Epoxy material performance prediction method, system, equipment and medium

Abstract

An epoxy material performance prediction method, system, device and medium, the method comprising: acquiring formula data of an epoxy material, the formula data comprising epoxy resin series matrix molecular structure data, epoxy resin curing agent category data and catalyst category data; predicting the formula data based on a pre-trained epoxy material performance prediction model to obtain predicted process parameter data; the predicted process parameter data comprises the band gap width, dielectric constant or glass transition temperature of the epoxy material; the pre-trained epoxy material performance prediction model is obtained by training a prediction model by taking formula data and process parameters as training data, and the prediction model comprises a feedforward neural network model. According to the method, the influence of formula characteristics and process parameters in the epoxy material is fully considered, and the epoxy material performance prediction model is constructed, so that high-precision quantitative prediction of the electronic energy gap, dielectric constant or glass transition temperature of the epoxy material is realized, and meanwhile, the calculation amount and the test amount are effectively reduced.

Description

Technical Field

[0001] This invention relates to the field of material performance prediction technology, and in particular to a method, system, device and medium for predicting the performance of epoxy materials. Background Technology

[0002] Epoxy materials are widely used in electronics, electrical engineering, and composite materials due to their excellent mechanical strength, chemical stability, and superior insulation properties. The demand for high-field-strength insulation materials in applications such as power system transformers and power electronic devices places increasingly higher demands on the performance of epoxy materials, requiring wider band gaps to improve breakdown strength. Finding epoxy materials with wide band gaps and high breakdown strength is crucial for efficient performance prediction and screening of epoxy materials and for guiding the design of electrical insulation materials.

[0003] Traditional testing of the electrical properties of epoxy materials relies heavily on extensive experimental measurements. However, this approach is hampered by long testing cycles, high costs, and complex sample preparation, making it difficult to promptly identify formulations that meet specific insulation requirements and to systematically analyze the complex interrelationships between components, processes, and properties. In recent years, with the development of computational materials science and materials genome engineering, some physical model-based simulation methods have been used to study the relationship between the microstructure and properties of materials. However, these methods involve enormous computational demands for complex polymer systems, making it difficult to simulate the electrical insulation properties at the macroscopic scale or under different formulation or process conditions. Summary of the Invention

[0004] To address the problem that existing methods involve enormous computational costs for complex polymer systems, making it difficult to simulate electrical insulation properties at the macroscopic scale or under different formulations or process conditions, this invention provides a method for predicting the properties of epoxy materials. This method includes: Obtain formulation data for epoxy materials, including epoxy resin series matrix molecular structure data, epoxy resin curing agent category data, and catalyst category data; The formulation data is predicted based on a pre-trained epoxy material performance prediction model to obtain predicted process parameter data. The predicted process parameter data includes the epoxy material bandgap width, dielectric constant, or glass transition temperature. The pre-trained epoxy material performance prediction model is obtained by training the prediction model with formulation data and process parameters as training data. The prediction model includes a feedforward neural network model.

[0005] Preferably, the training of the epoxy material performance prediction model includes: Based on the total amount of training data in the training set and the number of features in the training data, the initial parameters of the epoxy material performance prediction model to be trained are set, including the number of hidden layer neurons. Obtain the formulation data of the epoxy material and the corresponding process parameters; The formulation data in the training set is input into the epoxy material performance prediction model to be trained, and the predicted process parameter data is output. According to a predetermined loss function, the predicted process parameter data and the process parameter data corresponding to the formula data in the training set are calculated to determine the loss function value; The initial parameters of the epoxy material performance prediction model to be trained are adjusted based on the loss function value to obtain the updated epoxy material performance prediction model. The trained and updated epoxy material performance prediction model is evaluated using a test set to obtain evaluation indices; and the optimal parameters of the epoxy material performance prediction model are determined based on the evaluation indices to obtain the trained epoxy material performance prediction model.

[0006] Preferably, after obtaining the epoxy material formulation data and the corresponding process parameters, the method further includes: The formula data and the corresponding process parameters are preprocessed based on at least one of invalid sample removal, feature dimensionality reduction, outlier handling, and data standardization to obtain preprocessed training data.

[0007] Preferably, the predetermined loss function includes: mean squared error, mean absolute error, cross-entropy loss, logarithmic loss, and / or exponential loss.

[0008] Preferably, the step of evaluating the trained and updated epoxy material performance prediction model using a test set to obtain evaluation indicators; and determining the optimal parameters of the epoxy material performance prediction model based on the evaluation indicators to obtain the trained epoxy material performance prediction model, includes: The formulation data from the test set are substituted into the updated epoxy material performance prediction model to obtain the predicted process parameters. Based on the process parameters corresponding to the formula data in the test set and the predicted process parameters, evaluation indicators are calculated, including: accuracy, precision, recall, mean squared error, mean absolute error, and coefficient of determination. Based on the evaluation index values, an optimization algorithm is used to find the parameter values ​​that make the epoxy material performance prediction model perform optimally. The epoxy material performance prediction model corresponding to the optimal parameters is used as the trained epoxy material performance prediction model.

[0009] Preferably, the step of predicting the formulation data based on a pre-trained epoxy material performance prediction model to obtain predicted process parameter data further includes: In the process of predicting the formulation data based on the trained epoxy material performance prediction model, the contribution value of the formulation data is determined. When the contribution value of the formula data is greater than or equal to a preset threshold, the formula data is taken as the optimal data.

[0010] Preferably, in the process of predicting the formulation data based on the trained epoxy material performance prediction model, determining the contribution value of the formulation data includes: Based on each piece of data in the formula data, a first set and a second set are determined; wherein, the first set is a subset of the remaining data in the formula data excluding the current data; and the second set is composed of the first set and the current data. The value of the first set and the value of the second set are calculated based on a preset feature function; the value of the second set is subtracted from the value of the first set to obtain the change. The weight of the first set is determined by multiplying the factorial of the number of elements in the first set by the factorial of the difference between the total number of data and the number of elements in the first set minus one, and then dividing by the factorial of the total number of data. The weights of the first set and the change amount are weighted and combined to obtain the contribution value of each data point in the target data.

[0011] Preferably, the contribution value is calculated using the following formula: , in, Representative in addition to the first Summation of the subset of data other than the given data; The contribution value represents... The number of elements in the first set. The representation does not include the first The first set of data, This indicates that the second set consists of all the data in the target data. The value of the first set, For the value of the second set, The amount of data for the formula. } represents all data in the recipe data other than the current data.

[0012] Preferably, after obtaining the formulation data of the epoxy material, the method further includes: The formulation data is preprocessed by at least one of invalid sample removal, feature dimensionality reduction, outlier handling, and data standardization to obtain preprocessed formulation data.

[0013] Furthermore, this application also provides an epoxy material performance prediction system, comprising: The data acquisition module is used to acquire the formulation data of epoxy materials, including the molecular structure data of epoxy resin series matrix, the category data of epoxy resin curing agent, and the category data of catalyst. The prediction module is used to predict the formulation data based on a pre-trained epoxy material performance prediction model to obtain the predicted process parameter data. The predicted process parameter data includes the epoxy material bandgap width, dielectric constant, or glass transition temperature. The pre-trained epoxy material performance prediction model is obtained by training the prediction model with formulation data and process parameters as training data. The prediction model includes a feedforward neural network model.

[0014] Preferably, it further includes a training module for: Based on the total amount of training data in the training set and the number of features in the training data, the initial parameters of the epoxy material performance prediction model to be trained are set, including the number of hidden layer neurons. Obtain the formulation data of the epoxy material and the corresponding process parameters; The formulation data in the training set is input into the epoxy material performance prediction model to be trained, and the predicted process parameter data is output. According to a predetermined loss function, the predicted process parameter data and the process parameter data corresponding to the formula data in the training set are calculated to determine the loss function value; The initial parameters of the epoxy material performance prediction model to be trained are adjusted based on the loss function value to obtain the updated epoxy material performance prediction model. The trained and updated epoxy material performance prediction model is evaluated using a test set to obtain evaluation indices; and the optimal parameters of the epoxy material performance prediction model are determined based on the evaluation indices to obtain the trained epoxy material performance prediction model.

[0015] Preferably, the training module is further configured to: preprocess the formulation data and the process parameters corresponding to the formulation data based on at least one of invalid sample removal, feature dimensionality reduction, outlier handling, and data standardization to obtain preprocessed training data.

[0016] Preferably, the predetermined loss function includes: mean squared error, mean absolute error, cross-entropy loss, logarithmic loss, and / or exponential loss.

[0017] Preferably, the training module uses a test set to evaluate the trained and updated epoxy material performance prediction model to obtain evaluation indicators; and determines the optimal parameters of the epoxy material performance prediction model based on the evaluation indicators, thus obtaining the trained epoxy material performance prediction model. The implementation steps include: The formulation data from the test set are substituted into the updated epoxy material performance prediction model to obtain the predicted process parameters. Based on the process parameters corresponding to the formula data in the test set and the predicted process parameters, evaluation indicators are calculated, including: accuracy, precision, recall, mean squared error, mean absolute error, and coefficient of determination. Based on the evaluation index values, an optimization algorithm is used to find the parameter values ​​that make the epoxy material performance prediction model perform optimally. The epoxy material performance prediction model corresponding to the optimal parameters is used as the trained epoxy material performance prediction model.

[0018] Preferably, the system further includes a data selection module; the data selection module includes: The contribution value calculation submodule is used to determine the contribution value of the formulation data during the process of predicting the formulation data based on the trained epoxy material performance prediction model. The optimal judgment submodule is used to determine the optimal data when the contribution value of the formula data is greater than or equal to a preset threshold.

[0019] Preferably, the contribution value calculation submodule is specifically used for: Based on each piece of data in the formula data, a first set and a second set are determined; wherein, the first set is a subset of the remaining data in the formula data excluding the current data; and the second set is composed of the first set and the current data. The value of the first set and the value of the second set are calculated based on a preset feature function; the value of the second set is subtracted from the value of the first set to obtain the change. The weight of the first set is determined by multiplying the factorial of the number of elements in the first set by the factorial of the difference between the total number of data and the number of elements in the first set minus one, and then dividing by the factorial of the total number of data. The weights of the first set and the change amount are weighted and combined to obtain the contribution value of each data point in the target data.

[0020] In another aspect, this application also provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, an epoxy material performance prediction method as described above is implemented.

[0021] Furthermore, this application also provides a readable storage medium having an executable program stored thereon, which, when executed, implements the epoxy material performance prediction method as described above.

[0022] Compared with the prior art, the beneficial effects of this application are as follows: This application provides a method for predicting the performance of epoxy materials, comprising: acquiring epoxy material formulation data, wherein the formulation data includes epoxy resin series matrix molecular structure data, epoxy resin curing agent category data, and catalyst category data; predicting the formulation data based on a pre-trained epoxy material performance prediction model to obtain predicted process parameter data; wherein the predicted process parameter data includes the epoxy material bandgap width, dielectric constant, or glass transition temperature; the pre-trained epoxy material performance prediction model is obtained by training the prediction model using the formulation data and process parameters as training data, and the prediction model includes a feedforward neural network model. This technical solution fully considers the characteristics of key components in epoxy materials and the influence of process parameters, constructing a reasonable epoxy material performance prediction model to achieve high-precision quantitative prediction of the electronic bandgap, dielectric constant, or glass transition temperature of epoxy materials, while effectively reducing computational and experimental workload. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is an overall flowchart of a method for predicting the performance of epoxy materials according to this application; Figure 2 This is a flowchart of the epoxy material performance prediction method provided in Embodiment 2 of this application; Figure 3 This is a schematic diagram of the epoxy material performance prediction process provided in Embodiment 3 of this application; Figure 4 This is a flowchart of the epoxy material performance prediction process provided in Embodiment 4 of this application; Figure 5 This is a schematic diagram of an electronic device structure according to this application. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of this application.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] Example 1 This application provides a method for predicting the properties of epoxy materials, such as... Figure 1 As shown, the method includes: Step 1: Obtain the formulation data of the epoxy material, which includes the molecular structure data of the epoxy resin series matrix, the type data of the epoxy resin curing agent, and the type data of the catalyst; Step 2: Based on the pre-trained epoxy material performance prediction model, predict the formulation data to obtain the predicted process parameter data; The predicted process parameter data includes the epoxy material bandgap width, dielectric constant, or glass transition temperature. The pre-trained epoxy material performance prediction model is obtained by training the prediction model with formulation data and process parameters as training data. The prediction model includes a feedforward neural network model.

[0028] Preferably, the training of the epoxy material performance prediction model includes: Based on the total amount of training data in the training set and the number of features in the training data, the initial parameters of the epoxy material performance prediction model to be trained are set, including the number of hidden layer neurons. Obtain the formulation data of the epoxy material and the corresponding process parameters; The formulation data in the training set is input into the epoxy material performance prediction model to be trained, and the predicted process parameter data is output. According to a predetermined loss function, the predicted process parameter data and the process parameter data corresponding to the formula data in the training set are calculated to determine the loss function value; The initial parameters of the epoxy material performance prediction model to be trained are adjusted based on the loss function value to obtain the updated epoxy material performance prediction model. The trained and updated epoxy material performance prediction model is evaluated using a test set to obtain evaluation indices; and the optimal parameters of the epoxy material performance prediction model are determined based on the evaluation indices to obtain the trained epoxy material performance prediction model.

[0029] Preferably, after obtaining the epoxy material formulation data and the corresponding process parameters, the method further includes: The formula data and the corresponding process parameters are preprocessed based on at least one of invalid sample removal, feature dimensionality reduction, outlier handling, and data standardization to obtain preprocessed training data.

[0030] Preferably, the predetermined loss function includes: mean squared error, mean absolute error, cross-entropy loss, logarithmic loss, and / or exponential loss.

[0031] Preferably, the step of evaluating the trained and updated epoxy material performance prediction model using a test set to obtain evaluation indicators; and determining the optimal parameters of the epoxy material performance prediction model based on the evaluation indicators to obtain the trained epoxy material performance prediction model, includes: The formulation data from the test set are substituted into the updated epoxy material performance prediction model to obtain the predicted process parameters. The evaluation index is calculated between the process parameters corresponding to the formula data in the test set and the predicted process parameters. The evaluation index includes: accuracy, precision, recall, mean squared error, mean absolute error, and coefficient of determination. Based on the evaluation index values, an optimization algorithm is used to find the parameter values ​​that make the epoxy material performance prediction model perform optimally. The epoxy material performance prediction model corresponding to the optimal parameters is used as the trained epoxy material performance prediction model.

[0032] Preferably, the step of predicting the formulation data based on a pre-trained epoxy material performance prediction model to obtain predicted process parameter data further includes: In the process of predicting the formulation data based on the trained epoxy material performance prediction model, the contribution value of the formulation data is determined. When the contribution value of the formula data is greater than or equal to a preset threshold, the formula data is taken as the optimal data.

[0033] Preferably, in the process of predicting the formulation data based on the trained epoxy material performance prediction model, determining the contribution value of the formulation data includes: Based on each piece of data in the formula data, a first set and a second set are determined; wherein, the first set is a subset of the remaining data in the formula data excluding the current data; and the second set is composed of the first set and the current data. The value of the first set and the value of the second set are calculated based on a preset feature function; the value of the second set is subtracted from the value of the first set to obtain the change. The weight of the first set is determined by multiplying the factorial of the number of elements in the first set by the factorial of the difference between the total number of data and the number of elements in the first set minus one, and then dividing by the factorial of the total number of data. The weights of the first set and the change amount are weighted and combined to obtain the contribution value of each data point in the target data.

[0034] Preferably, the contribution value is calculated using the following formula: , in, Representative in addition to the first Summation of the subset of data other than the given data; The contribution value represents... The number of elements in the first set. The representation does not include the first The first set of data, This indicates that the second set consists of all the data in the target data. The value of the first set, For the value of the second set, The amount of data for the formula. } represents all data in the recipe data other than the current data.

[0035] Preferably, after obtaining the formulation data of the epoxy material, the method further includes: The formulation data is preprocessed by at least one of invalid sample removal, feature dimensionality reduction, outlier handling, and data standardization to obtain preprocessed formulation data.

[0036] Example 2 Figure 2 This is a flowchart of the epoxy material performance prediction method provided in Embodiment 1 of this application. This embodiment is applicable to predicting the band gap width of epoxy materials. In the embodiments of this application, the formulation data is referred to as target data. Figure 2 As shown, a method for predicting the properties of epoxy materials includes: S110. Obtain target data for epoxy materials, including epoxy resin series matrix molecular structure data, epoxy resin curing agent category data, and catalyst category data.

[0037] Epoxy materials are an important class of thermosetting resins, and epoxy resin is a key component of epoxy materials. Epoxy resin generally refers to organic polymers containing two or more epoxy groups in their molecules. Its molecular structure is characterized by the presence of reactive epoxy groups in the molecular chain; these epoxy groups can be located at the ends, in the middle, or in a cyclic structure of the molecular chain.

[0038] In this scheme, epoxy resin itself has a thermoplastic linear structure, requiring the addition of a curing agent to induce a cross-linking reaction, forming a three-dimensional network structure, thereby obtaining good physical and mechanical properties and chemical stability. The epoxy resin curing agent category data includes amine curing agent data, acid anhydride curing agent data, and latent curing agent data.

[0039] In this embodiment, during the curing process of epoxy resin, the catalyst can accelerate the curing reaction, lower the curing temperature, shorten the curing time, and also have a certain impact on the properties of the cured product. Catalyst category data includes tertiary amine catalyst data and metal salt catalyst data.

[0040] Furthermore, data on the molecular structure of epoxy resin matrix, the types of epoxy resin curing agents, and the types of catalysts are of great significance for the formulation design, process optimization, and performance control of epoxy resin materials. By accurately grasping and rationally applying this data, epoxy resin composite materials that meet different application requirements can be prepared.

[0041] Specifically, data on the molecular structure of epoxy resin matrix, the types of epoxy resin curing agents, and the types of catalysts can be obtained from the database.

[0042] In this scheme, the target data may also include filler data, which includes alumina, silica, surface-treated boron nitride or boron nitride nanosheets, silicon carbide, etc. Combining epoxy resin matrix molecular structure data with a certain proportion of filler data can form composite materials. By using composite material data, epoxy resin curing agent category data, and catalyst category data as input, the performance of the composite material can be predicted.

[0043] Furthermore, target data may also include electrothermal parameters such as tensile strength (MPa), Young's modulus (MPa), dielectric loss, water absorption rate (%), and decomposition temperature. Tensile strength refers to the maximum tensile stress a material can withstand before tensile fracture, reflecting its ability to resist tensile failure. Young's modulus is the ratio of tensile stress to longitudinal linear strain during the elastic deformation stage, characterizing the material's stiffness (its ability to resist elastic deformation). Dielectric loss refers to the energy loss of a dielectric material in an alternating electric field due to polarization relaxation or conductivity. Water absorption rate refers to the percentage of water absorbed by a material under standard atmospheric conditions (or specific conditions), reflecting the material's hygroscopicity. Decomposition temperature is the temperature at which a material begins to chemically decompose when heated, usually referring to the temperature at which mass loss reaches a certain value.

[0044] In this embodiment, filler data, tensile strength, Young's modulus, dielectric loss, water absorption rate, decomposition temperature, etc., can be obtained from the database.

[0045] Optionally, after obtaining the target data for the epoxy material, the method further includes: The target data is preprocessed by at least one of invalid sample removal, feature dimensionality reduction, outlier handling, and data standardization to obtain preprocessed target data.

[0046] In this embodiment, invalid sample removal refers to deleting samples that lack key parameters, and also removing records with abnormal epoxy resin component information.

[0047] Feature dimensionality reduction refers to the process of transforming a high-dimensional feature space into a low-dimensional feature space without losing (or with minimal loss) data information. Its purpose is to reduce the dimensionality of the data, decrease the complexity of data processing, avoid overfitting, and improve the model's generalization ability and training efficiency. Based on expert experience and statistical analysis, the 20% most relevant variables are selected, while variables with extremely low frequency or high noise are removed.

[0048] In this approach, outliers are data points in the dataset that are significantly different from other data points and deviate from the overall distribution. The purpose of handling outliers is to prevent them from negatively impacting data analysis and model training, thereby making the model more robust and accurate. (Ref. 3) The principle is to eliminate samples with unreasonable parameter values.

[0049] Data standardization involves transforming data according to certain rules to give it a specific mean and standard deviation, or mapping the data to a specific interval. This eliminates the influence of different features due to differences in units and value ranges, making the data comparable and improving the convergence speed and accuracy of the model. Specifically, normalization methods are used to map each input variable to the [0,1] interval, eliminating the uneven influence of different units on the model.

[0050] Preprocessing the target data can improve data quality, enhance model performance, reduce training time, and increase model stability.

[0051] S120. Based on the epoxy material performance prediction model, predict the target data and output the predicted data; wherein, the predicted data includes the epoxy material band gap width, dielectric constant or glass transition temperature.

[0052] In this embodiment, the band gap width of the epoxy material refers to the energy difference between the valence band and the conduction band in the epoxy material, and is used to measure the conductivity of the epoxy material.

[0053] The dielectric constant reflects the polarization capability of a dielectric material under the influence of an electric field. Polarization refers to the rearrangement of charges in a dielectric material under the influence of an electric field, forming electric dipoles. The larger the dielectric constant, the higher the degree of polarization of the dielectric material under the same electric field, the more polarization charges are generated, and thus the stronger the weakening effect on the original electric field.

[0054] Furthermore, the glass transition temperature refers to the temperature at which an amorphous polymer (including the non-crystalline portion of a crystalline polymer) transitions from a glassy state to a highly elastic state (or vice versa). Below the glass transition temperature, the polymer exhibits glass-like properties, being hard and brittle; above the glass transition temperature, the polymer exhibits rubber-like elasticity, possessing good flexibility and deformability.

[0055] In this scheme, the epoxy material performance prediction model is a feedforward artificial neural network (ANN) prediction model, which is a machine learning model based on artificial neural networks. The input layer of the feedforward ANN prediction model receives various input features related to the target epoxy material data. The hidden layer consists of multiple neurons that transmit information through weighted connections. The hidden layer performs nonlinear transformations and feature extraction on the input information, mapping the input data to a higher-dimensional feature space to better capture the complex relationship between the input features and the epoxy material's bandgap width, dielectric constant, or glass transition temperature. The output layer typically has the number of neurons corresponding to the epoxy material's bandgap width, dielectric constant, or glass transition temperature to be predicted. Through training, the model adjusts the weights and thresholds in the network to ensure the output layer can predict the epoxy material's bandgap width as accurately as possible.

[0056] In this embodiment, the epoxy material performance prediction model can also be a kernel ridge regression (KRR), Gaussian process regression (GPR), support vector regression (SVR), or a model based on semantic embedding and attention mechanisms (Transformer architecture). KRR maps low-dimensional input data to a high-dimensional feature space using kernel functions (such as radial basis functions (RBF) or multinomial kernels), transforming linearly inseparable problems into linearly separable ones, and then fitting the data using ridge regression (L2-regularized linear regression). Gaussian process regression treats the regression problem as a stochastic process, assuming all data points follow a joint Gaussian distribution. It defines the similarity between data points using kernel functions and outputs the mean and variance of the predicted values ​​(i.e., uncertainty estimates). Support vector regression maps data to a high-dimensional space using kernel functions, searching for the optimal hyperplane that minimizes the error between sample points and predicted values ​​to a preset threshold, while controlling model complexity through regularization terms.

[0057] Specifically, epoxy material systems are more complex in terms of component design and processing technology. Their macroscopic structural properties are affected by various process parameters such as resin matrix, curing agent, and catalyst, and may also be dynamically changed along with process parameters such as curing reaction conditions. Therefore, by taking the molecular structure data of epoxy resin series matrix, epoxy resin curing agent type data and catalyst type data as input, a pre-trained epoxy material performance prediction model is used to predict the molecular structure data of epoxy resin series matrix, epoxy resin curing agent type data and catalyst type data, and output the band gap width, dielectric constant or glass transition temperature of epoxy material.

[0058] Furthermore, using the molecular structure data of epoxy resin matrix, the category data of epoxy resin curing agent, and the category data of catalyst as input, the pre-trained epoxy material performance prediction model is used to predict the molecular structure data of epoxy resin matrix, the category data of epoxy resin curing agent, and the category data of catalyst, and outputs electrical conductivity, thermal conductivity, etc.

[0059] Among them, the epoxy material performance prediction model can predict the electrical or thermodynamic properties of epoxy materials.

[0060] Optionally, the parameters of the epoxy material performance prediction model will differ depending on the predicted data.

[0061] In this scheme, the predicted data includes the band gap width, dielectric constant, or glass transition temperature of the epoxy material. When the predicted data is the band gap width of the epoxy material, the parameter of the epoxy material performance prediction model is the first parameter; when the predicted data is the dielectric constant, the parameter of the epoxy material performance prediction model is the second parameter; and when the predicted data is the glass transition temperature, the parameter of the epoxy material performance prediction model is the third parameter. The first, second, and third parameters are determined by training the epoxy material performance prediction model.

[0062] The technical solution of this invention acquires target data for epoxy materials, including molecular structure data of epoxy resin matrix, type data of epoxy resin curing agents, and type data of catalysts. Then, based on an epoxy material performance prediction model, the target data is predicted to output the band gap width, dielectric constant, or glass transition temperature of the epoxy material. By implementing this technical solution, the characteristics of key components in the epoxy material and the influence of process parameters are fully considered, and a reasonable epoxy material performance prediction model is constructed to achieve high-precision quantitative prediction of the electronic band gap, dielectric constant, or glass transition temperature of the epoxy material, while effectively reducing the computational and experimental workload.

[0063] Example 3 Figure 3 This is a schematic diagram of the epoxy material performance prediction process provided in Embodiment 3 of the present invention. The relationship between this embodiment and the above embodiments is a detailed description of the prediction process. Figure 3 As shown, a method for predicting the properties of epoxy materials includes: S210. In the process of predicting the target data based on the epoxy material performance prediction model, the contribution value of the target data is determined; wherein, the contribution value is used to characterize the degree of influence of the target data on the prediction result of the epoxy material performance prediction model.

[0064] In this scheme, the contribution value is used to characterize the degree of influence of the target data on the prediction results of the epoxy material performance prediction model. The magnitude of the contribution value is positively correlated with the degree of influence of the target data on the prediction results of the epoxy material performance prediction model; that is, the larger the contribution value, the more critical the role of the data in the model prediction process, and the greater the influence on the final prediction result.

[0065] Furthermore, the contribution value of the target data can be determined based on a pre-defined contribution value prediction model; alternatively, interpretability analysis tools can be used to calculate the contribution value of the target data. Interpretability analysis tools include SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model - agnostic Exlanations). SHAP (SHapley Additive exPlanations) is a game theory-based model interpretation method that helps understand the decision-making process of complex machine learning models by calculating the marginal contribution of each feature to the prediction result. Its core principle is to view model prediction as a cooperative game process, with each feature as a participant in the game. By calculating the average marginal contribution of each feature across all possible feature combinations (i.e., the change in the prediction result when a feature is added), the contribution of a feature to the prediction is quantified. LIME (Local Interpretable Model-agnostic Explanations) is a tool for interpreting the predictions of complex machine learning models. Its core idea is to reveal the decision-making logic of black-box models by building locally interpretable models. It supports the interpretation of predictions from various models and is particularly good at handling structured data such as images and text.

[0066] Optionally, determining the contribution value of the target data includes: For each piece of data in the target data, a first set and a second set are determined; wherein, the first set is a subset of the remaining data in the target data excluding the current data; and the second set consists of all the data in the target data. The value of the first set and the value of the second set are calculated based on a preset feature function; wherein the feature function is used to describe the value generated by each data point. Subtract the value of the second set from the value of the first set to obtain the change; Determine the weights of the first set; wherein the weights are used to describe the importance of the first set; The weights of the first set and the change amount are weighted and combined to obtain the contribution value of each data point in the target data.

[0067] In this scheme, the first set consists of the remaining data in the target data excluding the current data, and can be used as follows: Indicates, that is The representation does not include the first The first set is a feature subset of the target data; the second set consists of all the data in the target data, and can be used as... express.

[0068] In this embodiment, the feature function is used to describe the value generated by each data point, and can be used as follows: The value of the first set is... The value of the second set is .

[0069] Among them, the change is - .

[0070] Furthermore, the weight of the first set is calculated based on the amount of target data and the number of elements in the first set. The weight of the first set is... .in, The amount of data for the target data. The number of elements in the first set.

[0071] Specifically, the contribution value of each data point in the target data is calculated using the following formula. : ,in, Representative in addition to the first Summation of the subset of data other than the given data; As contribution value, The number of elements in the first set. The representation does not include the first The first set of data, This indicates that the second set consists of all the data in the target data. The value of the first set, For the value of the second set, The amount of data for the formula. } represents all data in the recipe data other than the current data.

[0072] Furthermore, scatter plots can be used to visually demonstrate the positive or negative impact of different data on the model's prediction results.

[0073] By applying interpretability analysis tools, we can deeply analyze the trained epoxy material performance prediction model, quantifying and visualizing the specific contributions of each component and process parameter to the band gap, dielectric constant, or glass transition temperature of epoxy materials. Validated and interpretable models can not only rapidly predict the performance of new formulation and process combinations, accelerating material screening, but also provide data-driven guidance for the rational design of epoxy materials with superior performance based on interpretability analysis results.

[0074] S220. If the contribution value of the target data is greater than or equal to a preset threshold, then the target data is taken as the optimal data; wherein, the optimal data includes epoxy resin series matrix molecular structure data, epoxy resin curing agent category data, and catalyst category data.

[0075] In this scheme, the preset threshold can be set based on the epoxy material preparation requirements.

[0076] Specifically, the optimal data, whose contribution value is greater than or equal to a preset threshold, can provide data-driven guidance for designing epoxy materials with excellent performance.

[0077] The technical solution of this application, in the process of predicting target data based on an epoxy material performance prediction model, determines the contribution value of the target data and selects the target data with a contribution value greater than or equal to a preset threshold as the optimal data. By implementing this technical solution, through in-depth analysis of the model prediction results, and using quantitative calculation and visualization techniques, the influence mechanism of each component and process parameter on the target performance is accurately revealed. The model's predictive capability is organically combined with interpretability analysis, and by quantifying the contribution of each factor and visualizing the influence path, a scientific basis is provided for the rational design of novel epoxy insulation materials. This can efficiently screen out the optimal combination of material formulations and process conditions, significantly improving material research and development efficiency and contributing to the development of high-performance epoxy insulation materials.

[0078] Example 4 Figure 4 The flowchart of the epoxy material performance prediction process provided in Embodiment 4 of this application, and the relationship between this embodiment and the above embodiments, is a detailed description of the epoxy material performance prediction model construction process. Figure 4 As shown, the method includes: S310. Obtain training data for the epoxy material performance prediction model to be trained.

[0079] In this scheme, the training data includes target data and process parameter data. The target data includes the molecular structure data of epoxy resin matrix, the type of epoxy resin curing agent and the type of catalyst, the band gap width of epoxy material, and the dielectric constant or glass transition temperature. The process parameter data includes the band gap width of epoxy material, and the dielectric constant or glass transition temperature.

[0080] In this embodiment, training data for the epoxy material performance prediction model to be trained can be obtained from a database.

[0081] Furthermore, after obtaining the training data for the epoxy material performance prediction model to be trained, the training data is preprocessed based on at least one of invalid sample removal, feature dimensionality reduction, outlier handling, and data standardization to obtain preprocessed training data.

[0082] S320. The epoxy material performance prediction model to be trained is trained based on the training set in the training data to obtain the updated epoxy material performance prediction model.

[0083] Among them, the epoxy material performance prediction model to be trained is a feedforward artificial neural network prediction model.

[0084] Specifically, the training data is randomly divided into K subsets of approximately equal size. In each iteration, one subset is selected as the test set, and the remaining K-1 subsets are combined as the training set. The training set is used to train the epoxy material performance prediction model, resulting in the updated epoxy material performance prediction model. The value of K is set based on the training requirements of the epoxy material performance prediction model.

[0085] Optionally, the epoxy material performance prediction model to be trained is trained based on the training set in the training data to obtain a trained and updated epoxy material performance prediction model, including: Determine the initial parameters of the epoxy material performance prediction model to be trained; The training set from the training data is input into the epoxy material performance prediction model to be trained, and the prediction data is output. The predicted data and the predetermined real data are calculated according to a predetermined loss function to determine the objective function value; The initial parameters of the epoxy material performance prediction model to be trained are adjusted based on the objective function value to obtain the updated epoxy material performance prediction model.

[0086] In this embodiment, before starting model training, initial parameters need to be set for the epoxy material performance prediction model to be trained. Initial parameters include the model's weights, biases, and number of neurons. The selection of initial parameters can be random or based on prior knowledge or experience. For example, Xavier initialization or He initialization methods can be used to initialize the weights to help the model converge faster.

[0087] The training set data is used as input to the epoxy material performance prediction model to be trained. The model processes the input data according to the current parameters and outputs the predicted band gap width, dielectric constant, or glass transition temperature of the epoxy material.

[0088] Furthermore, the predetermined true epoxy material bandgap width, true dielectric constant, or true glass transition temperature are actual values ​​obtained through experimental measurements or other reliable methods.

[0089] In this approach, the objective function measures the difference between the model's predicted values ​​and the actual values. Common loss functions include mean squared error (MSE), mean absolute error (MAE), cross-entropy loss, logarithmic loss, and exponential loss.

[0090] In this embodiment, the initial parameters of the model are adjusted using an optimization algorithm (such as gradient descent, Adam optimizer, etc.) based on the objective function value. The goal of the optimization algorithm is to minimize the objective function value, that is, to make the model's predicted value as close as possible to the true value. The above steps are repeated until the objective function value converges or a preset number of training epochs is reached. Each iteration updates the model's parameters, continuously improving the model's performance.

[0091] By fully considering the characteristics of key components in epoxy materials and the influence of process parameters, a reasonable epoxy material performance prediction model is constructed to achieve high-precision quantitative prediction of the electronic bandgap, dielectric constant, or glass transition temperature of epoxy materials, while effectively reducing the amount of computation.

[0092] Optionally, determining the initial parameters of the epoxy material performance prediction model to be trained includes: The number of hidden layer neurons in the epoxy material performance prediction model to be trained is set according to the total amount of training data and the number of features in the training data.

[0093] Specifically, the number of hidden layer neurons in the epoxy material performance prediction model to be trained can be optimized based on the overall scale and feature dimensions of the training data.

[0094] By adjusting the number of hidden layer neurons in the epoxy material performance prediction model being trained, the model can better capture complex relationships and features in the data, thereby improving the accuracy of epoxy material performance prediction. This avoids overfitting due to too many neurons or underfitting due to too few neurons, ensuring the model performs well on both the training and test sets. Optimizing the number of hidden layer neurons prevents the model from becoming overly complex, thus reducing the computational resources and time required during training. When the number of neurons is appropriate, the model trains faster, converges more quickly, reduces training costs, and improves efficiency.

[0095] S330. The epoxy material performance prediction model is evaluated using the test set in the training data to obtain evaluation index values.

[0096] The evaluation metrics include accuracy, precision, recall, mean squared error, mean absolute error, and coefficient of determination.

[0097] Specifically, the training data is randomly divided into K subsets of approximately equal size. In each iteration, one subset is selected as the test set, and the remaining K-1 subsets are combined as the training set. The epoxy material performance prediction model is evaluated using the training and test sets to obtain evaluation index values. The value of K is set based on the training requirements of the epoxy material performance prediction model.

[0098] S340. Based on the evaluation index values, determine the optimal parameters of the epoxy material performance prediction model.

[0099] In this approach, based on the evaluation index values, optimization algorithms are used to find the parameter values ​​that optimize the model's performance. These optimization algorithms include genetic algorithms, gradient descent, and others.

[0100] The technical solution of this invention involves acquiring training data for a performance prediction model of epoxy materials to be trained, training the model based on the training set in the training data to obtain an updated performance prediction model, evaluating the model using the test set in the training data to obtain evaluation index values, and determining the optimal parameters of the model based on these evaluation index values. By implementing this technical solution, cross-validation can be performed multiple times on limited data through dividing the dataset into multiple subsets, thereby more comprehensively evaluating the model's performance. In cross-validation, the model is trained and tested on different subsets, reducing the risk of overfitting due to improper dataset partitioning. By performing cross-validation on each parameter combination, the performance of different parameter combinations can be evaluated more accurately, thus selecting the optimal parameter settings.

[0101] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and no limitation is made herein.

[0102] Example 5 like Figure 5As shown, this application also provides an electronic device, which may be a computer device, a microcontroller device, a smart mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, processor, and transceiver component are connected via a bus; the memory can be used to store executable programs, and an exemplary executable program may include instructions; the processor is used to execute the instructions stored in the memory. The memory can also be used to store data, which can be accessed and / or modified when instructions are executed.

[0103] The processor may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and it is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the storage medium to realize the corresponding method flow or corresponding function, so as to realize the steps of the epoxy material performance prediction method in the above embodiments.

[0104] Example 6 Based on the same inventive concept, this application also provides a readable storage medium, specifically an electronic device readable storage medium (Memory). This readable storage medium is a memory device within an electronic device used to store programs and data. It is understood that the storage medium here can include both built-in storage media within the electronic device and extended storage media supported by the electronic device. The storage medium provides storage space, which stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more executable programs (including program code). It should be noted that the storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. Loading and executing one or more instructions stored in the storage medium by the processor can implement the steps of the epoxy material performance prediction method described in the above embodiments.

[0105] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0106] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0107] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0108] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0109] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for predicting the properties of epoxy materials, characterized in that, include: Obtain formulation data for epoxy materials, including epoxy resin series matrix molecular structure data, epoxy resin curing agent category data, and catalyst category data; The formulation data is predicted based on a pre-trained epoxy material performance prediction model to obtain predicted process parameter data. The predicted process parameter data includes the epoxy material bandgap width, dielectric constant, or glass transition temperature. The pre-trained epoxy material performance prediction model is obtained by training the prediction model with formulation data and process parameters as training data. The prediction model includes a feedforward neural network model.

2. The method according to claim 1, characterized in that, The training of the epoxy material performance prediction model includes: Based on the total amount of training data in the training set and the number of features in the training data, the initial parameters of the epoxy material performance prediction model to be trained are set, including the number of hidden layer neurons. Obtain the formulation data of the epoxy material and the corresponding process parameters; The formulation data in the training set is input into the epoxy material performance prediction model to be trained, and the predicted process parameter data is output. According to a predetermined loss function, the predicted process parameter data and the process parameter data corresponding to the formula data in the training set are calculated to determine the loss function value; The initial parameters of the epoxy material performance prediction model to be trained are adjusted based on the loss function value to obtain the updated epoxy material performance prediction model. The trained and updated epoxy material performance prediction model is evaluated using a test set to obtain evaluation indices; and the optimal parameters of the epoxy material performance prediction model are determined based on the evaluation indices to obtain the trained epoxy material performance prediction model.

3. The method as described in claim 2, characterized in that, After obtaining the epoxy material formulation data and the corresponding process parameters, the process further includes: The formula data and the corresponding process parameters are preprocessed based on at least one of invalid sample removal, feature dimensionality reduction, outlier handling, and data standardization to obtain preprocessed training data.

4. The method as described in claim 2, characterized in that, The predetermined loss function includes: mean squared error, mean absolute error, cross-entropy loss, logarithmic loss, and / or exponential loss.

5. The method as described in claim 2, characterized in that, The test set is used to evaluate the trained and updated epoxy material performance prediction model to obtain evaluation indicators; Based on the evaluation indicators, the optimal parameters of the epoxy material performance prediction model are determined to obtain the trained epoxy material performance prediction model, including: The formulation data from the test set are substituted into the updated epoxy material performance prediction model to obtain the predicted process parameters. Based on the process parameters corresponding to the formula data in the test set and the predicted process parameters, evaluation indicators are calculated, including: accuracy, precision, recall, mean squared error, mean absolute error, and coefficient of determination. Based on the evaluation index values, an optimization algorithm is used to find the parameter values ​​that make the epoxy material performance prediction model perform optimally. The epoxy material performance prediction model corresponding to the optimal parameters is used as the trained epoxy material performance prediction model.

6. The method according to claim 1, characterized in that, The process involves using a pre-trained epoxy material performance prediction model to predict the formulation data and obtain the predicted process parameter data. This process further includes: In the process of predicting the formulation data based on the trained epoxy material performance prediction model, the contribution value of the formulation data is determined. When the contribution value of the formula data is greater than or equal to a preset threshold, the formula data is taken as the optimal data.

7. The method according to claim 6, characterized in that, In the process of predicting the formulation data based on the trained epoxy material performance prediction model, the contribution value of the formulation data is determined, including: Based on each piece of data in the formula data, a first set and a second set are determined; wherein, the first set is a subset of the remaining data in the formula data excluding the current data; and the second set is composed of the first set and the current data. The value of the first set and the value of the second set are calculated based on a preset feature function; the value of the second set is subtracted from the value of the first set to obtain the change. The weight of the first set is determined by multiplying the factorial of the number of elements in the first set by the factorial of the difference between the total number of data and the number of elements in the first set minus one, and then dividing by the factorial of the total number of data. The weights of the first set and the change amount are weighted and combined to obtain the contribution value of each data point in the target data.

8. The method according to claim 7, characterized in that, The contribution value is calculated using the following formula: ， in, Representative in addition to the first Summation of the subset of data other than the given data; As contribution value, The number of elements in the first set. The representation does not include the first The first set of data, This indicates that the second set consists of all the data in the target data. The value of the first set, For the value of the second set, The amount of data for the formula. } represents all data in the recipe data other than the current data.

9. The method according to claim 1, characterized in that, After obtaining the formulation data of the epoxy material, the method further includes: The formulation data is preprocessed by at least one of invalid sample removal, feature dimensionality reduction, outlier handling, and data standardization to obtain preprocessed formulation data.

10. A system for predicting the performance of epoxy materials, characterized in that, include: The data acquisition module is used to acquire the formulation data of epoxy materials, including the molecular structure data of epoxy resin series matrix, the category data of epoxy resin curing agent, and the category data of catalyst. The prediction module is used to predict the formulation data based on a pre-trained epoxy material performance prediction model to obtain the predicted process parameter data. The predicted process parameter data includes the epoxy material bandgap width, dielectric constant, or glass transition temperature. The pre-trained epoxy material performance prediction model is obtained by training the prediction model with formulation data and process parameters as training data. The prediction model includes a feedforward neural network model.

11. The system according to claim 10, characterized in that, It also includes a training module for: Based on the total amount of training data in the training set and the number of features in the training data, the initial parameters of the epoxy material performance prediction model to be trained are set, including the number of hidden layer neurons. Obtain the formulation data of the epoxy material and the corresponding process parameters; The formulation data in the training set is input into the epoxy material performance prediction model to be trained, and the predicted process parameter data is output. According to a predetermined loss function, the predicted process parameter data and the process parameter data corresponding to the formula data in the training set are calculated to determine the loss function value; The initial parameters of the epoxy material performance prediction model to be trained are adjusted based on the loss function value to obtain the updated epoxy material performance prediction model. The trained and updated epoxy material performance prediction model is evaluated using a test set to obtain evaluation indices; and the optimal parameters of the epoxy material performance prediction model are determined based on the evaluation indices to obtain the trained epoxy material performance prediction model.

12. The system as claimed in claim 11, characterized in that, The training module uses a test set to evaluate the updated epoxy material performance prediction model and obtain evaluation indicators. The steps for determining the optimal parameters of the epoxy material performance prediction model based on the evaluation indicators, and obtaining the trained epoxy material performance prediction model, include: The formulation data from the test set are substituted into the updated epoxy material performance prediction model to obtain the predicted process parameters. Based on the process parameters corresponding to the formula data in the test set and the predicted process parameters, evaluation indicators are calculated, including: accuracy, precision, recall, mean squared error, mean absolute error, and coefficient of determination. Based on the evaluation index values, an optimization algorithm is used to find the parameter values ​​that make the epoxy material performance prediction model perform optimally. The epoxy material performance prediction model corresponding to the optimal parameters is used as the trained epoxy material performance prediction model.

13. The system according to claim 10, characterized in that, The system also includes a data selection module; The data selection module includes: The contribution value calculation submodule is used to determine the contribution value of the formulation data during the process of predicting the formulation data based on the trained epoxy material performance prediction model. The optimal judgment submodule is used to determine the optimal data when the contribution value of the formula data is greater than or equal to a preset threshold.

14. The system according to claim 13, characterized in that, The contribution value calculation submodule is specifically used for: Based on each piece of data in the formula data, a first set and a second set are determined; wherein, the first set is a subset of the remaining data in the formula data excluding the current data; and the second set is composed of the first set and the current data. The value of the first set and the value of the second set are calculated based on a preset feature function; the value of the second set is subtracted from the value of the first set to obtain the change. The weight of the first set is determined by multiplying the factorial of the number of elements in the first set by the factorial of the difference between the total number of data and the number of elements in the first set minus one, and then dividing by the factorial of the total number of data. The weights of the first set and the change amount are weighted and combined to obtain the contribution value of each data point in the target data.

15. An electronic device, characterized in that, include: At least one processor and memory; The memory and processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, an epoxy material performance prediction method as described in any one of claims 1 to 9 is implemented.

16. A readable storage medium, characterized in that, It contains an executable program, which, when executed, implements a method for predicting the performance of epoxy materials as described in any one of claims 1 to 9.

Images