Bismaleimide resin performance prediction model training method, performance prediction method, structure design method and related product

Abstract

The invention discloses a performance prediction model training method, a performance prediction method, a structure design method and related products of bismaleimide resin, and relates to the technical field of bismaleimide resin.The performance prediction model training method comprises the steps that a molecular structure is used as input, test conditions are used as supplementary input characteristics, and a performance prediction model is obtained; taking the dielectric constant as a label and the predicted value of the dielectric constant as output, training a supervised learning model to obtain a dielectric constant prediction model; training a supervised learning model by taking the molecular structure as input, the melting point data as a label and the predicted value of the melting point data as output to obtain a melting point prediction model; the molecular structure serves as input, the 5% thermal decomposition temperature serves as a label, the predicted value of the 5% thermal decomposition temperature serves as output, a supervised learning model is trained, and a thermal decomposition temperature prediction model is obtained. According to the method, the three prediction models are obtained by training the supervised learning model, so that the prediction of various properties of the bismaleimide resin material can be carried out based on the theoretical model.

Description

Technical Field

[0001] The present application relates to the technical field of bismaleimide resins, and in particular to a performance prediction model training method, a performance prediction method, a structure design method and related products of bismaleimide resins. Background Art

[0002] In recent years, with the continued development of 5G communications technology, automotive electronics, and aerospace engineering, electronic devices in these fields have shown a significant trend towards miniaturization, lightweighting, and higher frequency and speed. This technological shift places higher demands on the comprehensive performance of electronic packaging materials, with low-k dielectric materials playing a particularly crucial role in high-frequency circuit systems. The development of resin-based composite materials that combine low k dielectric constant with high thermal stability has become a key research direction in driving technological advancements in the electronic information industry. Furthermore, the feasibility of material processing technology is a key factor in determining its engineering application.

[0003] Bismaleimide resins, a class of addition-polymerized thermosetting polyimide materials with maleimide as active end groups, exhibit no low-molecular-weight byproducts during the curing process, and exhibit short molecular chains with a high degree of crosslinking between crosslinks. These materials, with their excellent thermal stability, low hygroscopicity, and good dielectric properties, combined with their facile monomer synthesis, abundant raw material resources, and commercially viable cost advantages, present broad application prospects in large-scale electronic device manufacturing. However, the high symmetry of the molecular chains and the regularity of the crystal structure of most bismaleimide resins result in high melting points for the prepolymers, significantly impacting the material's processing and molding properties. Conventional strategies to improve processing properties by introducing asymmetric structures often result in increased polarity, decreased dielectric properties, and negatively impacting thermal stability. This interplay of properties has become a technical bottleneck restricting the development of high-performance bismaleimide resins. Therefore, research on molecular structure optimization is urgently needed to effectively enhance processing properties while maintaining high thermal stability and excellent dielectric properties.

[0004] To meet the comprehensive performance requirements of materials for engineering applications, researchers need to scientifically balance various performance indicators. Traditional resin material development relies primarily on trial and error. This empirical R&D process not only results in long material development cycles and high economic costs, but also, due to the complexity of the material's structure-property relationship, makes it difficult to simultaneously reduce the dielectric constant and improve heat resistance and processing performance. Summary of the Invention

[0005] The purpose of this application is to provide a performance prediction model training method, performance prediction method, structural design method and related products of bismaleimide resin, which can predict the performance of bismaleimide resin materials based on theoretical models and can simultaneously consider multiple properties such as dielectric constant and heat resistance.

[0006] To achieve the above objectives, this application provides the following solutions:

[0007] In a first aspect, the present application provides a method for training a performance prediction model for a bismaleimide resin, comprising:

[0008] Obtain bismaleimide data and supplementary data; the bismaleimide data includes: the molecular structure of bismaleimide resin and corresponding experimental data; the experimental data includes: dielectric constant data, melting point, and 5% thermal decomposition temperature; the supplementary data includes: dielectric constant supplementary data and temperature supplementary data; the dielectric constant supplementary data includes: the molecular structure of several types of resins and corresponding dielectric constant data; the temperature supplementary data includes: the molecular structure of polyimide resin and corresponding melting point data and 5% thermal decomposition temperature data; the dielectric constant data includes: dielectric constant and corresponding test conditions;

[0009] The molecular structure is used as input, the test condition is used as a supplementary input feature, the dielectric constant is used as a label, and the predicted value of the dielectric constant is used as an output, and a supervised learning model is trained to obtain a dielectric constant prediction model;

[0010] The molecular structure is used as input, the melting point data is used as a label, and the predicted value of the melting point data is output, and a supervised learning model is trained to obtain a melting point prediction model;

[0011] The molecular structure is used as input, the 5% thermal decomposition temperature is used as a label, and the predicted value of the 5% thermal decomposition temperature is used as an output, and a supervised learning model is trained to obtain a thermal decomposition temperature prediction model;

[0012] The performance prediction model includes: the dielectric constant prediction model, the melting point prediction model and the thermal decomposition temperature prediction model.

[0013] In a second aspect, the present application provides a method for predicting the performance of a bismaleimide resin, comprising:

[0014] Obtaining the molecular structure of the bismaleimide resin to be predicted;

[0015] Inputting the molecular structure into a dielectric constant prediction model to obtain a predicted value of the dielectric constant of the bismaleimide resin to be predicted; the dielectric constant prediction model is trained by the performance prediction model training method of the bismaleimide resin described above;

[0016] Inputting the molecular structure into a melting point prediction model to obtain a predicted value of the melting point data of the bismaleimide resin to be predicted; the melting point prediction model is trained by the performance prediction model training method of the bismaleimide resin described above;

[0017] Inputting the molecular structure into a thermal decomposition temperature prediction model to obtain a predicted value of the 5% thermal decomposition temperature of the bismaleimide resin to be predicted; the thermal decomposition temperature prediction model is trained by the performance prediction model training method of the bismaleimide resin described above;

[0018] The predicted value of the dielectric constant, the predicted value of the melting point data and the predicted value of the 5% thermal decomposition temperature of the bismaleimide resin to be predicted are performance prediction data.

[0019] In a third aspect, the present application provides a method for structural design of a bismaleimide resin, comprising:

[0020] Construct several virtual molecular structures of bismaleimide resins;

[0021] Calculating a synthetic accessibility score for each of the virtual molecular structures;

[0022] Inputting the virtual molecular structure into a dielectric constant prediction model to obtain corresponding dielectric constant prediction values; the dielectric constant prediction model is trained by the performance prediction model training method of bismaleimide resin described above;

[0023] Inputting the virtual molecular structure into a melting point prediction model to obtain corresponding melting point data prediction values; the melting point prediction model is trained by the performance prediction model training method of the bismaleimide resin described above;

[0024] Inputting the virtual molecular structure into a thermal decomposition temperature prediction model to obtain corresponding 5% thermal decomposition temperature prediction values; the thermal decomposition temperature prediction model is trained by the performance prediction model training method of the bismaleimide resin described above;

[0025] Calculating a comprehensive score for each virtual molecular structure according to the predicted dielectric constant value, the predicted melting point data value, the predicted 5% thermal decomposition temperature value, and the synthetic accessibility score;

[0026] The design structure of the bismaleimide resin is determined according to the comprehensive score.

[0027] In a fourth aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for training a performance prediction model of bismaleimide resin or the above-described method for predicting the performance of bismaleimide resin or the above-described method for structural design of bismaleimide resin.

[0028] In a fifth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the performance prediction model training method for bismaleimide resin described above, or the performance prediction method for bismaleimide resin described above, or the structural design method for bismaleimide resin described above.

[0029] In a sixth aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements the performance prediction model training method for bismaleimide resin described above, or the performance prediction method for bismaleimide resin described above, or the structural design method for bismaleimide resin described above.

[0030] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0031] The present application provides a performance prediction model training method, a performance prediction method, a structural design method and related products of a bismaleimide resin. The performance prediction model training method comprises: obtaining bismaleimide data and supplementary data; the bismaleimide data comprises: the molecular structure of the bismaleimide resin and the corresponding experimental data; the experimental data comprises: dielectric constant data, melting point and 5% thermal decomposition temperature; the supplementary data comprises: dielectric constant supplementary data and temperature supplementary data; the dielectric constant supplementary data comprises: the molecular structure of several types of resins and the corresponding dielectric constant data; the temperature supplementary data comprises: the molecular structure of the polyimide resin and the corresponding melting point data and 5% thermal decomposition temperature data; the dielectric constant ... The data includes: dielectric constant and corresponding test conditions; taking the molecular structure as input, the test conditions as supplementary input features, the dielectric constant as a label, and the predicted value of the dielectric constant as output, training a supervised learning model to obtain a dielectric constant prediction model; taking the molecular structure as input, the melting point data as a label, and the predicted value of the melting point data as output, training a supervised learning model to obtain a melting point prediction model; taking the molecular structure as input, the 5% thermal decomposition temperature as a label, and the predicted value of the 5% thermal decomposition temperature as output, training a supervised learning model to obtain a thermal decomposition temperature prediction model; the performance prediction model includes: the dielectric constant prediction model, the melting point prediction model and the thermal decomposition temperature prediction model. This application obtains three prediction models by training a supervised learning model. These three prediction models can simultaneously consider multiple properties such as dielectric constant and heat resistance, so that various properties of bismaleimide resin materials can be predicted based on the theoretical model. At the same time, this application improves the accuracy of the machine learning model in predicting the dielectric constant, melting point and 5% thermal decomposition temperature of bismaleimide resin by introducing experimental data of various types of resins as a supplement, thereby solving the problem of small material data. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0033] Figure 1 A schematic flow chart of a method for training a performance prediction model for a bismaleimide resin according to an embodiment of the present application;

[0034] Figure 2 A schematic diagram of the data distribution of the dielectric constant, melting point, and 5% thermal decomposition temperature data set provided in one embodiment of the present application;

[0035] Figure 3 A schematic diagram of a process for constructing a graph convolutional neural network combined with a multi-layer perceptron performance prediction model provided in one embodiment of the present application;

[0036] Figure 4 A schematic diagram of the accuracy of the bismaleimide resin performance prediction model provided in one embodiment of the present application;

[0037] Figure 5 A schematic diagram of a virtual bismaleimide resin structure combination method provided in one embodiment of the present application;

[0038] Figure 6 A schematic diagram of the results of verifying the reliability of the model provided in one embodiment of the present application;

[0039] Figure 7 A schematic diagram comparing the performance of the models of Example 4, Comparative Examples 1 and 2 provided in one embodiment of the present application;

[0040] Figure 8 A schematic diagram of the performance space of the predicted candidate bismaleimide resin structure provided in one embodiment of the present application;

[0041] Figure 9 A schematic diagram of the structure of a preferred bismaleimide resin obtained by screening according to an embodiment of the present application;

[0042] Figure 10 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0044] Constructing a material property prediction method based on theoretical models to achieve rapid screening and design of bismaleimide resin molecular structures with low dielectric constant, high heat resistance and easy processability has become a key issue that needs to be urgently addressed in this field.

[0045] The prediction method of this application includes the following steps: inputting the molecular graph and adjacency matrix of the bismaleimide resin material into a prediction model to obtain performance prediction values of the bismaleimide resin; the prediction model is trained using a dataset; wherein the dataset includes SMILES and performance parameters of the bismaleimide resin. This application utilizes data-driven and material genome technology to guide the rational design of resin structure, which can significantly improve the efficiency and accuracy of material research and development, and efficiently develop bismaleimide resins with low economic cost and strong practicality, opening up new paths for the development and application of low-dielectric, easy-to-process, and high-temperature resistant bismaleimide resins.

[0046] This application addresses the technical bottlenecks of low R&D efficiency and high economic costs in the traditional trial-and-error method for designing bismaleimide resins with conflicting properties. By constructing a high-precision prediction model, the rationality and efficiency of bismaleimide resin molecular structure design can be significantly improved, while reducing R&D costs.

[0047] However, experimental data for bismaleimide resin systems is scarce, and most of it comes from multi-component systems or actual product-level testing. This makes it difficult for existing experimental data to accurately reflect the material's intrinsic structure-property relationship. Directly using this data to build machine learning models significantly reduces the model's predictive accuracy and generalization capabilities.

[0048] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0049] Example 1:

[0050] In an exemplary embodiment, Figure 1 As shown, a method for training a performance prediction model for a bismaleimide resin is provided. The method is executed by a computer device, specifically a computer device such as a terminal or a server, or a terminal and a server. In the embodiment of the present application, the method is applied to a server as an example for illustration, and includes the following steps S1 to S4. Wherein:

[0051] S1. Obtain bismaleimide data and supplementary data. The bismaleimide data includes the molecular structure of bismaleimide resin and corresponding experimental data. The experimental data includes dielectric constant data, melting point, and 5% thermal decomposition temperature. The supplementary data includes supplementary dielectric constant data and supplementary temperature data. The supplementary dielectric constant data includes the molecular structure and corresponding dielectric constant data of several types of resins. The supplementary temperature data includes the molecular structure of polyimide resin and its corresponding melting point data and 5% thermal decomposition temperature data. The dielectric constant data includes the dielectric constant and corresponding test conditions. In this embodiment, the test conditions refer to the test frequency and test temperature.

[0052] The experimental data for bismaleimide resin materials is relatively small, mostly based on multi-component measurements, lacking more accurate experimental-level structural and performance data. Direct modeling based on this data results in low model accuracy and weak generalization capabilities. To improve model accuracy, this example incorporates data from multiple resin structures as supplementary data during data collection, based on the structural characteristics of bismaleimide resin. The dielectric constant of a polymer is mainly determined by its structure and is usually temperature- and frequency-dependent. For the dielectric constant model, this embodiment also cites dielectric constant data of various types of resins, including thermosetting and thermoplastic resins, as supplements, and collects test frequency and test temperature data (dielectric properties are associated with polymer structure and test conditions. Using dielectric constant data of different types of polymers (including thermoplastics and thermosetting resins) for data enhancement helps the model to reasonably learn the relationship between structure and dielectric constant, thereby improving the reliability and generalization ability of the model); for the melting point and 5% thermal decomposition temperature, this embodiment also cites the melting point and 5% thermal decomposition temperature data of polyimide resins that also have imide as supplements (for the melting point and 5% thermal decomposition temperature, polyimide and bismaleimide have the same imide group. This structural similarity gives them similar properties and common rules in thermal properties. Therefore, introducing polyimide data for data enhancement helps improve the generalization ability of the model).

[0053] Then the above data set is cleaned to improve the data quality.

[0054] The method of cleaning the data set is preferably to cover the test frequency of the dielectric constant from 1*10 -5 Hz to 1.4*10 8Hz, test temperature data from -196 ° C to 380 ° C, the data distribution range is wide and conforms to the normal distribution, for the dielectric constant test frequency with too large distribution skewness, the data after taking lg (taking the logarithm with base e) replaces the original data, ensuring that the model better learns the relationship between the dielectric constant and the material structure and test conditions; for the same structure with different melting point data, when the maximum and minimum values differ by more than 50 ° C, the structure is discarded, and the remaining structures use the average melting point value as the target performance; the test atmosphere for the 5% thermal decomposition temperature of the resin material is set to a nitrogen atmosphere; for the same structure with different 5% thermal decomposition temperatures, when the maximum and minimum values differ by more than 50 ° C, the structure is discarded, and the remaining structures use the average 5% thermal decomposition temperature value as the target performance.

[0055] The data volume of the data set is generally more than 100. The experimental data of the dielectric constant, melting point and 5% thermal decomposition temperature of the bismaleimide resin are preferably 30-60, such as 60 experimental data of the dielectric constant of the bismaleimide resin, 51 experimental data of the melting points of the bismaleimide resin, and 30 experimental data of 5% thermal decomposition temperatures of the bismaleimide resin. The data set also includes 100-2500 supplementary performance data except the experimental data of the bismaleimide resin, such as 2344 experimental data of the dielectric constant of various types of resins (polyolefins, polyurethanes, etc.), 475 experimental data of 5% thermal decomposition temperatures of polyimides, and 141 melting point experimental data of polyimides. The introduction of the supplementary data can expand the chemical space of model learning and enhance the generalization ability of the model. Those skilled in the art know that the performance data in the data set can generally be obtained by collecting in public literature and / or obtaining in public databases.

[0056] S2. Using the molecular structure as input, the test conditions as supplementary input features, the dielectric constant as a label, and the predicted value of the dielectric constant as output, a supervised learning model is trained to obtain a dielectric constant prediction model.

[0057] S3. Take the molecular structure as input, the melting point data as a label, and the predicted value of the melting point data as output, train a supervised learning model, and obtain a melting point prediction model.

[0058] S4. Using the molecular structure as input, the 5% thermal decomposition temperature as a label, and the predicted value of the 5% thermal decomposition temperature as an output, a supervised learning model is trained to obtain a thermal decomposition temperature prediction model. The performance prediction model includes: the dielectric constant prediction model, the melting point prediction model, and the thermal decomposition temperature prediction model.

[0059] It is worth noting that there is no specific limitation on the order of steps S2 to S4.

[0060] The supervised learning model selected in this embodiment includes: an input layer, an intermediate layer and a fully connected layer; the input layer is used to encode the molecular structure in the form of a molecular graph; the intermediate layer is used to update the embedding vector of the target node by aggregating neighbor node information through continuous convolution and batch normalization operations; the fully connected layer is used to use the experimental performance data corresponding to the molecular structure as the model output.

[0061] Among them, the intermediate layer can include graph convolution layer, batch normalization layer and pooling layer. The graph convolution layer realizes the aggregation of feature relationships of adjacent nodes and updates the node feature representation; the batch normalization layer normalizes the features to accelerate the convergence of the model; the pooling layer sums all node features to obtain the feature representation of the entire graph.

[0062] The model used in this example is a learning framework based on a graph convolutional neural network combined with a multilayer perceptron. The multilayer perceptron (MLP) layer is composed of a stack of fully connected layers and activation functions. The addition of two MLP layers allows for nonlinear transformation of features, capturing more complex information and improving the model's expressiveness. Finally, a fully connected layer is used for output.

[0063] Specifically, this application combines the cleaned data set to construct a machine learning model of a graph convolutional neural network combined with a multi-layer perceptron to automatically learn the relationship between the structure and performance of bismaleimide resin, and obtain a performance prediction model for bismaleimide resin. Compared with the machine learning model built based on Gaussian process regression and the XGBoost regression model based on the gradient boosting tree, its prediction accuracy is higher.

[0064] In this embodiment, a learning framework based on a graph convolutional neural network combined with a multi-layer perceptron is used. The bismaleimide resin structure is encoded in the form of a molecular graph as the input of the model. The embedding vector of the target node is updated by aggregating neighbor node information through continuous convolution and batch normalization operations. Finally, two fully connected layers are connected, and the experimental performance data corresponding to the structure is used as the model output. The data set is divided into a training:test ratio of 8:2 to automatically learn the potential relationship between the structure and performance of bismaleimide resin.

[0065] Since the dielectric constant of a polymer is generally frequency-dependent and temperature-dependent, this embodiment adds the test frequency and temperature of the dielectric constant as supplementary features to the model input to help the model capture the law of change of the dielectric constant with environmental conditions. This improves the model's prediction accuracy while reflecting the direct relationship between the dielectric constant of the bismaleimide resin polymer and its macroscopic behavior.

[0066] The method of encoding the bismaleimide resin structure in the form of a molecular graph preferably includes: representation of atomic information in the molecule, representation of bond information, and representation of a molecular adjacency matrix.

[0067] The updating of the target node's embedding vector by aggregating neighbor node information through continuous convolution and batch normalization operations preferably includes: using batch normalization immediately after each layer of graph convolution to accelerate the convergence speed of the model and improve the stability of the model.

[0068] The two fully connected layers are preferably connected after the GCN, with two fully connected layers having 300 nodes. This combined use of GCN and MLP can capture local and global features in graph data, thereby improving the expressive power of the model.

[0069] In this embodiment, the dielectric constant, melting point, and 5% thermal decomposition temperature data of the bismaleimide resin are expressed in a material property space.

[0070] Those skilled in the art should know that the 5% thermal decomposition temperature reflects the high temperature resistance of the resin material; and the melting point reflects the processing performance of the resin material.

[0071] This example introduces experimental data of various types of resins as supplementary data to improve the accuracy of the machine learning model in predicting the dielectric constant, melting point, and 5% thermal decomposition temperature of bismaleimide resin, solving the problem of small material data. In terms of dielectric constant prediction, the R 2 The average absolute error is 0.19; in terms of melting point prediction, the R 2 The R value of the test set is 0.88, with an average absolute error of 28°C. 2 The average absolute error is 0.82, and the average absolute error is 22°C.

[0072] Example 2:

[0073] Based on the same inventive concept, the present embodiment provides a method for predicting the performance of a bismaleimide resin, comprising:

[0074] A1. Obtain the molecular structure of the bismaleimide resin to be predicted.

[0075] A2. Inputting the molecular structure into a dielectric constant prediction model to obtain a predicted value of the dielectric constant of the bismaleimide resin to be predicted; the dielectric constant prediction model is trained by the performance prediction model training method of the bismaleimide resin described in Example 1.

[0076] A3. Inputting the molecular structure into a melting point prediction model to obtain a predicted value of the melting point data of the bismaleimide resin to be predicted; the melting point prediction model is trained by the performance prediction model training method of the bismaleimide resin described in Example 1.

[0077] A4. Inputting the molecular structure into a thermal decomposition temperature prediction model to obtain a predicted value of the 5% thermal decomposition temperature of the bismaleimide resin to be predicted; the thermal decomposition temperature prediction model is trained by the performance prediction model training method of the bismaleimide resin described in Example 1.

[0078] The predicted value of the dielectric constant, the predicted value of the melting point data and the predicted value of the 5% thermal decomposition temperature of the bismaleimide resin to be predicted are performance prediction data.

[0079] Example 3:

[0080] Based on the same inventive concept, the present embodiment provides a method for structural design of a bismaleimide resin, comprising:

[0081] B1. Construct several virtual molecular structures of bismaleimide resins.

[0082] Specifically:

[0083] B11. Determine the genetic type of the bismaleimide resin structure.

[0084] In this embodiment, the gene types of the bismaleimide resin structure are defined; wherein the gene types preferably include: 1) maleic anhydride gene; 2) diamine gene with two and only two amino groups at both ends of the main chain; 3) ammonium acid gene with a benzene ring in the main chain, and the benzene ring is connected to an amino group and a formic acid; 4) diphenol gene with a benzene ring in the main chain, and two hydroxyl groups are connected to the benzene ring; 5) maleic acid diamine gene; 6) halogenated hydrocarbon gene with halogen atoms at both ends of the main chain; 7) amino alcohol gene with an amino group and a hydroxyl group connected to each end of the main chain; 8) diacid gene with two and only two formic acid groups at both ends of the main chain.

[0085] B12. Obtain several corresponding genes from the database according to the gene type.

[0086] Genes were searched using public databases to obtain a diverse gene pool.

[0087] Among them, the gene pool may include 700-800 diamine genes, 1-100 ammonium acid genes, 100-200 diphenol genes, 1-50 halogenated hydrocarbon genes, 1-50 amino alcohol genes, and 100-200 diacid genes; for example, 788 diamine genes, 50 ammonium acid genes, 101 diphenol genes, 33 halogenated hydrocarbon genes, 48 amino alcohol genes, and 111 diacid genes.

[0088] B13. Synthesizing several of the genes into several bismaleimide resins according to a preset template to obtain several virtual molecular structures of the bismaleimide resins.

[0089] In this embodiment, based on the gene pool, a virtual molecular structure of bismaleimide resin is obtained by combination; the combination method of the virtual molecular structure is to use the SMARTS (SMiles ARbitrary Target Specification) module in RDKit to construct a chemical reaction template, which is implemented by Python programming; SMARTS is a language in RDKit for describing molecular structures, providing specific symbols to describe the atoms, bonds and the connection relationship between them in molecules. In this embodiment, by using the molecular connection method of the SMARTS module in Python, four reaction templates of rules (1)-(4) are constructed, and the reaction method (connection method) such as (1)-(4) is realized. Specifically, it is: defining the SMARTS pattern of reactants and products; combining reactants and products and creating a SMARTS format template for the reaction; using the reaction template to execute the chemical reaction of the molecule.

[0090] In addition to the common synthesis pathway of bismaleimide resins by reacting maleic anhydride and diamine, this example summarizes other possible experimental synthesis pathways and defines three new reaction templates. This not only enriches the synthesis methods of bismaleimide resins but also provides more possibilities for the design of subsequent high-performance resin materials. The chemical templates include the following combination rules:

[0091] Rule (1):

[0092]

[0093] Among them, maleimide and diamine are used as reactants and react in a 2:1 ratio to generate bismaleimide (common synthesis route).

[0094] Rule (2):

[0095]

[0096] Among them, maleimide, ammonium phosphate and diphenol phosphate are used as reactants and react in a ratio of 2:2:1 to generate bismaleimide resin.

[0097] Rule (3):

[0098]

[0099] Among them, maleic acid diamine and halogenated hydrocarbon react in a ratio of 2:1 to generate bismaleimide resin.

[0100] Rule (4):

[0101]

[0102] Among them, maleimide, amino alcohol gene and diacid gene react in a ratio of 2:2:1 to generate bismaleimide resin.

[0103] The virtual molecular structures of the bismaleimide resin may be more than 10,000, preferably 10,000-15,000, for example 10,744.

[0104] B2. Calculate the synthetic accessibility score of each virtual molecular structure.

[0105] In this embodiment, the method for evaluating the synthesis difficulty of the virtual molecular structure can adopt the conventional synthesis accessibility score and Beilstein complexity index in the art, preferably the synthesis accessibility score (SA).

[0106] B3. Inputting the virtual molecular structure into a dielectric constant prediction model to obtain corresponding dielectric constant prediction values; the dielectric constant prediction model is trained by the performance prediction model training method of the bismaleimide resin described in Example 1.

[0107] B4. Inputting the virtual molecular structure into a melting point prediction model to obtain corresponding melting point data prediction values; the melting point prediction model is trained by the performance prediction model training method of the bismaleimide resin described in Example 1.

[0108] B5. Input the virtual molecular structure into a thermal decomposition temperature prediction model to obtain corresponding 5% thermal decomposition temperature prediction values; the thermal decomposition temperature prediction model is trained by the performance prediction model training method of the bismaleimide resin described in Example 1.

[0109] B6. Calculate a comprehensive score for each virtual molecular structure based on the predicted dielectric constant, the predicted melting point data, the predicted 5% thermal decomposition temperature, and the synthetic accessibility score.

[0110] The calculation formula for the comprehensive score is:

[0111]

[0112]

[0113] Among them, Score, ε, T M 、T d 、S A Respectively represent the comprehensive score, dielectric constant of the virtual molecular structure, melting point, 5% thermal decomposition temperature, and synthetic accessibility score; ε min , ε max represent the minimum and maximum values of the dielectric constant of the virtual molecular structure; T Mmin 、T Mmax Respectively represent the minimum and maximum values of the melting point of the virtual molecular structure; T dmin 、T dmax They represent the minimum and maximum values of the 5% thermal decomposition temperature of the virtual molecular structure; S Amin 、S Amax represent the minimum and maximum values of the synthetic accessibility scores of the virtual molecular structures, respectively.

[0114] B7. Determine the design structure of the bismaleimide resin based on the comprehensive score.

[0115] In this embodiment, the design structure of the bismaleimide resin material is determined by sorting in order of comprehensive score.

[0116] After obtaining the material performance space, the method further includes screening the performance space, wherein the screening method adopts a weighted scalar function.

[0117] Those skilled in the art will appreciate that the structural information of a bismaleimide resin is generally obtained by converting the structure of the bismaleimide resin to be analyzed into the corresponding SMILES input. SMILES, short for Simplified Molecular Input Line Entry Specification, is a specification that explicitly describes a molecular structure using ASCII strings.

[0118] In addition to the traditional diamine genes, the bismaleimide resin gene types defined in this embodiment also include ammonium acid genes, diphenol genes, maleic acid diamine genes, amino alcohol genes and diacid genes, which greatly expands the chemical space of virtual resin structures. The selected preferred structures are expected to be widely used in the aerospace field.

[0119] Example 4:

[0120] This example conducts a performance prediction method and reliability verification of bismaleimide resin:

[0121] C1. Collect the performance data sets of dielectric constant, melting point and 5% thermal decomposition temperature from the literature. The distribution of the performance is as follows Figure 2 As shown, Figure 2 (a) is the dielectric constant performance, Figure 2 (b) is the melting point performance, Figure 2 (c) is the 5% thermal decomposition temperature performance, where the dielectric constant data includes the test frequency from 1*10 -5 Hz to 1.4*10 8 Hz, test temperatures from -196 ° C to 380 ° C, including 60 bismaleimide resins and 2344 various types of resins (polyolefins, polyurethanes, etc.) experimental data; 5% thermal decomposition temperature data including 30 bismaleimide resins and 475 polyimide resins experimental data; melting point data including 51 bismaleimide resins and 141 polyimide resins experimental data.

[0122] C2. Data cleaning of the collected data sets. For the dielectric constant test frequency, the original data was replaced with the data after taking the value of lg. For the same structure with different melting point data, if the difference between the maximum and minimum values was greater than 50°C, the structure was discarded. The average melting point value of the remaining structures was used as the target performance. The test atmosphere for the 5% thermal decomposition temperature of the resin material was set to a nitrogen atmosphere. For the same structure with different 5% thermal decomposition temperatures, if the difference between the maximum and minimum values was greater than 50°C, the structure was discarded. The average 5% thermal decomposition temperature value of the remaining structures was used as the target performance.

[0123] C3. Based on the cleaned data set, a graph convolutional neural network combined with a multi-layer perceptron performance prediction model is constructed. The process is as follows: Figure 3 ,include:

[0124] (1) The dataset is divided into training set and test set in a ratio of 8:2.

[0125] (2) Represent the molecular structure in the form of a molecular graph, specifically: through the molecular adjacency matrix and the atomic number, aromaticity, number of hydrogen atoms and implicit valence of each atom in the molecule, as well as the type, spatial configuration, direction and aromaticity of the bonds in the molecule.

[0126] (3) The characteristic vector after molecular graph encoding and the test conditions of dielectric constant are used as model input, and the corresponding performance is used as model output. Batch normalization is used after each graph convolution layer in the model, and finally two fully connected layers are connected. The performance prediction model of bismaleimide resin is obtained through forward deduction and back propagation training. The accuracy of the performance prediction model in this embodiment is as follows Figure 4 As shown, Figure 4 (a) is the dielectric constant prediction accuracy, Figure 4 (b) is the melting point prediction accuracy, Figure 4(c) is the 5% thermal decomposition temperature prediction accuracy.

[0127] C4. Inputting the structural information of the bismaleimide resin to be analyzed into the performance prediction model to obtain the dielectric constant, melting point and 5% thermal decomposition temperature performance data of the bismaleimide resin to be analyzed.

[0128] The bismaleimide resin to be analyzed includes an existing bismaleimide resin and a virtual bismaleimide resin, wherein the method for obtaining the virtual bismaleimide resin comprises the following steps:

[0129] (1) Define the gene types of bismaleimide resins, including maleic anhydride gene, diamine gene, ammonium acid gene, diphenol gene, maleic acid diamine gene, halogenated hydrocarbon gene, amino alcohol gene and diacid gene.

[0130] (2) Genes were searched on the public databases SciFinder, PubChem, and ChemSpider, and a gene pool of 788 diamine genes, 50 ammonium acid genes, 101 diphenol genes, 33 halogenated hydrocarbon genes, 48 amino alcohol genes, and 111 diacid genes was obtained.

[0131] (3) By combining maleic anhydride genes with diamine genes, ammonium acid genes, diphenol genes, maleic acid diamine genes, halogenated hydrocarbon genes, amino alcohol genes and diacid genes, 10744 virtual bismaleimide resin structures were obtained. The combination methods are as follows: Figure 5 shown.

[0132] S5. Collect experimental data of existing bismaleimide resins that are not in the model and compare them with the predicted values of the model, such as Figure 6 As shown. Regarding the dielectric constant data, since radar operating frequencies are high in MHz and operating temperatures are mostly room temperature, room temperature, 1 MHz data was used for comparison. A structure not used for model training was collected. The experimental and predicted dielectric constant values were 2.94 and 2.96, respectively. The experimental and predicted melting points were 160°C and 145°C, respectively. The experimental and predicted 5% thermal decomposition temperatures were 476°C and 440°C, respectively. The predicted values for all three models were relatively close to the true values, validating the reliability of the method.

[0133] Comparative Example 1:

[0134] The same data set as step C3 in Example 4 is used for modeling. The difference from Example 4 is that the machine learning model is constructed by using the Mordred descriptor of SMILES through XGBoost regression based on the gradient boosting tree.

[0135] Comparative Example 2:

[0136] The same data set as step C3 in Example 4 is used for modeling. The difference from Example 4 is that the machine learning model is constructed by Gaussian process regression (GPR) using the Mordred descriptor of SMILES.

[0137] Application Example 4

[0138] The models of Example 4 and Comparative Example 1 and Comparative Example 2 were constructed and randomly run 100 times. The R 2 (coefficient of determination) and MAE (mean absolute error) such as Figure 7 As shown, Figure 7 (a) is the comparison of dielectric constant performance, Figure 7 (b) is a comparison of melting point performance, Figure 7 (c) is the performance comparison of 5% thermal decomposition temperature. Figure 7 It can be seen that for the dielectric constant prediction model, the average R 2 and MAE is 0.80 and 0.22, R 2 It is higher than the average value of 0.79 of the XGBoost model in comparative example 1 and higher than the average value of 0.57 of the GPR model in comparative example 2. The MAE is lower than the average value of 0.23 of the XGBoost model in comparative example 1 and lower than the average value of 0.40 of the GPR model in comparative example 2. For the melting point prediction model, the average R 2 is 0.88, the average MAE is 29.5℃, and the average R 2 is 0.61, the average MAE is 48.8°C, and the average R 2 is 0.59, and the average MAE is 52.4℃; the average R 2 Higher, lower average MAE; for the 5% thermal decomposition temperature prediction model, the average R 2 is 0.84, the average MAE is 22.3℃, and the average R 2 Higher than the average R of the XGBoost model test set in comparison example 1 2 0.60, which is higher than the average R of the GPR model test set in comparative example 2. 2 The average MAE is 0.66, which is 28°C lower than the average MAE of the XGBoost model test set in Comparative Example 1 and 24.5°C lower than the average MAE of the GPR model test set in Comparative Example 2. Compared with XGBoost and GPR, the GCN model has higher accuracy, smaller error, and stronger generalization ability, demonstrating the advantages of the method proposed in this example.

[0139] Embodiment 5:

[0140] Structure design of high-performance bismaleimide resin based on material genetics approach

[0141] In this example, based on the framework of the material gene method, the performance prediction model constructed in Example 4 was applied to quickly screen and obtain a new bismaleimide resin structure with low dielectric constant, low melting point and high 5% thermal decomposition temperature. The steps include:

[0142] D1. Use the performance prediction model constructed in Example 4 to predict the 10,744 candidate structures of virtual bismaleimide resins in Example 4 to obtain their performance data.

[0143] D2. A weighted scalar function is used to represent the three properties as a comprehensive score, where the weight ratio of dielectric constant, melting point, 5% thermal decomposition temperature, and SASocre is 3:3:3:1, such as Figure 8 As shown, the spheres from small to large represent comprehensive performance from low to high.

[0144]

[0145] Among them, Score, ε, T M 、T d , SA represent the comprehensive score, dielectric constant of virtual structure, melting point, 5% thermal decomposition temperature, and synthetic accessibility score, respectively; ε min , ε max represent the minimum and maximum values of the dielectric constant in the virtual structure respectively; T Mmin 、T Mmax represent the minimum and maximum values of the melting point in the virtual structure; T dmin 、T dmax Represent the minimum and maximum values of 5% thermal decomposition temperature in the virtual structure; SA min 、SA max represent the minimum and maximum values of the composite accessibility scores in the virtual structure, respectively.

[0146] D3. Based on the comprehensive score, select 10 new bismaleimide resin structures from the top 5% preferred structures. The structures and the corresponding dielectric constants at room temperature and 1 MHz, melting points, and 5% thermal decomposition temperature performance predictions for each structure are as follows: Figure 9 As shown, its excellent dielectric properties, high temperature resistance and easy processing characteristics have broad application prospects in the fields of electronic communications and aerospace.

[0147] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 10As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a performance prediction model training method for a bismaleimide resin, a performance prediction method for a bismaleimide resin, or a structural design method for a bismaleimide resin is implemented.

[0148] Those skilled in the art will understand that Figure 10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0149] In an exemplary embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0150] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0151] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0152] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0153] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A performance prediction model training method for bismaleimide resin, characterized in that: include: Obtain bismaleimide data and supplementary data; the bismaleimide data includes: the molecular structure of bismaleimide resin and corresponding experimental data; the experimental data includes: dielectric constant data, melting point, and 5% thermal decomposition temperature; the supplementary data includes: dielectric constant supplementary data and temperature supplementary data; the dielectric constant supplementary data includes: the molecular structure of several types of resins and corresponding dielectric constant data; the temperature supplementary data includes: the molecular structure of polyimide resin and corresponding melting point data and 5% thermal decomposition temperature data; the dielectric constant data includes: dielectric constant and corresponding test conditions; The molecular structure is used as input, the test condition is used as a supplementary input feature, the dielectric constant is used as a label, and the predicted value of the dielectric constant is used as an output, and a supervised learning model is trained to obtain a dielectric constant prediction model; The molecular structure is used as input, the melting point data is used as a label, and the predicted value of the melting point data is output, and a supervised learning model is trained to obtain a melting point prediction model; The molecular structure is used as input, the 5% thermal decomposition temperature is used as a label, and the predicted value of the 5% thermal decomposition temperature is used as an output, and a supervised learning model is trained to obtain a thermal decomposition temperature prediction model; The performance prediction model includes: the dielectric constant prediction model, the melting point prediction model and the thermal decomposition temperature prediction model.

2. The performance prediction model training method of bismaleimide resin according to claim 1, wherein The supervised learning model includes: an input layer, an intermediate layer and a fully connected layer; The input layer is used to encode the molecular structure in the form of a molecular graph; The intermediate layer is used to update the embedding vector of the target node by aggregating neighbor node information through continuous convolution and batch normalization operations; The fully connected layer is used to use the experimental performance data corresponding to the molecular structure as the model output.

3. A method for predicting the performance of a bismaleimide resin, characterized in that: include: Obtaining the molecular structure of the bismaleimide resin to be predicted; Inputting the molecular structure into a dielectric constant prediction model to obtain a predicted value of the dielectric constant of the bismaleimide resin to be predicted; the dielectric constant prediction model is trained by the performance prediction model training method of the bismaleimide resin according to claim 1 or 2; Inputting the molecular structure into a melting point prediction model to obtain a predicted value of the melting point data of the bismaleimide resin to be predicted; the melting point prediction model is trained by the performance prediction model training method of the bismaleimide resin according to claim 1 or 2; Inputting the molecular structure into a thermal decomposition temperature prediction model to obtain a predicted value of the 5% thermal decomposition temperature of the bismaleimide resin to be predicted; the thermal decomposition temperature prediction model is trained by the performance prediction model training method of the bismaleimide resin according to claim 1 or 2; The predicted value of the dielectric constant, the predicted value of the melting point data and the predicted value of the 5% thermal decomposition temperature of the bismaleimide resin to be predicted are performance prediction data.

4. A method for structural design of bismaleimide resin, characterized in that: include: Construct several virtual molecular structures of bismaleimide resins; Calculating a synthetic accessibility score for each of the virtual molecular structures; Inputting the virtual molecular structure into a dielectric constant prediction model to obtain corresponding dielectric constant prediction values; the dielectric constant prediction model is trained by the performance prediction model training method of the bismaleimide resin according to claim 1 or 2; Inputting the virtual molecular structure into a melting point prediction model to obtain corresponding melting point data prediction values; the melting point prediction model is trained by the performance prediction model training method of the bismaleimide resin according to claim 1 or 2; Inputting the virtual molecular structure into a thermal decomposition temperature prediction model to obtain corresponding 5% thermal decomposition temperature prediction values; the thermal decomposition temperature prediction model is trained by the performance prediction model training method of the bismaleimide resin according to claim 1 or 2; Calculating a comprehensive score for each virtual molecular structure according to the predicted dielectric constant value, the predicted melting point data value, the predicted 5% thermal decomposition temperature value, and the synthetic accessibility score; The design structure of the bismaleimide resin is determined according to the comprehensive score.

5. The structural design method of bismaleimide resin according to claim 4, characterized in that: The construction of several virtual molecular structures of bismaleimide resins specifically includes: Determine the genotype of the bismaleimide resin structure; Obtaining a number of corresponding genes in a database according to the gene type; Several genes are synthesized into several bismaleimide resins according to a preset template to obtain several virtual molecular structures of the bismaleimide resins.

6. The method for structural design of bismaleimide resin according to claim 5, wherein: The genotypes include: Maleic anhydride gene; diamine gene with only two amino groups at both ends of the main chain; ammonium acid gene with a benzene ring in the main chain connected to an amino group and a formic acid; diphenol gene with a benzene ring in the main chain connected to two hydroxyl groups; maleic acid diamine gene; halogenated hydrocarbon gene with halogen atoms at both ends of the main chain; amino alcohol gene with an amino group and a hydroxyl group connected to both ends of the main chain; diacid gene with only two formic acids at both ends of the main chain.

7. The structural design method of bismaleimide resin according to claim 4, characterized in that: The calculation formula for the comprehensive score is: Among them, Score, ε, T M 、T d 、S A Respectively represent the comprehensive score, dielectric constant of the virtual molecular structure, melting point, 5% thermal decomposition temperature, and synthetic accessibility score; ε min , ε max represent the minimum and maximum values of the dielectric constant of the virtual molecular structure; T Mmin 、T Mmax Respectively represent the minimum and maximum values of the melting point of the virtual molecular structure; T dmin 、T dmax They represent the minimum and maximum values of the 5% thermal decomposition temperature of the virtual molecular structure; S Amin 、S Amax represent the minimum and maximum values of the synthetic accessibility scores of the virtual molecular structures, respectively.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the performance prediction model training method for the bismaleimide resin according to any one of claims 1 to 2, the performance prediction method for the bismaleimide resin according to claim 3, or the structural design method for the bismaleimide resin according to any one of claims 4 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for training a performance prediction model of a bismaleimide resin according to any one of claims 1 to 2 or the method for predicting the performance of a bismaleimide resin according to claim 3 or the method for designing a structure of a bismaleimide resin according to any one of claims 4 to 7 is implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for training a performance prediction model of a bismaleimide resin according to any one of claims 1 to 2 or the method for predicting the performance of a bismaleimide resin according to claim 3 or the method for designing a structure of a bismaleimide resin according to any one of claims 4 to 7 is implemented.

Images