Method and system for establishing multi-scale prediction model of composite material pore expansion rate
Through the combination of multi-scale model and machine learning, the composite porosity prediction model is constructed, which solves the problem of insufficient model accuracy and applicability in the existing technology, and realizes high-precision and high-reliability prediction of composite materials. Real-time and dynamic adaptability have promoted the advancement of materials science and engineering technology.
Patent Information
- Application Number
- CN202411385227.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2024-12-31
- Estimated Expiration
- 2044-09-30
AI Technical Summary
The existing composite porosity prediction technology has insufficient model accuracy, limited applicability and poor ability to capture complex materials behaviors, making it difficult to achieve high-precision and high-reliability prediction, especially under new materials or new processes. Quickly adapt to different production environments.
Using the multi-scale prediction model establishment method, by collecting and pre-processing the experimental data of composite materials, microscopic, mesoscopic and macroscopic models are constructed, and a fusion feature matrix is formed in combination with machine learning models, and the porosity prediction model is trained, considering that the materials are different. Performance changes under environmental conditions, with real-time and dynamic adaptability.
It improves the accuracy and reliability of composite porosity prediction, can deeply learn complex material behavior, realize real-time prediction of different environmental conditions, and promotes the advancement of materials science and engineering technology.
Smart Images
Figure CN118888066B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of hole expansion rate prediction, and in particular to a method and system for establishing a multi-scale prediction model for the hole expansion rate of composite materials. Background Art
[0002] Composite materials are widely used in aerospace, automobile and construction fields due to their excellent mechanical properties and lightweight characteristics. As an important indicator to measure the processing and performance of materials, the porosity directly affects their structural integrity and durability. The porosity refers to the degree of pore expansion in a material under specific conditions, which is affected by many factors, including the microstructure, stress state and environmental conditions of the material. Therefore, in-depth research on the porosity of composite materials will not only help optimize the design and processing technology of the material, but also improve its reliability and safety in practical applications, thereby promoting technological progress and innovation in this field.
[0003] Although existing hole expansion rate prediction technologies have made some progress in many fields, there are still some significant shortcomings, such as the complexity of the model and the high sensitivity to input parameters, which makes it difficult to achieve high-precision and high-reliability predictions in practical applications. In addition, many methods rely on empirical data and lack sufficient theoretical basis, which limits their applicability under new materials or new process conditions and cannot quickly adapt to different production environments, which limits their widespread promotion and effectiveness in industrial applications.
[0004] The prior art has the following deficiencies:
[0005] Existing hole expansion rate prediction technologies generally have problems such as insufficient model accuracy, limited applicability, and weak ability to capture complex material behaviors. These deficiencies often lead to the inability to effectively predict the hole expansion phenomenon of materials in practical applications. First, many existing models rely on simplified assumptions and linear relationships, which may not accurately capture complex material behaviors and nonlinear characteristics. Secondly, the data acquisition process may be limited by experimental conditions, resulting in insufficient samples or unbalanced data, which affects the generalization ability of the model. In addition, the differences in microstructure and mechanical properties of different types of composite materials limit the applicability of general models. Finally, current prediction methods often lack real-time and dynamic adaptability, and are difficult to cope with rapidly changing working environments and load conditions. Therefore, further improving the accuracy and applicability of the model remains an urgent problem to be solved.
[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention
[0007] The purpose of the present invention is to provide a method and system for establishing a multi-scale prediction model of the pore expansion rate of a composite material, so as to solve the problems raised in the above-mentioned background technology.
[0008] To achieve the above object, the present invention provides a method for establishing a multi-scale prediction model of the pore expansion rate of a composite material, the specific steps comprising:
[0009] Step 1: Collect experimental data of composite materials, including material type, volume, density, temperature, stress, strain, elastic modulus, Poisson's ratio, microstructure information and pore expansion rate, and pre-process the collected data;
[0010] Step 2: Analyze the experimental data of the pretreated composite material and construct a multi-scale model, which includes a microscopic model, a mesoscopic model and a macroscopic model;
[0011] Step 3: Build a hole expansion rate prediction model based on the machine learning model, fuse the input and output of the multi-scale model to form a fusion feature matrix, and use the fusion feature matrix as the input of the hole expansion rate prediction model, and use the corresponding hole expansion rate as a label to train the hole expansion rate prediction model;
[0012] Step 4: Obtain relevant experimental data of the composite material to be predicted, and input it into the trained multi-scale model and pore expansion rate prediction model to obtain the predicted pore expansion rate;
[0013] Step 5: Analyze the output of the hole expansion rate prediction model and the temperature of the current material to obtain the hole expansion rate of the composite material to be predicted at the current moment.
[0014] Furthermore, the specific logic for collecting experimental data of materials is as follows:
[0015] Acquire experimental data of various composite materials, record the material type, volume, density, temperature, stress, strain, elastic modulus, Poisson's ratio, microstructure information and pore expansion rate of each composite material, and record the area of the composite material during the experiment, the force during the tensile test and the shear modulus during the shear test, calculate the mesoscopic stress and tensile strength, and the microstructure information includes the diameter and length of the fiber, the diameter and length of the pores, and the contact force between the fibers. Normalize the collected data and store the processed data as a sample data set;
[0016] The stiffness matrix of the composite material is constructed by analyzing the elastic modulus, Poisson's ratio and the shear modulus during the experiment. The formula for obtaining the mesoscopic stress is:
[0017] ;
[0018] in, represents the mesoscopic stress, represents the mesoscopic strain, is the stiffness matrix of the composite material, is the volume;
[0019] The formula for obtaining the tensile strength is:
[0020] ;
[0021] in, is the tensile strength, is the force, is the force bearing area.
[0022] Furthermore, the specific logic for constructing a multi-scale model is as follows:
[0023] The sample data set is divided into a training set and a test set with a ratio of 7:3, which are used to build microscopic models, mesoscopic models and macroscopic models respectively. The logic for building the microscopic model is as follows:
[0024] Input the material's density, temperature, elastic modulus and microstructure information into the micromodel, use molecular dynamics software to calculate, and use the corresponding porosity, stress, strain, and contact force between fibers as labels to train the micromodel;
[0025] The logic behind constructing the mesoscopic model is:
[0026] The output of the microscopic model is standardized, and the porosity and contact force between fibers output by the microscopic model are stored as eigenvalues of the mesoscopic model. A feature matrix is constructed, in which each row represents a sample and each column represents an eigenvalue. The feature matrix is checked for consistency and rationality and input into the mesoscopic model. The elastic modulus, mesoscopic stress, and temperature of the corresponding material are used as labels to train the mesoscopic model.
[0027] The logic behind building the macro model is:
[0028] The elastic modulus, mesoscopic stress, and temperature output by the mesoscopic model are used as material property inputs of the macroscopic model, and the corresponding tensile strength is used as a label to establish a finite element model of the macroscopic structure and train the macroscopic model.
[0029] Furthermore, the specific logic for constructing the hole expansion rate prediction model is as follows:
[0030] Perform feature extraction on the above multi-scale model, splice the input features and output labels of the microscopic model, each row represents a sample, each column represents a feature, obtain the microscopic feature matrix, and similarly obtain the mesoscopic feature matrix and macroscopic feature matrix. Combine the microscopic feature matrix, mesoscopic feature matrix and macroscopic feature matrix into a fusion feature matrix, and select The standardized processing is performed, and the processed fusion feature matrix is used as input. The hole expansion rate of the corresponding material is used as the label. The support vector machine is used as the machine learning model for training. The hyperparameters of the support vector machine are optimized using the grid search method. The mean square error and determination coefficient are used to evaluate the model performance. When the mean square error is In the range, the coefficient of determination is When it is within the range, it indicates that the model training is completed, and then the test set is input into the hole expansion rate prediction model for testing.
[0031] Furthermore, the specific logic of applying the hole expansion rate prediction model is as follows:
[0032] The relevant experimental data of the composite material to be predicted is obtained, and the data is input into the multi-scale model after preprocessing, and the features of the input and output data of the multi-scale model are extracted to obtain a fused feature matrix, which is input into the hole expansion rate prediction model to obtain the predicted hole expansion rate of the composite material.
[0033] Furthermore, the specific logic for obtaining the predicted value of the hole expansion rate at the current moment is:
[0034] By analyzing the predicted hole expansion rate output by the model, the formula for obtaining the predicted value of the hole expansion rate at the current moment is:
[0035] ;
[0036] in, is the predicted value of hole expansion rate at the current moment, is the temperature of the composite material at the current moment, is the predicted hole expansion rate output by the model, is the weight coefficient, and , is a constant term.
[0037] The present invention further provides a system for establishing a multi-scale prediction model for the pore expansion rate of a composite material. The system for establishing a multi-scale prediction model for the pore expansion rate of a composite material is used to implement the method for establishing the multi-scale prediction model for the pore expansion rate of a composite material, comprising:
[0038] Data acquisition module, used to collect experimental data of composite materials, including material type, volume, density, temperature, stress, strain, elastic modulus, Poisson's ratio, microstructure information and pore expansion rate, and perform preprocessing operations on the collected data;
[0039] Multi-scale model building module, used to analyze the experimental data of pre-treated composite materials and build multi-scale models, which are microscopic model, mesoscopic model and macroscopic model;
[0040] A hole expansion rate prediction model construction module is used to construct a hole expansion rate prediction model based on a machine learning model, fuse the input and output of the multi-scale model to form a fusion feature matrix, and use the fusion feature matrix as the input of the hole expansion rate prediction model, and use the corresponding hole expansion rate as a label to train the hole expansion rate prediction model;
[0041] A hole expansion rate prediction module is used to obtain relevant experimental data of the composite material to be predicted, and input it into the trained multi-scale model and hole expansion rate prediction model to obtain the predicted hole expansion rate;
[0042] The prediction result analysis module is used to analyze the output of the hole expansion rate prediction model and the temperature of the current material to obtain the hole expansion rate of the composite material to be predicted at the current moment.
[0043] In the above technical solution, the technical effects and advantages provided by the present invention are:
[0044] The method for establishing the multi-scale prediction model of the pore expansion rate of composite materials has significant technical effects and advantages. It can build a multi-scale model by combining micro, meso and macro levels through systematic data collection and preprocessing. By analyzing the multi-scale model, a machine learning model is constructed to obtain the predicted pore expansion rate, and then combined with the real-time temperature of the material, the pore expansion rate of the composite material is accurately predicted. This method not only improves the accuracy and reliability of the prediction, but also realizes the deep learning and self-optimization of the behavior of complex materials through the application of machine learning algorithms. In addition, the model flexibly considers the performance changes of materials under different environmental conditions, has real-time and dynamic adaptability, and provides a scientific basis for the design and application of composite materials, thereby promoting the progress of materials science and engineering technology. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] Figure 1 It is a schematic diagram of the overall method flow of the present invention;
[0046] Figure 2 Schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION
[0047] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.
[0048] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0049] Example:
[0050] See also Figure 1 , the present invention provides a technical solution:
[0051] The method for establishing a multi-scale prediction model for the pore expansion rate of composite materials includes the following specific steps:
[0052] Step 1: Collect experimental data of composite materials, including material type, volume, density, temperature, stress, strain, elastic modulus, Poisson's ratio, microstructure information and pore expansion rate, and pre-process the collected data;
[0053] In this embodiment, the specific logic for collecting experimental data of materials is as follows:
[0054] Acquire experimental data of various composite materials, record the material type, volume, density, temperature, stress, strain, elastic modulus, Poisson's ratio, microstructure information and pore expansion rate of each composite material, and record the area of the composite material during the experiment, the force during the tensile test and the shear modulus during the shear test, calculate the mesoscopic stress and tensile strength, and the microstructure information includes the diameter and length of the fiber, the diameter and length of the pores, and the contact force between the fibers. Normalize the collected data and store the processed data as a sample data set;
[0055] The stiffness matrix of the composite material is constructed by analyzing the elastic modulus, Poisson's ratio and the shear modulus during the experiment. The formula for obtaining the mesoscopic stress is:
[0056] ;
[0057] in, represents the mesoscopic stress, represents the mesoscopic strain, is the stiffness matrix of the composite material, is the volume;
[0058] The formula for obtaining the tensile strength is:
[0059] ;
[0060] in, is the tensile strength, is the force, is the force bearing area.
[0061] By recording the basic characteristics of the material (such as type, volume, density, etc.) and mechanical properties (such as stress, strain, elastic modulus, etc.), the necessary basic data is provided for the establishment of subsequent models. Collecting microstructural information (such as fiber diameter, length, pore characteristics, etc.) can reveal the complexity of the material and help scientists and engineers better understand the key factors affecting the pore expansion rate, so as to make more accurate predictions. Normalization processing, normalization preprocessing data helps to eliminate the influence of different dimensions and scales, making various features more comparable and stable in model training, thereby improving the model effect.
[0062] Step 2: Analyze the experimental data of the pretreated composite material and construct a multi-scale model, which includes a microscopic model, a mesoscopic model and a macroscopic model;
[0063] In this embodiment, the specific logic for constructing the multi-scale model is as follows:
[0064] The sample data set is divided into a training set and a test set with a ratio of 7:3, which are used to build microscopic models, mesoscopic models and macroscopic models respectively. The logic for building the microscopic model is as follows:
[0065] Input the material's density, temperature, elastic modulus and microstructure information into the micromodel, use molecular dynamics software to calculate, and use the corresponding porosity, stress, strain, and contact force between fibers as labels to train the micromodel;
[0066] The logic behind constructing the mesoscopic model is:
[0067] The output of the microscopic model is standardized, and the porosity and contact force between fibers output by the microscopic model are stored as eigenvalues of the mesoscopic model. A feature matrix is constructed, in which each row represents a sample and each column represents an eigenvalue. The feature matrix is checked for consistency and rationality and input into the mesoscopic model. The elastic modulus, mesoscopic stress, and temperature of the corresponding material are used as labels to train the mesoscopic model.
[0068] The logic behind building the macro model is:
[0069] The elastic modulus, mesoscopic stress, and temperature output by the mesoscopic model are used as material property inputs of the macroscopic model, and the corresponding tensile strength is used as a label to establish a finite element model of the macroscopic structure and train the macroscopic model.
[0070] The performance of composite materials is often affected by their microscopic, mesoscopic and macroscopic structures. By constructing a multiscale model, the complexity of material behavior can be fully captured, making the prediction results more accurate and reliable. For example, the microscopic model can reveal the interaction between fibers, while the mesoscopic model and the macroscopic model help to understand how these microscopic characteristics affect the overall material performance. Dividing the data set into a training set and a test set with a ratio of 7:3 can ensure the effectiveness of model training and the accuracy of evaluation. An effective training set can help the model better learn the characteristics of the material, while an independent test set can test the generalization ability of the model and avoid overfitting. By using molecular dynamics software to calculate the microscopic model, fine-grained material behavior information, including porosity, stress and strain, can be obtained. This method can provide more accurate microscopic data, thereby providing solid data support for subsequent model construction. Check whether the feature matrix constructed by the microscopic model data is consistent with the microscopic model data, and ensure that the data in the feature matrix are within the normal range. Standardize and check the consistency of the microscopic model output to ensure the rationality of the feature matrix, which will help improve the stability and accuracy of the subsequent model. The effectiveness of the eigenvalue directly affects the training effect of the mesoscopic and macroscopic models. By gradually constructing microscopic, mesoscopic and macroscopic models, a systematic understanding from microscopic to macroscopic can be achieved. This hierarchical modeling approach allows each model to focus on specific physical phenomena while forming overall consistency through data transfer. Applying the finite element method in the macroscopic model can make full use of existing engineering analysis tools and improve the implementation effect of the model. The macroscopic model can not only provide important predictions for the tensile strength of the material, but also provide a scientific basis for the design and optimization in practical engineering applications. Through the combination of multi-scale models, the behavior of the material at different levels can be better captured, thereby improving the accuracy of the prediction of the hole expansion rate, which is of great significance for optimizing the design of composite materials, improving production processes and improving material properties.
[0071] Step 3: Build a hole expansion rate prediction model based on the machine learning model, fuse the input and output of the multi-scale model to form a fusion feature matrix, and use the fusion feature matrix as the input of the hole expansion rate prediction model, and use the corresponding hole expansion rate as a label to train the hole expansion rate prediction model;
[0072] In this embodiment, the specific logic for constructing the hole expansion rate prediction model is:
[0073] Perform feature extraction on the above multi-scale model, splice the input features and output labels of the microscopic model, each row represents a sample, each column represents a feature, obtain the microscopic feature matrix, and similarly obtain the mesoscopic feature matrix and macroscopic feature matrix. Combine the microscopic feature matrix, mesoscopic feature matrix and macroscopic feature matrix into a fusion feature matrix, and select The standardized processing is performed, and the processed fusion feature matrix is used as input. The hole expansion rate of the corresponding material is used as the label. The support vector machine is used as the machine learning model for training. The hyperparameters of the support vector machine are optimized using the grid search method. The mean square error and determination coefficient are used to evaluate the model performance. When the mean square error is In the range, the coefficient of determination is When it is within the range, it indicates that the model training is completed, and then the test set is input into the hole expansion rate prediction model for testing.
[0074] The multi-scale model integrates information at the microscopic, mesoscopic and macroscopic levels, making the features related to the hole expansion rate very complex. These complex nonlinear relationships can be effectively captured through machine learning models. Machine learning methods can automatically identify patterns based on large-scale data sets and predict the hole expansion rate. This data-driven approach can reduce reliance on manual experience and improve the accuracy and reliability of predictions. By merging the features of models at different levels into fused features, the advantages of models at each level can be utilized to improve the overall performance of the model. Information extracted at different scales may contain complementary knowledge, which helps to improve the accuracy of hole expansion rate prediction. Through feature extraction and fusion, the support vector machine can better learn the potential patterns of the hole expansion rate, thereby improving prediction accuracy. Use Standardized processing features can eliminate the dimensional effects between different features, ensure the stability of the model training process, and improve the model convergence speed. Grid search can systematically explore the hyperparameter space and find the optimal hyperparameter combination, thereby improving the performance of the support vector machine model. Using mean square error and determination coefficient as evaluation indicators, the prediction performance of the model can be quantified to ensure that it meets the needs of practical applications. After training, the generalization ability of the model is verified through the test set to ensure the prediction performance of the model on unseen data and improve the credibility of practical applications.
[0075] Step 4: Obtain relevant experimental data of the composite material to be predicted, and input it into the trained multi-scale model and pore expansion rate prediction model to obtain the predicted pore expansion rate;
[0076] In this embodiment, the specific logic of applying the hole expansion rate prediction model is:
[0077] The relevant experimental data of the composite material to be predicted is obtained, and the data is input into the multi-scale model after preprocessing, and the features of the input and output data of the multi-scale model are extracted to obtain a fused feature matrix, which is input into the hole expansion rate prediction model to obtain the predicted hole expansion rate of the composite material.
[0078] Step 5: Analyze the output of the hole expansion rate prediction model and the temperature of the current material to obtain the hole expansion rate of the composite material to be predicted at the current moment.
[0079] In this embodiment, the specific logic for obtaining the predicted value of the hole expansion rate at the current moment is:
[0080] By analyzing the predicted hole expansion rate output by the model, the formula for obtaining the predicted value of the hole expansion rate at the current moment is:
[0081] ;
[0082] in, is the predicted value of hole expansion rate at the current moment, is the temperature of the composite material at the current moment, is the predicted hole expansion rate output by the model, is the weight coefficient, and , is a constant term.
[0083] Temperature is an important factor affecting the performance of composite materials. It directly affects the mechanical properties and hole expansion behavior of the material. Therefore, incorporating temperature into the hole expansion rate calculation can improve the accuracy of the prediction. In practical applications, composite materials may be used under different temperature conditions, so it is necessary to obtain the predicted value of the hole expansion rate at the current moment in real time to adapt to dynamically changing environmental conditions. This step allows real-time hole expansion rate prediction under different temperature conditions during the use of the material, providing more timely decision support, and using linear models to analyze the relationship between the hole expansion rate, temperature and predicted values, making the model easy to understand and apply, and facilitating engineers to make decisions. Taking the current material temperature into account can help improve the accuracy of the hole expansion rate prediction and avoid prediction deviations caused by ignoring the temperature factor. Here the weight coefficient is set to It can help maintain the stability of the model and prevent the drastic fluctuation of the hole expansion rate prediction caused by the drastic change of instantaneous temperature. Although the effect of temperature on the hole expansion rate of the composite material is taken into account here, the weight of temperature on the prediction is still lower than the weight of the model prediction. Therefore, it is set here. By setting The weight coefficients are normalized to maintain the rationality and consistency of the prediction results, which means that no matter how the input temperature or the hole expansion rate output by the model changes, the prediction results will be within a controllable range.
[0084] See also Figure 2 The present invention further provides a system for establishing a multi-scale prediction model for the pore expansion rate of a composite material. The system for establishing a multi-scale prediction model for the pore expansion rate of a composite material is used to implement the method for establishing the multi-scale prediction model for the pore expansion rate of a composite material, comprising:
[0085] Data acquisition module, used to collect experimental data of composite materials, including material type, volume, density, temperature, stress, strain, elastic modulus, Poisson's ratio, microstructure information and pore expansion rate, and perform preprocessing operations on the collected data;
[0086] Multi-scale model building module, used to analyze the experimental data of pre-treated composite materials and build multi-scale models, which are microscopic model, mesoscopic model and macroscopic model;
[0087] A hole expansion rate prediction model construction module is used to construct a hole expansion rate prediction model based on a machine learning model, fuse the input and output of the multi-scale model to form a fusion feature matrix, and use the fusion feature matrix as the input of the hole expansion rate prediction model, and use the corresponding hole expansion rate as a label to train the hole expansion rate prediction model;
[0088] A hole expansion rate prediction module is used to obtain relevant experimental data of the composite material to be predicted, and input it into the trained multi-scale model and hole expansion rate prediction model to obtain the predicted hole expansion rate;
[0089] The prediction result analysis module is used to analyze the output of the hole expansion rate prediction model and the temperature of the current material to obtain the hole expansion rate of the composite material to be predicted at the current moment.
[0090] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.
[0091] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.
[0092] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0093] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.
Claims
1. A method for establishing a multi-scale prediction model for the pore expansion rate of composite materials, characterized in that: The specific steps include: Step 1: Collect experimental data of composite materials, including material type, volume, density, temperature, stress, strain, elastic modulus, Poisson's ratio, microstructure information and pore expansion rate, and pre-process the collected data; Step 2: Analyze the experimental data of the pretreated composite material and construct a multi-scale model, which includes a microscopic model, a mesoscopic model and a macroscopic model; Step 3: Build a hole expansion rate prediction model based on the machine learning model, fuse the input and output of the multi-scale model to form a fusion feature matrix, and use the fusion feature matrix as the input of the hole expansion rate prediction model, and use the corresponding hole expansion rate as a label to train the hole expansion rate prediction model; Step 4: Obtain relevant experimental data of the composite material to be predicted, and input it into the trained multi-scale model and pore expansion rate prediction model to obtain the predicted pore expansion rate; Step 5: Analyze the output of the hole expansion rate prediction model and the temperature of the current material to obtain the hole expansion rate of the composite material to be predicted at the current moment; Acquire experimental data of various composite materials, record the material type, volume, density, temperature, stress, strain, elastic modulus, Poisson's ratio, microstructure information and pore expansion rate of each composite material, and record the area of the composite material during the experiment, the force during the tensile test and the shear modulus during the shear test, calculate the mesoscopic stress and tensile strength, and the microstructure information includes the diameter and length of the fiber, the diameter and length of the pores, and the contact force between the fibers. Normalize the collected data and store the processed data as a sample data set; The stiffness matrix of the composite material is constructed by analyzing the elastic modulus, Poisson's ratio and the shear modulus during the experiment. The formula for obtaining the mesoscopic stress is: in, represents the mesoscopic stress, represents the mesoscopic strain, is the stiffness matrix of the composite material, is the volume; The formula for obtaining the tensile strength is: in, is the tensile strength, is the force, is the force-bearing area; The sample data set is divided into a training set and a test set with a ratio of 7:3, which are used to build microscopic models, mesoscopic models and macroscopic models respectively. The logic for building the microscopic model is as follows: Input the material's density, temperature, elastic modulus and microstructure information into the micromodel, use molecular dynamics software to calculate, and use the corresponding porosity, stress, strain, and contact force between fibers as labels to train the micromodel; The logic behind constructing the mesoscopic model is: The output of the microscopic model is standardized, and the porosity and contact force between fibers output by the microscopic model are stored as eigenvalues of the mesoscopic model. A feature matrix is constructed, in which each row represents a sample and each column represents an eigenvalue. The feature matrix is checked for consistency and rationality and input into the mesoscopic model. The elastic modulus, mesoscopic stress, and temperature of the corresponding material are used as labels to train the mesoscopic model. The logic behind building the macro model is: The elastic modulus, mesoscopic stress, and temperature output by the mesoscopic model are used as the material property inputs of the macroscopic model, and the corresponding tensile strength is used as a label to establish a finite element model of the macroscopic structure and train the macroscopic model. Perform feature extraction on the above multi-scale model, splice the input features and output labels of the microscopic model, each row represents a sample, each column represents a feature, obtain the microscopic feature matrix, and similarly obtain the mesoscopic feature matrix and macroscopic feature matrix. Combine the microscopic feature matrix, mesoscopic feature matrix and macroscopic feature matrix into a fusion feature matrix, and select The standardized processing is performed, and the processed fusion feature matrix is used as input. The hole expansion rate of the corresponding material is used as the label. The support vector machine is used as the machine learning model for training. The hyperparameters of the support vector machine are optimized using the grid search method. The mean square error and determination coefficient are used to evaluate the model performance. When the mean square error is In the range, the coefficient of determination is When it is within the range, it indicates that the model training is completed, and then the test set is input into the hole expansion rate prediction model for testing; Obtain relevant experimental data of the composite material to be predicted, input the data into the multi-scale model after preprocessing, perform feature extraction on the input and output data of the multi-scale model, obtain a fusion feature matrix, input the fusion feature matrix into the hole expansion rate prediction model, and obtain the predicted hole expansion rate of the composite material; By analyzing the predicted hole expansion rate output by the model, the formula for obtaining the predicted value of the hole expansion rate at the current moment is: in, is the predicted value of hole expansion rate at the current moment, is the temperature of the composite material at the current moment, is the predicted hole expansion rate output by the model, is the weight coefficient, and , is a constant term.
2. A method for establishing a multi-scale prediction model for the pore expansion rate of composite materials, characterized in that: The system for establishing a multi-scale prediction model for the pore expansion rate of a composite material is used to implement the method for establishing a multi-scale prediction model for the pore expansion rate of a composite material according to claim 1, comprising: Data acquisition module, used to collect experimental data of composite materials, including material type, volume, density, temperature, stress, strain, elastic modulus, Poisson's ratio, microstructure information and pore expansion rate, and perform preprocessing operations on the collected data; Multi-scale model building module, used to analyze the experimental data of pre-treated composite materials and build multi-scale models, which are microscopic model, mesoscopic model and macroscopic model; A hole expansion rate prediction model construction module is used to construct a hole expansion rate prediction model based on a machine learning model, fuse the input and output of the multi-scale model to form a fusion feature matrix, and use the fusion feature matrix as the input of the hole expansion rate prediction model, and use the corresponding hole expansion rate as a label to train the hole expansion rate prediction model; A hole expansion rate prediction module is used to obtain relevant experimental data of the composite material to be predicted, and input it into the trained multi-scale model and hole expansion rate prediction model to obtain the predicted hole expansion rate; The prediction result analysis module is used to analyze the output of the hole expansion rate prediction model and the temperature of the current material to obtain the hole expansion rate of the composite material to be predicted at the current moment.