A data-driven composite damage failure rapid calculation analysis method

Abstract

The application relates to a data-driven composite material damage failure rapid calculation analysis method, which comprises the following steps: S1, acquiring a composite material to construct a complex mesoscopic unit cell model; S2, applying a strain load to the RVE unit in the complex mesoscopic unit cell model as the composite material under a microscale to acquire a damage variable of the RVE unit; S3, uniformly processing the damage variable to respectively acquire a damage variable of the composite material under a mesoscale and a damage variable of the composite material under a macroscale; S4, applying a strain load to the complex mesoscopic unit cell model to acquire the strain of each integral point under the macroscale, transmitting the strain of the integral point to the RVE unit as a boundary condition to update and correct the model parameters of the RVE unit; and S5, taking the corrected RVE unit as the composite material under the microscale, repeating steps S2 and S3 to acquire the final damage variables of the composite material under the microscale, the mesoscale and the macroscale.

Description

Technical Field

[0001] This invention relates to the field of composite material analysis technology, and in particular to a data-driven method for rapid calculation and analysis of composite material damage and failure. Background Technology

[0002] Advanced composite material structures, with their high production efficiency, strong designability, and excellent near-net-shape forming capability for complex structures, have been widely used in aerospace, transportation, and marine engineering. However, due to the inherent multi-scale characteristics of composite material structures—which exhibit different structural forms at macro-micro-scales—and the fact that their failure behavior also displays multi-scale, multi-loss mode coupling failure characteristics, the simulation of composite material failure processes is extremely difficult.

[0003] For composite material structures, traditional phenomenological models can only describe the equivalent mechanical properties at the macroscopic scale, making it difficult to characterize local mechanical behaviors such as matrix plasticity and component damage. Furthermore, the material parameters involved in composite materials are relatively complex, and traditional research methods, which are costly and time-consuming, are no longer suitable for the current development trends of composite materials.

[0004] Therefore, there is an urgent need for a data-driven method for rapid calculation and analysis of composite material damage and failure. Based on the data-driven analysis method, a concurrent multi-scale simulation framework coupling macro-micro, micro-micro, and macro-micro-micro scales of composite materials was established to improve the computational efficiency of composite material failure analysis. Summary of the Invention

[0005] The purpose of this invention is to propose a data-driven method for rapid computational analysis of damage and failure in composite materials, in order to solve the problems existing in the prior art. Taking the complex microscopic unit cell of woven composite materials as the research object, the elastoplastic progressive failure behavior of the material under uniaxial tensile and compressive loads is examined. An efficient reduced-order model based on the data-driven analysis method is proposed. On this basis, a concurrent multi-scale simulation framework of macro-micro, micro-micro, and macro-micro-micro coupling of composite materials is established, thereby improving the computational efficiency of composite material failure analysis.

[0006] To achieve the above objectives, the present invention provides the following solution: A data-driven method for rapid calculation and analysis of damage and failure in composite materials includes: S1. Obtain the composite material, and construct a complex micro-cell model based on the composite material; the complex micro-cell model is composed of several RVE units; S2. Apply strain load to the RVE unit in the complex micro-scale unit cell model as a composite material at the microscale, and obtain the damage variable of the RVE unit. S3. The damage variables of the RVE unit are homogenized to obtain the damage variables of the composite material at the mesoscale and the damage variables of the composite material at the macroscale. S4. Apply strain load to the complex micro-scale unit cell model, obtain the strain at each integration point at the macro scale, and use the strain at the integration point as a boundary condition to update the RVE element, thereby correcting the model parameters of the RVE element. S5. Using the modified RVE unit as the composite material at the microscale, repeat steps S2 and S3 to obtain the final damage variables of the composite material at the microscale, mesoscale, and macroscale.

[0007] Optionally, constructing the complex mesoscopic unit cell model includes: Based on the composite material, cross-sections with different geometric parameters are obtained as fibers; By gradually adding fibers according to a preset fiber content, the composite material is woven, and cross-sectional image information of the woven composite material is obtained to generate the complex microscopic unit cell model.

[0008] Optionally, obtaining the damage variables of the RVE unit includes: An equivalent boundary strain is applied to the complex microscopic unit cell model, and the displacement of the element nodes is obtained by finite element method. The strain tensor is then derived from the displacement. Based on the strain tensor and the constitutive relation of the constituent materials, the stress tensor is obtained. Based on the stress tensor, the damage of the RVE element is identified using the initial criterion. That is, when the principal stress or shear stress of the stress tensor reaches or exceeds the strength limit corresponding to the material, it is proven that the RVE element has initial damage, and the damage variable of the RVE element is obtained. The strength limits include: longitudinal tensile strength, transverse compressive strength and shear strength of the yarn.

[0009] Optionally, identifying the damage to the RVE unit using initial criteria includes: The initial criteria are used to determine whether the RVE element has been damaged. If no damage has occurred, the damage variable of the RVE element is calculated using a damage model, and the material properties of the RVE element are reduced to obtain the stiffness of the RVE element after reduction. Based on the stiffness, the strain load is updated, the stress and strain of the RVE element are recalculated, and the RVE element is again determined to be damaged. This process continues until the RVE element is damaged. If damage has occurred, the damage variable of the RVE element is calculated using a damage model, and the material properties of the RVE element are reduced to obtain the stiffness of the RVE element after reduction. Based on the stiffness, the strain load is updated, the stress and strain of the RVE element are recalculated, and the RVE element is again determined to be damaged. This process continues until the iteration conditions are met.

[0010] Optionally, the iteration conditions include: Determine whether the Voxe1 unit is equal to all units. If it is not equal to all units, increment the number of Voxe1 units by 1 and repeat the determination of whether the RVE unit has suffered secondary damage. If it is equal to all units, end the determination.

[0011] Optionally, the damage variables include: longitudinal damage variables, transverse damage variables, and shear damage variables.

[0012] Optionally, the method for calculating the damage variables of the RVE unit using a damage model is as follows: ; ; ; in, For longitudinal damage variables, For lateral damage variables, For shear damage variables, and These represent the principal stress and stress components of the yarn, respectively. , , , and These represent the longitudinal tensile strength, transverse tensile strength, longitudinal compressive strength, transverse compressive strength, and shear strength of the yarn, respectively.

[0013] Optionally, modifying the model parameters of the RVE unit includes: Set the integral point strain boundary condition, and calculate and update the equivalent stiffness matrix, local damage variable distribution field, and element orientation tensor of the RVE element based on the applied integral point strain boundary condition. Among them, the stiffness matrix update is used to reflect the degradation characteristics of material properties as damage evolves; the damage variable distribution is used to determine the local response state of the RVE element; and the orientation tensor is used to maintain the physical consistency between the local coordinate system and the global coordinate system, and to transmit macroscopic and mesoscopic information between different scales.

[0014] Optionally, obtaining the damage variables of the composite material at the mesoscale and the damage variables of the composite material at the macroscale includes: The damage variables of the RVE unit are homogenized to obtain the damage variables of the composite material at the microscale. The damage variables of the composite material at the microscale are then homogenized to obtain the damage variables of the composite material at the macroscale.

[0015] The beneficial effects of this invention are as follows: This invention establishes a concurrent multi-scale simulation framework for macro-micro, micro-micro, and macro-micro-micro coupled composite materials, enabling real-time prediction of multi-scale damage behavior of composite material structures and providing a new approach for the design and application of composite materials. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in 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.

[0017] Figure 1 This is a flowchart of a data-driven rapid calculation and analysis method for composite material damage and failure according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the multi-scale analysis of the woven composite material from the micro to the macro level in an embodiment of the present invention; Figure 3 This is a schematic diagram of the concurrent multi-scale method according to an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] like Figure 1 As shown, this embodiment discloses a data-driven method for rapid calculation and analysis of composite material damage and failure, including: S1, acquiring the composite material and constructing a complex mesoscopic unit cell model based on the composite material; the complex mesoscopic unit cell model consists of several RVE elements; S2, applying strain loads to the RVE elements in the complex mesoscopic unit cell model as the composite material at the microscale, and obtaining the damage variables of the RVE elements; S3, homogenizing the damage variables of the RVE elements, and obtaining the damage variables of the composite material at the mesoscopic scale and the damage variables of the composite material at the macroscopic scale respectively; S4, applying strain loads to the complex mesoscopic unit cell model, obtaining the strain at each integration point at the macroscopic scale, and using the strain at the integration point as a boundary condition to update the RVE elements, correcting the model parameters of the RVE elements; S5, using the corrected RVE elements as the composite material at the microscopic scale, repeating steps S2 and S3, and obtaining the final damage variables of the composite material at the microscopic, mesoscopic, and macroscopic scales, specifically: First, a rapid parametric modeling method was used to construct the woven composite material. This rapid parametric modeling primarily involved using Python scripts to define the value ranges of relevant dimensional variables, thereby quickly generating a batch of finite element model databases. Next, a damage analysis database was constructed based on the failure modes of the composite material to calculate its damage and failure behavior. Finally, the superior computational efficiency of the multi-scale method was verified by comparing the consistent multi-scale method with direct multi-scale numerical methods.

[0021] Constructing a complex mesoscopic unit cell model includes: obtaining cross-sections with different geometric parameters as fibers based on the composite material; gradually adding fibers according to a preset fiber content to weave the composite material; obtaining cross-sectional image information of the woven composite material to generate a complex mesoscopic unit cell model, specifically: A rapid parametric modeling method is used to construct complex microscopic unit cell models (RVEs) of braided composite materials: First, a modeling script with a specified format is written in Python to generate a series of cross-sections with different geometric parameters as fibers; second, fibers are added step by step according to the specified fiber content; based on the cross-sectional image information of these composite materials, the required three-dimensional computational unit model can be directly generated.

[0022] Constructing fiber-matrix interface elements: Interface elements are used to simulate the thin-layer structure between the fiber and the matrix, playing a crucial transitional role in mechanical properties. The interface is typically set as a ring-shaped shell element with a thickness of 1 to 5 micrometers, and a modulus gradient is set in the thickness direction to reflect non-uniform properties. During the modeling phase, interface elements achieve differentiated expressions of mechanical behavior by defining contact properties or custom material models. In the finite element solution, they participate as independent elements in stress field calculations. Although not independently invoked in the main workflow, they play an auxiliary role in stiffness summarization and damage accumulation.

[0023] The damage variables of the RVE element are obtained by: applying equivalent boundary strain to the complex mesoscopic unit cell model, obtaining the displacement of the element nodes using finite element method, and deriving the strain tensor from the displacement; obtaining the stress tensor based on the strain tensor and the constitutive relation of the constituent materials; and identifying the damage of the RVE element using the initial criterion based on the stress tensor, i.e., when the principal stress or shear stress of the stress tensor reaches or exceeds the strength limit corresponding to the material, it is proven that the RVE element has undergone initial damage, and the damage variables of the RVE element are obtained; among which, the strength limit includes: the longitudinal tensile strength, transverse compressive strength, and shear strength of the yarn.

[0024] The method for identifying damage to RVE elements using initial criteria includes: determining whether the RVE element has been damaged using initial criteria; if no damage has occurred, calculating the damage variable of the RVE element using a damage model, reducing the material properties of the RVE element, obtaining the stiffness of the reduced RVE element, updating the strain load based on the stiffness, recalculating the stress and strain of the RVE element, and determining whether the RVE element has been damaged again, until the RVE element is damaged; if damage has occurred, calculating the damage variable of the RVE element using a damage model, reducing the material properties of the RVE element, obtaining the stiffness of the reduced RVE element, updating the strain load based on the stiffness, recalculating the stress and strain of the RVE element, and determining whether the RVE element has suffered secondary damage, until the iteration conditions are met.

[0025] The iteration conditions include: determining whether the Voxe1 element is equal to all elements; if not, incrementing the number of Voxe1 elements by 1, and repeating the check to see if secondary damage has occurred to the RVE element; if equal to all elements, the process ends. Specifically: Damage Model Construction: The failure modes of braided composite materials mainly include yarn failure and matrix failure. A simple stiffness reduction model is selected as the damage model to study the damage evolution of yarn and matrix in braided composite materials. Three damage variables of the yarn (longitudinal damage variable) are considered in the local coordinate system. I L, lateral damage variable I T and shear damage variables I S).

[0026] The method for calculating the damage variables of RVE elements using a damage model is as follows: ; ; ; in, For longitudinal damage variables, For lateral damage variables, For shear damage variables, and These represent the principal stress and stress components of the yarn, respectively. , , , and These represent the longitudinal tensile strength, transverse tensile strength, longitudinal compressive strength, transverse compressive strength, and shear strength of the yarn, respectively.

[0027] The modification of the RVE element model parameters includes: setting the integral point strain boundary condition; and, based on the applied integral point strain boundary condition, calculating and updating the equivalent stiffness matrix, local damage variable distribution field, and element orientation tensor of the RVE element. Among these, the stiffness matrix update is used to reflect the degradation characteristics of material properties with damage evolution; the damage variable distribution is used to determine the local response state of the RVE element; and the orientation tensor is used to maintain the physical consistency between the local coordinate system and the global coordinate system, and to transfer macroscopic and mesoscopic information between different scales.

[0028] Obtaining the damage variables of the composite material at both the mesoscale and macroscale scales includes: homogenizing the damage variables of the RVE elements to obtain the damage variables of the composite material at the mesoscale, and homogenizing the damage variables of the composite material at the mesoscale to obtain the damage variables of the composite material at the macroscale. Specifically: like Figure 2-3 As shown, a concurrent multi-scale simulation framework is constructed: the multi-scale framework includes top-down localization and bottom-up homogenization. Braided composite materials exhibit typical hierarchical structural characteristics, and the entire structure or sample is considered to be at the macroscopic scale. A periodic RVE extracted from the sample is used to predict macroscopic mechanical properties. The braided yarns inside the braided composite material are composed of thousands of fibers, which are considered to be at the microscopic scale. The homogenization result is transferred from the low-scale to the high-scale.

[0029] Damage Analysis of Braided Composite Materials: For the braided composite material structure in this embodiment, at the macroscopic scale, the braided composite material structure is considered to be a homogeneous material and discretized into a finite element mesh, which is solved using the Newton-Raphson iterative method. The integration point of each element corresponds to the braided composite material stiffness matrix (RVE) at the mesoscopic scale, and information transfer between the two scales is established through the integration point and the RVE. The strain at each integration point at the macroscopic scale is used as a boundary condition and transferred to the mesoscopic RVE model. The response of the mesoscopic RVE is then calculated using the damage model constructed in the third step. Unit strain boundary conditions in different directions are applied to the mesoscopic RVE elements to obtain their response stress. Combining the coordinate tensor transformation formula, the RVE stiffness response is transformed into an equivalent stiffness tensor in global coordinates. Volume averaging is performed on the stiffness tensors of all RVE elements to finally generate the macroscopic stiffness matrix corresponding to the integration point. The damage state in the RVE element is represented by three damage variables (longitudinal damage variable...). Horizontal damage variables Shear damage variables The process involves characterizing the damage variable. In each iteration, the increase in the current damage variable is calculated based on the principal stress or strain state at the integration point, and the material stiffness is updated accordingly. If the iteration result still does not meet the convergence condition, the iteration continues until the damage variable converges or a preset termination condition is reached. The updated stiffness is then used in the next stress field calculation to achieve coupled local-global updates.

[0030] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A data-driven method for rapid calculation and analysis of damage and failure in composite materials, characterized in that, include: S1. Obtain the composite material, and construct a complex micro-cell model based on the composite material; the complex micro-cell model is composed of several RVE units; S2. Apply strain load to the RVE unit in the complex micro-scale unit cell model as a composite material at the microscale, and obtain the damage variable of the RVE unit. S3. The damage variables of the RVE unit are homogenized to obtain the damage variables of the composite material at the mesoscale and the damage variables of the composite material at the macroscale. S4. Apply strain load to the complex micro-scale unit cell model, obtain the strain at each integration point at the macro scale, and use the strain at the integration point as a boundary condition to update the RVE element, thereby correcting the model parameters of the RVE element. S5. Using the modified RVE unit as the composite material at the microscale, repeat steps S2 and S3 to obtain the final damage variables of the composite material at the microscale, mesoscale, and macroscale.

2. The data-driven rapid calculation and analysis method for composite material damage and failure according to claim 1, characterized in that, Constructing the complex mesoscopic unit cell model includes: Based on the composite material, cross-sections with different geometric parameters are obtained as fibers; By gradually adding fibers according to a preset fiber content, the composite material is woven, and cross-sectional image information of the woven composite material is obtained to generate the complex microscopic unit cell model.

3. The data-driven rapid calculation and analysis method for composite material damage and failure according to claim 1, characterized in that, Obtaining the damage variables of the RVE unit includes: An equivalent boundary strain is applied to the complex microstructure unit cell model, and the displacement of the element nodes is obtained by finite element method. The strain tensor is then derived from the displacement. Based on the strain tensor and the constitutive relation of the constituent materials, the stress tensor is obtained. Based on the stress tensor, the damage of the RVE element is identified using the initial criterion. That is, when the principal stress or shear stress of the stress tensor reaches or exceeds the strength limit corresponding to the material, it is proven that the RVE element has initial damage, and the damage variable of the RVE element is obtained. The strength limits include: longitudinal tensile strength, transverse compressive strength and shear strength of the yarn.

4. The data-driven rapid calculation and analysis method for composite material damage and failure according to claim 3, characterized in that, Identifying the damage to the RVE unit using initial criteria includes: The initial criteria are used to determine whether the RVE element has been damaged. If no damage has occurred, the damage variable of the RVE element is calculated using a damage model, and the material properties of the RVE element are reduced to obtain the stiffness of the RVE element after reduction. Based on the stiffness, the strain load is updated, the stress and strain of the RVE element are recalculated, and the RVE element is again determined to be damaged. This process continues until the RVE element is damaged. If damage has occurred, the damage variable of the RVE element is calculated using a damage model, and the material properties of the RVE element are reduced to obtain the stiffness of the RVE element after reduction. Based on the stiffness, the strain load is updated, the stress and strain of the RVE element are recalculated, and the RVE element is again determined to be damaged. This process continues until the iteration conditions are met.

5. The data-driven rapid calculation and analysis method for composite material damage and failure according to claim 4, characterized in that, The iteration conditions include: Determine whether the Voxe1 unit is equal to all units. If it is not equal to all units, increment the number of Voxe1 units by 1 and repeat the determination of whether the RVE unit has suffered secondary damage. If it is equal to all units, end the determination.

6. The data-driven rapid calculation and analysis method for composite material damage and failure according to claim 1, characterized in that, The damage variables include: longitudinal damage variables, transverse damage variables, and shear damage variables.

7. The data-driven rapid calculation and analysis method for composite material damage and failure according to claim 4, characterized in that, The method for calculating the damage variables of the RVE unit using the damage model is as follows: ； ； ； in, For longitudinal damage variables, For lateral damage variables, For shear damage variables, and These represent the principal stress and stress components of the yarn, respectively. , , , and These represent the longitudinal tensile strength, transverse tensile strength, longitudinal compressive strength, transverse compressive strength, and shear strength of the yarn, respectively.

8. The data-driven rapid calculation and analysis method for composite material damage and failure according to claim 1, characterized in that, The model parameters of the RVE unit are corrected as follows: Set the integral point strain boundary condition, and calculate and update the equivalent stiffness matrix, local damage variable distribution field, and element orientation tensor of the RVE element based on the applied integral point strain boundary condition. Among them, the stiffness matrix update is used to reflect the degradation characteristics of material properties as damage evolves; the damage variable distribution is used to determine the local response state of the RVE element; and the orientation tensor is used to maintain the physical consistency between the local coordinate system and the global coordinate system, and to transmit macroscopic and mesoscopic information between different scales.

9. The data-driven rapid calculation and analysis method for composite material damage and failure according to claim 1, characterized in that, Obtaining damage variables of composite materials at both the mesoscale and macroscale scales includes: The damage variables of the RVE unit are homogenized to obtain the damage variables of the composite material at the microscale. The damage variables of the composite material at the microscale are then homogenized to obtain the damage variables of the composite material at the macroscale.

Images