , The range of values ​​is Discrete variables are in a woven configuration. Four representative 2.5D weaving configurations were selected. , , and The four configurations correspond to different warp yarn ratios, and their differences reflect the varying levels of axial load-bearing capacity, lateral constraint, and thickness-direction interlocking in the weaving structure. By unifying the aforementioned continuous variables and discrete configurations into the same screening framework, the fragmented approach of separately optimizing material parameters and subsequently supplementing the configuration judgment in existing methods can be avoided.

[0030] In this embodiment, as Figure 2 As shown, the establishment of the multi-scale equivalent performance database adopts a serial analysis process from microscopic fiber bundle unit cells to mesoscopic braided unit cells. First, a representative unit cell model of the fiber bundle is established at the microscopic level. This microscopic model uses a regular hexagonal array fiber arrangement to balance the efficiency of parametric modeling with the stability of the equivalent average response. For any given... Fiber radius and unit cell thickness The fiber center distance can be determined from the fiber volume fraction relationship, generating the corresponding fiber-matrix periodic unit cell. At the microscopic level, the equivalent stiffness matrix and equivalent engineering constants of the yarn are obtained by applying periodic boundary conditions and average strain control loads to the unit cell. Since the goal of this invention is design screening rather than just mechanism analysis, the main function of the microscopic level is to transform changes in material parameters at the component level into equivalent yarn parameters that can be directly accessed at the microscopic level, thereby establishing a clear, stable, and repeatable data transfer relationship between scales.

[0031] At a detailed level, this embodiment establishes braided unit cell models for four representative 2.5D braiding configurations. The differences between the configurations mainly lie in the different distribution ratios and path morphologies of warp and weft yarns within the unit cell, resulting in differences in the sensitivity of the overall equivalent properties to changes in material parameters. For any configuration The equivalent yarn parameters obtained at the microscopic level are assigned to the yarn region in the mesoscopic braided unit cell, and homogenization analysis is performed under uniform boundary conditions and average strain loading to obtain the overall mesoscopic engineering constants. To fully preserve structural response information, in this embodiment, the response vector at the mesoscopic level is denoted as: ; in, , , These are the elastic moduli in the three principal directions. , , Poisson's ratio, , , This refers to the shear modulus. Considering that this invention currently primarily focuses on target engineering constant-driven screening, the subsequent feasible region identification and ranking process preferably revolves around... , , The three principal moduli are expanded, but other engineering constants can still be retained in the database for direct use when expanding to multi-index collaborative constraints later.

[0032] To enable the discrete sample database to be used for fast searches in the continuous parameter space, this embodiment establishes a database based on continuous material parameters for each configuration. A surrogate model is used to obtain the target engineering constant. Since the response variation patterns differ under different configurations, mixing all configuration samples in the same surrogate model can easily mask configuration differences and reduce fitting accuracy. Therefore, this embodiment adopts a configuration-separate modeling approach, i.e., for... , , and Independent response surfaces were fitted separately. To minimize the impact of differences in the dimensions of different variables on the stability of the fit, it was preferable to fit continuous design variables. After standardization, its expression is: ; in, and They represent the first The mean and standard deviation of each design variable in the configuration sample set.

[0033] For any response variable: , Its second-order response surface can be expressed as: ; Where β0 is a constant term, β i Standardize design variables The coefficient of the first term, β ij The coefficients are the quadratic terms of each design variable. Each coefficient is obtained by fitting discrete sample data under the corresponding weaving configuration and is used to characterize the response relationship of the target engineering constant with the change of design variables. Equivalent engineering constant The predicted values ​​are: i,j=1,2,3.

[0034] By using the discrete response surfaces of the above configuration, the finite element results that originally only existed on discrete sample points can be extended to approximate prediction functions in continuous parameter space, providing a foundation for subsequent feasible region identification and parameter search.

[0035] In this embodiment, the preferred set of target engineering constants is: ; The corresponding target value is preferably set as follows: ; The reason for selecting the above three indicators as the target set is that, for 2.5D braided composite materials, the three-dimensional principal modulus can relatively comprehensively characterize the stiffness level of the structure in the main load-bearing direction, the transverse constraint direction, and the thickness direction, and can reflect the main mechanical characteristics under the coupling effect of material parameters and configuration. Compared with using only the modulus in a single direction as the screening criterion, conducting reverse screening with three-objective constraints better reflects the engineering applicability of this invention under the synergistic constraints of multiple indicators.

[0036] Once the target value is given, the parameter points that satisfy the constraints are searched in the continuous parameter space corresponding to each weaving configuration. For any set of candidate parameters... If the following conditions are met: j=1,2,3; Then the set of parameters is determined to be a feasible solution, where, Indicates the first Target values ​​of project engineering constants The equivalent engineering constants predicted by the surrogate model under this parameter set are represented by the parameter points that meet all the conditions, forming the parameter feasible region for this configuration. Since different configurations respond differently to material parameter disturbances, the feasible region shape, area range, and parameter clustering region of the four configurations are usually different. For configurations with a low warp yarn ratio, the feasible region often relies more on a higher volume fraction to compensate for insufficient axial load; for configurations with a high warp yarn ratio, the target stiffness requirement can often be achieved under a lower volume fraction condition through a higher longitudinal modulus. Based on these differences between configurations, this invention adopts separate searches for each configuration, rather than treating all structures as a single category.

[0037] Considering that the feasible region typically contains multiple parameter points that satisfy the objective constraints, and that different parameter points, although all satisfying the constraints, have varying degrees of closeness to the objective value, this embodiment further introduces a ranking criterion for feasible solutions. Let the first... The relative ratio of the project target is: j=1,2,3; And set weighting coefficients ,satisfy: j=1,2,3; In this embodiment, the preferred method is: ; The reason for this design is because As the modulus of the main load-bearing direction, it usually has a more significant impact on the structural service performance, and therefore is given a higher weight. It reflects the lateral load-bearing capacity and the structural in-plane coordination capacity, with the second highest weight. While thickness-direction stiffness is equally important, its role is slightly less significant than the other two in some component scenarios, thus it is given a lower weight. Under this weighting system, the feasible solution scoring function is defined as: j=1,2,3; when The smaller the value, the lower the redundancy and the closer the candidate parameter point is to the target value. Therefore, it is more suitable as the optimal solution output.

[0038] like Figure 3 As shown, in actual design, there may be situations where no single configuration fully satisfies all objective constraints. To avoid the drawback of terminating the design process upon finding no solution in existing technologies, this embodiment further introduces an approximate solution evaluation mechanism. For any objective item, the shortage degree is defined as: j=1,2,3; Based on this, an approximate scheme scoring function is established: j=1,2,3; When there is no strictly feasible solution under a certain configuration, it can be utilized The approximate solution that best matches the target requirement is selected from all candidate parameter points. This setting allows the invention to output design suggestions with engineering reference value even when the strictly feasible region is empty, rather than simply giving a conclusion that there is no solution. Especially when multi-objective constraints are tight and there is a natural upper limit to configuration capability, this approximate solution mechanism can provide a clear basis for subsequent configuration adjustments, target fine-tuning, or material compensation.

[0039] Since the number of candidate points obtained from continuous parameter space search is usually large, and dense points with very similar parameters and highly similar predicted responses often appear in local feasible regions, this embodiment further performs representative screening of candidate points after sorting. Specifically, under a preset parameter distance threshold or response distance threshold, candidate points that are too close are grouped into the same cluster, and only the parameter point with the best score or closest to the target is retained as a representative solution. This process can effectively reduce redundancy caused by repeated parameter output and improve the efficiency of subsequent positive verification and engineering selection. The number of solutions after representative screening can be set according to actual needs. Preferably, at least one representative solution is retained for each configuration. If the feasible region is large, multiple representative solutions from different regions can also be retained for designers to further compare.

[0040] In a preferred embodiment of this example, after performing reverse screening on the four representative configurations, the following results were obtained: , and The configuration can obtain a feasible solution under the three objective constraints, and The configuration does not have a strictly feasible solution under the current objective combination; only an optimal approximate solution can be obtained. Further, after sorting and representative selection, one representative solution is retained from each of the four configurations. The preferred representative solution is: The parameters of the approximate scheme corresponding to the configuration are: ; The parameters of the feasible scheme corresponding to the configuration are: ; The parameters of the feasible scheme corresponding to the configuration are: ; The parameters of the feasible scheme corresponding to the configuration are: ; As can be seen from the above results, there are significant differences in the optimal parameter combinations corresponding to different configurations, indicating that the present invention does not simply provide globally unique material parameters, but is able to identify the influence of configuration differences on the optimal material path.

[0041] To ensure the reliability of the surrogate model selection results, this embodiment further performs forward finite element verification on the representative schemes obtained. The forward verification serves two purposes: first, to verify the consistency between the surrogate model prediction results and the high-fidelity finite element recalculation results; and second, to compare the degree of satisfaction between the final scheme and the target engineering constants. For any representative scheme, the relative deviation between the surrogate prediction value and the forward verification value can be defined as: j=1,2,3; At the same time, the deviation of the verification value from the target value can also be defined as: j=1,2,3; in, For the surrogate model's predicted values, These are the values ​​calculated using the forward finite element method. These correspond to the target value. Using the two deviation indicators mentioned above, the selection scheme can be evaluated from two aspects: model prediction accuracy and the degree to which the target is met.

[0042] In a preferred embodiment, after performing forward finite element analysis on the four representative schemes, the recalculation results and the selected prediction results are in good agreement overall, and the absolute values ​​of the prediction deviations of the three principal moduli do not exceed [a certain value]. This result demonstrates that the configuration-discrete surrogate model constructed in this invention can accurately characterize the response of the three principal moduli in the continuous parameter space to changes in fiber parameters, and also shows that the process of feasible region identification, ranking optimization, and representativeness retention does not compromise the reliability of the results. The configuration, although it failed to obtain a strictly feasible solution under the current three-objective combination, through... The function can still obtain an approximate solution that is closest to the target, which can serve as an important reference for further adjusting the target value or switching configurations in engineering.

[0043] This embodiment further demonstrates that the parameter paths required to achieve the target engineering constants are not the same for different configurations. (The example also mentions a lower proportion of warp yarns.) Configurations tend to improve and reduce The reason for achieving the target requirements in this way is that a higher volume fraction can compensate for the insufficient axial load-bearing capacity of the configuration as a whole, while a lower longitudinal modulus can still meet the target while avoiding excessive redundancy; conversely, a higher warp yarn proportion and The configuration tends to adopt a relatively low Highly compatible This is because its configuration itself already possesses strong support capacity in the main load-bearing direction, thus achieving the target at a lower volume fraction by increasing the longitudinal stiffness of the fibers. This demonstrates that the present invention not only provides the final parameter scheme but also reveals the parameter matching rules behind the configuration differences, making the screening process more interpretable.

[0044] To facilitate the explanation of the complete closed loop of the present invention, the overall method flow is further summarized in this embodiment as follows: Figure 4As shown: First, a multi-scale equivalent performance database is established based on the preset fiber parameter range and four weaving configurations; second, the response surface of the principal modulus with respect to continuous parameters is established for each configuration; then, the target engineering constants are input. and weighting coefficients Search for feasible solutions and calculate them under different configurations. If a certain configuration has no feasible solution, then calculate... The process involves outputting the optimal approximate solution, then retaining representative solutions from the selection process, and finally performing a forward finite element method (FEM) verification on the representative solutions, evaluating their accuracy and target satisfaction using a deviation index. This workflow achieves a complete integration from database construction, surrogate prediction, target constraint selection, solution comparison to result verification.

[0045] The method for screening parameters of braided composite materials oriented towards target engineering constants provided in this embodiment is as follows: Figure 1 As shown, the specific steps include the following: Step 1: Database Construction. Establish a multi-scale equivalent property database for 2.5D braided composite materials. The database includes continuous material parameters, discrete braiding configurations, and corresponding yarn equivalent properties and mesoscopic equivalent engineering constants.

[0046] Among them, the continuous material parameters include at least fiber volume fraction, fiber longitudinal elastic modulus, and fiber transverse elastic modulus, and their design variable vector is: .

[0047] The discrete weave configuration is denoted as The 2.5D weave configuration includes at least four representative configurations. , , and Different configurations correspond to different warp yarn ratios and micro-geometric features, and independent parameter screening models are established for each.

[0048] The multi-scale equivalent property database was obtained through serial analysis of microscopic fiber bundle units and fine-scale braided units. At the microscopic level, parameters such as fiber volume fraction, fiber longitudinal elastic modulus, and fiber transverse elastic modulus were input to obtain the equivalent properties of the yarn: equivalent elastic modulus. , , Poisson's ratio , , shear modulus , , Microscopic level: The equivalent properties of the yarn are transferred to the microscopic weave unit cell to obtain the microscopic equivalent engineering constants under different weave configurations. The equivalent engineering constants include at least the elastic modulus in the principal direction. , and This forms a data foundation composed of continuous material parameters, discrete weaving configurations, and engineering constant responses.

[0049] In this embodiment, the multi-scale equivalent performance database preferably adopts a configuration-hierarchical storage method. Specifically, an independent data subset can be established for each weaving configuration, and each sample in each subset contains at least the following fields: sample number, configuration number, and fiber volume fraction. Fiber longitudinal elastic modulus Fiber transverse elastic modulus Microscopic equivalent elastic modulus , , Poisson's ratio , , shear modulus , , In a preferred embodiment, equivalent parameters of the micro-yarn can also be retained for direct retrieval when the screening process needs to be traced back to a more fundamental mechanism. This database format, organized hierarchically by configuration, has two advantages: firstly, it facilitates direct surrogate model fitting of data for a single configuration; secondly, it facilitates the side-by-side comparison of feasible domain features, candidate schemes, and representative retention schemes for different configurations during the output stage.

[0050] Step 2: Surrogate Model Training. For different weaving configurations, surrogate models are established to represent continuous material parameters to target engineering constants, thus expanding the discrete sample response into a continuous parameter space expression. The target engineering constant constraints are input, and feasible material parameter regions satisfying the constraints are identified within the continuous material parameter space.

[0051] In this embodiment, the establishment of the proxy model is not limited to fitting each output independently; it can also be adjusted according to actual needs. , , They can be managed using either single response surface fitting or a unified approach with multiple outputs.

[0052] Preferably, a quadratic response surface surrogate model is established for each weaving configuration, and for any target engineering constant... The prediction expression is: ;in, These are the standardized design variables.

[0053] After the fitting is completed, the determination coefficient, root mean square error and residual distribution of each model can be further calculated to verify the fitting quality of the surrogate model.

[0054] After completing the proxy model construction, input the target engineering constant set. and corresponding target values Search for parameter points that satisfy the constraints in the continuous parameter space corresponding to different weaving configurations. When any parameter point... satisfy When a solution is found to be feasible, it is considered a feasible solution; all feasible solutions constitute the parametric feasible region under the objective engineering constant constraints. For each feasible solution, a weighting coefficient corresponding to each objective engineering constant is introduced. Relative ratio to target And a weighted scoring function is used. The results are sorted to identify preferred candidate solutions that are closer to the target engineering constants.

[0055] In this embodiment, when searching for feasible solutions in a continuous parameter space, at least one of grid search, random sampling search, or adaptive encrypted search is preferably used: For grid search, you can , , Discrete candidate points are generated within a preset range at a selected step size, and then substituted into the response surfaces of each configuration in turn for calculation. , , ; For random sampling search, candidate points can be randomly generated within the parameter range to improve efficiency in large-scale searches; For adaptive encrypted search, a coarse search can be performed globally first, and then the sampling density can be increased in the region close to the target to improve the accuracy of feasible domain boundary identification.

[0056] When no feasible solution that fully satisfies the constraints exists under a certain configuration, the shortage degree is further defined. And an approximate scheme scoring function is adopted. The system evaluates substandard solutions to output the optimal approximate solution. This setting not only filters parameter combinations that meet the objectives but also provides a design reference that most closely approximates the target requirements when a strictly feasible solution does not exist.

[0057] Step 3: Candidate Scheme Clustering and Retention. The candidate parameter combinations are sorted, and representative schemes are output. That is, after obtaining the sorting results, candidate points in the parameter space that are too close and have similar performance responses are representatively filtered out, eliminating duplicate schemes that are too close and have similar performance responses, in order to retain parameter combinations and weaving configuration schemes that are differentiated and representative.

[0058] In this embodiment, to avoid excessive concentration of candidate solutions, the present invention can set a minimum separation distance for candidate points in the parameter space. For any two candidate points... and Define the normalized parameter distance as: ; in, Indicates the first The scale difference of each variable within a preset search range. Less than the preset threshold When considering two candidate points, they can be regarded as similar solutions within the same local cluster, and only the one with the better score is retained. In this way, the present invention can compress a large number of local nearest neighbor solutions into a small number of representative design suggestions, avoiding the problem of a large number of output results with serious information redundancy. For the final representative solutions, it is preferable to sort and output them according to configuration number and scoring order, so as to facilitate designers to directly compare the preferred solutions under different configurations.

[0059] Step 4: Tabular Output. Perform forward finite element verification on representative schemes to obtain the final fiber parameters and weaving configuration selection results. Compare the verification results with the target values ​​and surrogate model prediction values ​​respectively, and verify the effectiveness of the method based on the degree of target satisfaction and prediction deviation, thus forming a complete closed loop consisting of target setting, reverse screening, representative optimization, and forward verification.

[0060] To facilitate engineering applications of the output results of this invention, this embodiment preferably employs tabular output. Specifically, after screening, at least the following types of result tables can be generated: a parameter range and configuration definition table to clarify the screening boundaries; a target engineering constant and weight setting table to explain the current design task; a representative candidate scheme table to list the parameter combinations and predicted values ​​corresponding to each configuration; a forward finite element verification result table to list the recalculated values ​​and deviations; and a scheme evaluation summary table to display the feasible, near-feasible, representative retention, and final recommendation results side-by-side. This tabular organization not only facilitates patent text writing but also makes it easier to transform into software tools or engineering screening modules in the future.

[0061] In this embodiment, it is preferable to use Approximate schemes and , , All feasible solutions are listed in the same overall results table, with a solution status field to distinguish between feasible and near-feasible solutions. For example, the solution status can be set to "near-feasible" or "feasible" in the table, along with the corresponding scoring function value. or This method allows for comparison of different configurations even if they meet varying degrees of requirements, all within the same view. This facilitates designers in making a final decision by comprehensively considering factors such as manufacturing difficulty, material costs, and deviations from the target.

[0062] This embodiment further illustrates that forward finite element verification is not limited to simple recalculation of the three principal moduli; it can also supplementarily verify other engineering constants or local stress distributions when necessary. For example, in some application scenarios, although the three principal moduli meet the target, if... or If the changing trend does not meet the structural durability requirements, a verification of the thickness shear modulus can be added at the final recommendation stage. Similarly, for components subjected to loads in a specific direction, the verification can be extended to the distribution of local high-stress areas to further check whether the selected parameter combination will cause unreasonable stress concentration. Thus, this invention focuses on the target engineering constant, but its methodological framework can be extended to a wider range of auxiliary verification indicators.

[0063] This embodiment also illustrates that the present invention is not limited to outputting only a single final solution. In a preferred embodiment, a first candidate set can be output according to the principle of at least one representative solution for each configuration, followed by a second manual screening based on manufacturing costs, material availability, or service condition requirements. For example, when designers prioritize axial load-bearing capacity, they can prioritize options with a higher warp yarn ratio. or Configuration; when designers are more concerned with manufacturing stability under high volume fraction conditions, they can prioritize configurations with more concentrated parameter distributions. Configuration; when strict objectives cannot be relaxed, options that only provide approximate solutions can be directly excluded. Configuration; when a certain indicator can be moderately relaxed, it can also be... The optimal approximate solution is retained as an alternative design. Therefore, the output of this invention is not merely a single parameter point, but a structured, interpretable, and comparable design decision result.

[0064] The essence of this invention lies in: based on a multi-scale equivalent performance database, using a configuration discrete surrogate model as a bridge, driven by target engineering constant constraints, and guaranteed by feasible solution ranking, approximate scheme output, and forward verification, achieving parameter screening of 2.5D braided composite materials within the joint space of continuous fiber parameters and discrete braiding configurations. Compared with existing methods that rely on iterative forward calculations, this invention can significantly improve the parameter identification efficiency in target performance design and enhance the interpretability of configuration comparisons and the engineering operability of the results output.

[0065] The parameter screening method for braided composite materials oriented towards target engineering constants provided in this embodiment first establishes a serial multi-scale analysis database from micro-fiber bundle cells to meso-braid cells. Fiber volume fraction, fiber longitudinal elastic modulus, and fiber transverse elastic modulus are used as continuous material parameters, while the 2.5D braid configuration is used as a discrete structural variable. Meso-equivalent engineering constants under different parameter combinations are obtained. Subsequently, surrogate models of the target engineering constants with respect to the continuous material parameters are constructed for different braid configurations, mapping discrete samples to the continuous parameter space. The feasible region of the parameters is identified under the given constraints of the target engineering constants. Further, candidate solutions that satisfy the constraints are sorted according to weighted deviations. When no feasible solution fully satisfies the constraints, a shortage evaluation is introduced to output the optimal approximate solution. Then, representative candidate solutions are deduplicated and screened, and positive finite element verification is performed, forming a closed loop consisting of target setting, reverse screening, positive verification, and result comparison. It can simultaneously handle the coupling relationship between fiber parameters and braiding configuration, realize rapid screening and optimization of target engineering constants, reduce the cost of repetitive finite element calculations, and provide reusable method support for the determination of fiber parameters, configuration selection and target performance design of 2.5D braided composite materials.

[0066] The parameter screening method for braided composite materials oriented towards target engineering constants provided in this embodiment has the following advantages: First, it can carry out synergistic screening in the joint space of continuous material parameters and discrete braiding configurations, rather than treating the two separately; second, it can transform discrete finite element samples into a feasible region identification problem in continuous parameter space, thereby significantly improving the efficiency of target performance design; third, it can still output the optimal approximate solution when there is no completely feasible solution, improving the engineering applicability of the method; and fourth, it enhances the reliability and reusability of the screening results by constructing a verification closed loop through forward finite element verification.

[0067] Example 2 The braided composite material parameter screening system oriented towards target engineering constants provided in this embodiment includes: The database construction module is configured to: establish a multi-scale equivalent performance database of braided composite materials, wherein the equivalent performance database includes continuous material parameters, discrete braiding configurations, and equivalent engineering constants; The identification module is configured to: establish a proxy model from continuous material parameters to equivalent engineering constants for different weaving configurations, input the target engineering constant constraints, and identify feasible material parameters that satisfy the constraints in the continuous material parameter space; The screening module is configured to sort feasible material parameters, perform representative screening, and perform positive finite element verification on representative schemes.

[0068] It should be noted that each module in this embodiment corresponds one-to-one with each step in Embodiment 1, and their specific implementation processes are the same, so they will not be repeated here.

[0069] Example 3 This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the method for screening parameters of braided composite materials oriented towards target engineering constants as described in Embodiment 1 above.

[0070] Example 4 This embodiment provides a computer device, such as... Figure 5 As shown, the system includes a computer-readable storage medium 1003, a processor 1001, a communication interface 1002, and a computer program stored on the computer-readable storage medium 1003 and executable on the processor 1001. The processor 1001, communication interface 1002, and computer-readable storage medium 1003 can be connected via a bus or other means. The communication interface 1002 is used to receive and send data. When the processor 1001 executes the program, it implements the steps in the method for screening parameters of braided composite materials oriented towards target engineering constants as described in Embodiment 1 above.

[0071] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for screening parameters of braided composite materials oriented towards target engineering constants, characterized in that, include: A multi-scale equivalent property database of braided composite materials is established, which includes continuous material parameters, discrete braiding configurations, and equivalent engineering constants. For different weaving configurations, surrogate models from continuous material parameters to equivalent engineering constants are established. The target engineering constant constraints are input, and feasible material parameters that satisfy the constraints are identified in the continuous material parameter space. After sorting the feasible material parameters, representative screening is carried out, and the representative schemes are verified by positive finite element method. The proxy model adopts a quadratic response surface proxy model for any equivalent engineering constant. The prediction expression is: in, For standardized design variables, β0 is a constant term, β i Standardize design variables The coefficient of the first term, β ij These are the coefficients of the quadratic terms of each design variable; Equivalent engineering constant The predicted value; The representative screening includes: for any two candidate points and Define the normalized parameter distance as: ;in, Indicates the first The scale difference of each variable within a preset search range; when Less than the preset threshold When two candidate points are considered as similar schemes in the same local cluster, only the one with the better score is retained; The positive finite element verification is based on the relative deviation between the surrogate prediction value and the positive verification value. The deviation of the verification value from the target value is: ;in, For the surrogate model's predicted values, These are the values ​​calculated using the forward finite element method. This is the target value.

2. The method for screening parameters of braided composite materials oriented towards target engineering constants as described in claim 1, characterized in that, The continuous material parameters include fiber volume fraction, fiber longitudinal elastic modulus, and fiber transverse elastic modulus; The equivalent engineering constants include the principal elastic modulus, Poisson's ratio, and shear modulus.

3. The method for screening parameters of braided composite materials oriented towards target engineering constants as described in claim 1, characterized in that, For feasible material parameters, weighting coefficients corresponding to each equivalent engineering constant are introduced. Relative ratio to target And a weighted scoring function is used. Sort; where, For each target value, These are the predicted values ​​of the surrogate model for each equivalent engineering constant.

4. The method for screening parameters of braided composite materials oriented towards target engineering constants as described in claim 3, characterized in that, When there are no feasible material parameters that fully satisfy the constraints under a certain weaving configuration, a shortage degree is defined. And adopt an approximate scheme scoring function. The substandard solutions are evaluated to output the optimal approximate solution.

5. A parameter screening system for braided composite materials oriented towards target engineering constants, characterized in that it includes: The database construction module is configured to: establish a multi-scale equivalent performance database of braided composite materials, wherein the equivalent performance database includes continuous material parameters, discrete braiding configurations, and equivalent engineering constants; The identification module is configured to: establish a proxy model from continuous material parameters to equivalent engineering constants for different weaving configurations, input the target engineering constant constraints, and identify feasible material parameters that satisfy the constraints in the continuous material parameter space; The screening module is configured to: sort feasible material parameters, perform representative screening, and perform positive finite element verification on representative schemes; The proxy model adopts a quadratic response surface proxy model for any equivalent engineering constant. The prediction expression is: in, For standardized design variables, β0 is a constant term, β i Standardize design variables The coefficient of the first term, β ij These are the coefficients of the quadratic terms of each design variable; Equivalent engineering constant The predicted value; The representative screening includes: for any two candidate points and Define the normalized parameter distance as: ;in, Indicates the first The scale difference of each variable within a preset search range; when Less than the preset threshold When two candidate points are considered as similar schemes in the same local cluster, only the one with the better score is retained; The positive finite element verification is based on the relative deviation between the surrogate prediction value and the positive verification value. The deviation of the verification value from the target value is: ;in, For the surrogate model's predicted values, These are the values ​​calculated using the forward finite element method. This is the target value.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the method for screening parameters of braided composite materials oriented towards target engineering constants as described in any one of claims 1-4.

7. A computer device comprising a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the method for screening parameters of braided composite materials oriented towards target engineering constants as described in any one of claims 1-4.

Images