Polymer material secondary transformation temperature prediction method based on machine learning

Abstract

The invention discloses a polymer material secondary transformation temperature prediction method based on machine learning. The method comprises the following steps: obtaining a to-be-detected polymer material; constructing a secondary transformation temperature prediction model; inputting a to-be-detected polymer material into the secondary transformation temperature prediction model for prediction to obtain a prediction result; wherein the secondary transformation temperature prediction model is obtained after training through a training set and testing through a testing set. According to the method, the internal relation between the molecular structure of the polymer and the secondary transformation temperature (Tbeta) can be effectively captured under the condition of less data, and the rotation essence of the secondary transformation temperature (Tbeta) in the temperature change process of the polymer is disclosed.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence material research and development design technology, and in particular to a method for predicting the secondary transition temperature of a polymer material based on machine learning. Background Art

[0002] Secondary transition temperature (T β ) is an important thermodynamic property of polymer materials during temperature changes, reflecting the local mobility and flexibility of polymer chain segments, and is of great significance for material processing, performance regulation and application development. Secondary transition temperature (T β ) is crucial for the design and optimization of new materials. However, traditional experimental methods such as differential scanning calorimetry (DSC) and dynamic mechanical analysis (DMA) have limitations such as cumbersome sample preparation, long testing cycle, and difficult data interpretation. In addition, the secondary transition temperature (T β ) is scarce, which severely limits our in-depth understanding of its physical mechanism and its wide application.

[0003] In recent years, machine learning (ML), as an emerging data-driven method, has become an important tool in the field of materials science with its powerful data analysis capabilities. Through machine learning algorithms, we can mine the chemical structure and secondary transition temperature (T β ) and establish an efficient prediction model based on the existing experimental data, which greatly improves the accuracy and reliability of the prediction. β ) Due to the scarcity of data and its complex nonlinear relationship with molecular structure, the application of machine learning models in this field still faces certain challenges.

[0004] Therefore, how to make full use of limited experimental data and combine machine learning methods to develop a high-precision and high-generalization secondary transition temperature (T β ) prediction model has become an important topic in current materials science research. Summary of the invention

[0005] In order to solve the technical problems existing in the above-mentioned prior art, the present invention proposes a method for predicting the secondary transition temperature of polymer materials based on machine learning. Through this new method, the shortcomings of traditional experimental means can be effectively compensated, and a scientific basis can be provided for the design, processing and application of polymer materials, thereby promoting the widespread application of polymer materials in high-performance fields.

[0006] To achieve the above object, the present invention provides a method for predicting the secondary transition temperature of a polymer material based on machine learning, comprising:

[0007] obtaining a polymer material to be tested;

[0008] Construct a prediction model for secondary transition temperature;

[0009] Inputting the polymer material to be tested into the secondary transition temperature prediction model for prediction to obtain a prediction result;

[0010] The secondary transition temperature prediction model is obtained by training with a training set and testing with a test set.

[0011] Preferably, constructing the secondary transition temperature prediction model comprises:

[0012] Collecting the chemical structure formula of the target polymer and the corresponding secondary transition temperature data, dividing the secondary transition temperature data to obtain divided data;

[0013] Converting the chemical structure of the target polymer into a SMILES string, parsing the SMILES string, and extracting key structural features in the target polymer molecule;

[0014] The secondary transition temperature prediction model is established based on the key structural features and the divided data.

[0015] Preferably, dividing the secondary transition temperature data comprises:

[0016] The secondary transition temperature data are divided into data affected by the main chain and data affected by the side chain according to the rotation mechanism of the secondary transition temperature.

[0017] Preferably, converting the chemical structure formula of the target polymer into a SMILES string comprises:

[0018] The molecular structure diagram of the target polymer is drawn by CHEMDRAW software, and the two-dimensional structure of the target polymer is graphically represented;

[0019] The molecular structure diagram is converted into a simplified molecular input linear representation, namely the SMILES string, by CHEMDRAW software, and the quality control of the SMILES string is performed.

[0020] Preferably, parsing the SMILES string to extract key structural features in the polymer molecule includes:

[0021] Decomposing the given SMILES string into basic chemical structural units by molecular fragmentation method, extracting heteroatoms, chemical bond types and ring structure characteristics in the basic chemical structural units, and calculating the physicochemical characteristics of different structural units, wherein the physicochemical characteristics include but are not limited to rotational freedom, polarity, and flexibility;

[0022] Standardizing the heteroatoms, chemical bond types, and ring structure features, and performing data cleaning and outlier removal;

[0023] The extracted features are converted and processed according to different processing methods to generate feature sets corresponding to each method, and the different feature sets are integrated into comprehensive features of the target polymer, that is, key structural features in the target polymer molecule.

[0024] Preferably, the key structural features in the target polymer molecule include:

[0025] Chemical descriptors: total number of high electronegativity, number of fused rings, number of six-membered rings, number of five-membered rings, number of carbonyl groups, total number of single bonds, number of single bonds on the main chain, number of hydrogen bonds;

[0026] Based on the molecular weight at both ends of a single bond: the sum of the molecular weight differences of the single bonds with side chain groups at both ends, the sum of the molecular weights of the single bonds with side chain groups at both ends, the sum of the molecular weight differences of the single bonds without side chain groups on the main chain, and the sum of the molecular weights of the groups on the side chains.

[0027] Preferably, the different processing methods include molecular fragmentation method, condensed ring assignment method, benzene ring assignment method and atom normalization plus condensed ring / benzene ring assignment method.

[0028] Preferably, training the secondary transition temperature prediction model comprises:

[0029] Integrate the key structural features with the experimentally measured secondary transition temperature data to form an input data set and an output data set for model training, wherein the input data is the molecular features extracted by the molecular fragmentation method, and the output data is the secondary transition temperature value predicted by the model;

[0030] A machine learning algorithm is selected, modeling is performed according to the characteristics of the data, the secondary transition temperature prediction model is trained through a training set, and tested through a test set, and the prediction accuracy of the model is evaluated.

[0031] Preferably, the evaluation parameters used to test the secondary transition temperature prediction model are:

[0032]

[0033] In the formula, R 2 is the coefficient of determination; y i is the true value, is the predicted value, is the average value of the true value, n is the total number of samples, and i is the sample index value.

[0034] Compared with the prior art, the present invention has the following advantages and technical effects:

[0035] Compared with the traditional secondary transition temperature (T β ) data source, the present invention uses the fragmentation method to extract molecular structure features and combines it with machine learning models for training, which can provide higher prediction accuracy when data is scarce. By analyzing the chemical structure and physicochemical properties of the polymer, the method of the present invention can capture the rotational nature of the secondary transition temperature, surpassing the limitations of traditional experimental methods. In addition, the present invention has high flexibility, is applicable to different types of polymers, and can be widely used in the design and optimization of new materials, significantly improving the efficiency and accuracy of secondary transition temperature prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0037] Figure 1 This is a flow chart of a method for predicting secondary transition temperature of polymer materials based on machine learning according to an embodiment of the present invention;

[0038] Figure 2 A data distribution diagram affected by the main chain of the method for predicting the secondary transition temperature of polymer materials based on machine learning according to an embodiment of the present invention;

[0039] Figure 3 This is a data distribution diagram affected by side chains in the method for predicting the secondary transition temperature of polymer materials based on machine learning according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0041] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0042] At the secondary transition temperature (T β) range, the mobility of the polymer chain segments is significantly enhanced, which is manifested as local vibration and small single bond rotation, followed by a gradual transition to a larger range of segment motion. The rotation mechanism of the main chain is mainly affected by the groups connected to it (such as methyl, phenyl, acyl, etc.). The steric hindrance, charge distribution and polarity of these groups directly affect the rotational freedom and flexibility of the main chain. In this process, larger groups or polar groups usually weaken the rotation ability of the main chain, resulting in a higher secondary transition temperature (T β ) increases. At the same time, the rotation of the side chain is mainly dominated by the volume, polarity and high electronegative atoms of the side group. In particular, there is a close relationship between the molecular mobility of the side chain and the volume and polarity of the group. These factors further regulate the secondary transition temperature (T) by adjusting the rotational freedom of the single bond. β ). Secondary transition temperature (T β ) is essentially due to the rotational freedom of the single bond, so studying the behavior of single bonds is an important way to understand and predict the secondary transition temperature (T β ). By deeply analyzing the influence of single bond rotational freedom, the relationship between polymer chain segment mobility and secondary transition temperature (T β ) between them.

[0043] The present invention proposes a method for predicting the secondary transition temperature of polymer materials based on machine learning, comprising:

[0044] obtaining a polymer material to be tested;

[0045] Construct a prediction model for secondary transition temperature;

[0046] Inputting the polymer material to be tested into the secondary transition temperature prediction model for prediction to obtain a prediction result;

[0047] Among them, the secondary transition temperature prediction model is obtained after training with the training set and testing with the test set;

[0048] In this embodiment, the secondary transition temperature prediction model can also be used to predict the secondary transition temperature value of an unknown sample.

[0049] Further, constructing a secondary transition temperature prediction model includes (e.g. Figure 1 ):

[0050] Collecting the chemical structure formula of the target polymer and the corresponding secondary transition temperature data, dividing the secondary transition temperature data to obtain divided data;

[0051] Convert the chemical structure of the target polymer into a SMILES string, parse the SMILES string, and extract key structural features in the target polymer molecule;

[0052] The secondary transition temperature prediction model is established based on key structural features and the divided data.

[0053] This example combines regression analysis with fragmentation of polymer molecules to decompose them into multiple basic structural units to further study the secondary transition temperature (T β ) and explored the influence of the secondary transition temperature (T β ), such as the rotational freedom of single bonds, the flexibility of molecular segments, and intramolecular interactions. In order to effectively capture these influencing factors, molecular fragmentation method, condensed ring assignment method, benzene ring assignment method, atomic normalization and other methods and techniques were used.

[0054] Specifically, the chemical structure formula and the corresponding secondary transition temperature (T) of the target polymer were collected from literature or books. β ) data, and according to the secondary transition temperature (T β )’s rotation mechanism divides the data into two categories: data affected by the main chain and data affected by the side chain. The distribution of data sets affected by the main chain is as follows: Figure 2 As shown, the distribution of data sets affected by side chains is as follows Figure 3 shown.

[0055] Further, converting the chemical structure of the target polymer into a SMILES string includes:

[0056] Use CHEMDRAW software to draw the molecular structure diagram of the polymer and graphically represent the two-dimensional structure of the polymer, providing a basis for subsequent data processing and analysis;

[0057] The drawn molecular structure diagram is converted into a simplified molecular input linear representation (SMILES string) through CHEMDRAW software to facilitate subsequent calculations and analysis;

[0058] The generated SMILES string is quality controlled to ensure that the converted SMILES string is accurate, standardized, and can reflect the true chemical structure of the polymer.

[0059] The chemical structure of a specific polymer is converted into a SMILES string. The processing process includes:

[0060] Convert the chemical structure of the polymer into a standard SMILES string, ensuring accurate representation of each atom and bond;

[0061] In the SMILES structure, the "*" symbol is used to mark the two ends of the polymer to indicate the start and end of the polymer chain, and is also used to indicate the repeating unit part;

[0062] For polymers with ring structures, ensure accurate representation of ring closure and size, and avoid duplication or loss of ring labels; when dealing with polymers containing stereochemical information, add stereochemical symbols for cis / trans isomers as appropriate;

[0063] Clean and standardize the structure formula to remove redundant or unnecessary repeated information and make it conform to the SMILES standard;

[0064] The polymer structure is further broken down to identify the main chain and side chain parts, preparing for subsequent classification and analysis.

[0065] Furthermore, the SMILES string is parsed to extract the key structural features of the polymer molecule, including:

[0066] By molecular fragmentation method, a given SMILES string is decomposed into basic chemical structural units, heteroatoms, chemical bond types and ring structure characteristics in the basic chemical structural units are extracted, and the physicochemical characteristics of different structural units are calculated, wherein the physicochemical characteristics include but are not limited to rotational freedom, polarity, and flexibility;

[0067] Standardize the types of heteroatoms, chemical bonds, and ring structure features, and perform data cleaning and outlier removal;

[0068] The extracted features are converted and processed according to different processing methods to generate feature sets corresponding to each method, and the different feature sets are integrated into the comprehensive features of the target polymer, that is, the key structural features in the target polymer molecule.

[0069] During the assignment process, the influence of various structural units on the secondary transition temperature (T β ) and transform these influencing factors into numerical features that can be used by the model through quantitative analysis, so as to facilitate subsequent machine learning modeling and secondary transition temperature (T β )Predictive analysis.

[0070] Specifically, the characteristic value is extracted by the molecular fragmentation method, and the processing process specifically includes:

[0071] Split the polymer molecular structure into individual fragments or substructures (such as rings, heteroatoms, etc.);

[0072] Identify and classify individual chemical fragments, ensuring they are accurately labeled as single bonds, double bonds, ring structures, or heteroatoms;

[0073] Calculate the physicochemical characteristics of each fragment, including but not limited to rotational degrees of freedom, polarity, flexibility, etc.;

[0074] Standardize the extracted fragments to ensure that each fragment complies with the rules of molecular fragmentation;

[0075] Perform data cleaning and outlier removal to ensure the accuracy and consistency of feature values;

[0076] The extracted features are converted and processed according to different processing methods to generate feature sets corresponding to each method. Finally, these processed feature sets are integrated into comprehensive features of polymers as input data for subsequent modeling and analysis.

[0077] Furthermore, the data feature descriptors affected by the main chain specifically include: the total number of high electronegativity, the number of condensed rings, the number of six-membered rings, the number of five-membered rings, the number of carbonyl groups, the number of single bonds, the number of single bonds on the main chain, the number of hydrogen bonds, the sum of the molecular weight difference of the side chain groups at both ends of the single bond, the sum of the molecular weight difference of the side chain groups at both ends of the single bond, the sum of the molecular weight difference of the main chain single bond without side chain groups, and the sum of the molecular weight of the groups on the side chains. Among them, the high electronegativity atoms include O, N, F, Cl, and Br.

[0078] The characteristic descriptors of the side chain data specifically include: the total number of high electronegativity, the number of carbonyl groups, the number of hydrogen bonds, the number of branched single bonds, and the sum of molecular weights on the branches. Among them, high electronegativity atoms include O, N, F, Cl, and Br.

[0079] Furthermore, different processing methods include molecular fragmentation method, condensed ring assignment method, benzene ring assignment method, and atom normalization plus condensed ring / benzene ring assignment method.

[0080] Specifically, the molecular fragmentation method, the characteristic descriptors of this method include: the sum of the molecular weight difference of the side chain groups at both ends of the single bond of the polymer, the sum of the molecular weight of the side chain groups at both ends of the single bond, the total number of high electronegativity, the number of carbonyl groups, the number of hydrogen bonds, and the number of single bonds. Among them, high electronegativity atoms include O, N, F, Cl, and Br.

[0081] The condensed ring assignment method extracts specific structural features by identifying and assigning values ​​to condensed ring structures in polymer molecules, and quantifies them according to the size, number and atomic composition characteristics of the rings, thereby providing key structural information for polymer property prediction and model building. The characteristic descriptors of this method include: the total number of high electronegativity, the number of condensed rings, the number of six-membered rings, the number of five-membered rings, the number of carbonyl groups, the number of single bonds, the number of single bonds on the main chain, the number of hydrogen bonds, the sum of the molecular weight differences of the groups at both ends of the single bonds on the main chain, and the sum of the molecular weights on the side chains. Among them, the high electronegativity species are O, N, F, Cl, and Br.

[0082] The benzene ring assignment method extracts specific structural features by identifying and assigning values ​​to the benzene ring structure in the polymer molecule, and quantifies them according to the number, position and connection mode of the benzene rings with other groups, providing important structural information for the prediction of polymer properties and model building. The characteristic descriptors of this method include: the total number of high electronegativity, the number of five-membered rings, the number of non-five-membered rings, the number of carbonyl groups, the number of single bonds, the number of single bonds on the main chain, the number of hydrogen bonds, the sum of the molecular weight differences of the groups at both ends of the single bonds on the main chain, and the sum of the molecular weights on the side chains. Among them, the high electronegativity species are O, N, F, Cl, and Br.

[0083] Atomic normalization plus condensed ring / benzene ring assignment method, atomic normalization standardizes the characteristics of each atom in the polymer molecule, eliminates the influence of different atomic types or molecular structure scale differences on the characteristic values, ensures that the contribution of each atom in the model is comparable, and thus improves the stability and prediction accuracy of the model. The characteristic descriptors of this method include: the total number of high electronegativity, the number of condensed rings, the number of five-membered rings, the number of non-five-membered rings, the number of carbonyl groups, the number of single bonds, the number of single bonds on the main chain, the number of hydrogen bonds, the sum of the molecular weight differences of the groups at both ends of the single bonds on the main chain, and the sum of the molecular weights on the side chains. Among them, the high electronegativity species are O, N, F, Cl, and Br.

[0084] Finally, all extracted eigenvalues ​​were normalized and then expanded 100 times, providing consistent, comparable, and amplified structural information for subsequent polymer property prediction and model building.

[0085] Based on the selection of eigenvalues ​​and corresponding data processing, the data affected by the main chain and the side chain are analyzed and shown in Table 1.

[0086] Table 1

[0087]

[0088]

[0089]

[0090] Furthermore, the condensed ring assignment method uses the following formula for calculation:

[0091] For the number of C on the ring = 6, the corresponding assignment is a;

[0092] For the number of C on the ring < 6, the corresponding value is 0.5a;

[0093] For the number of C on the ring > 6, the corresponding assignment is 2a; where the value of a is 5.

[0094] In the atomic normalization plus condensed ring assignment method, normalization is calculated using the following formula:

[0095]

[0096] The benzene ring assignment method introduces several different algorithms, including only assigning rings, main chain single bonds plus hydrogen bonds, main chain single bonds without hydrogen bonds, single bonds plus hydrogen bonds, and single bonds without hydrogen bonds. The following equation is used for calculation:

[0097]

[0098] Among them, N i refers to the number of atoms appearing in the repeating unit, ζ i Refers to the effective mobility of each atom in the repeating unit. For a single bond, ζ i is 0.2a, and the ζ corresponding to the hydrogen bond i is 0.4a, and the ζ corresponding to a single bond i is 0.2a, for high electronegativity corresponding to ζ i is 0.6a, and the ζ corresponding to the carbonyl group i is 0.7a, and the ζ corresponding to the five-membered ring i is 0.5a, and for non-five-membered rings, ζ i a+w a (n-1), where n refers to the number of polycyclic rings, a is 5, and w a Take 1.5.

[0099] The results show that the secondary transition temperature (T β ) is mainly the ease of rotation at both ends of the single bond. The ease of rotation can be quantified by multiple key structural features, including high electronegative atoms, C=O, hydrogen bonds, molecular weight at both ends of the single bond, molecular weight of the branched chain and other factors. Specifically, high electronegative atoms will attract the electron cloud in the covalent bond, resulting in an uneven distribution of electron density at both ends of the single bond. Due to the asymmetric distribution of this electron cloud, the polarity of the single bond is enhanced, resulting in stronger electrostatic interactions between molecules. These interactions increase the energy barrier for single bond rotation, making rotation at both ends of the single bond more difficult, thereby limiting the flexibility and rotational freedom of the chain segment.

[0100] The C=O group itself has a strong polarity, which makes the two ends of the single bond connected to it also show a strong polarity difference. This polarity difference increases the energy barrier of the single bond, making the rotation of the single bond more difficult, thereby further limiting the rotational freedom of the single bond.

[0101] Hydrogen bonds enhance the rigidity of the chain segments through intermolecular forces, strengthen the stability of the molecular structure, and reduce the rotational freedom of the chain segments, thereby increasing the secondary transition temperature (T β) will increase. The molecular weight of the groups at both ends of the single bond is also a key factor affecting the ease of rotation. Larger groups are usually accompanied by stronger intermolecular interactions, which enhances the rigidity of the molecule, making the single bond rotation of the main chain face greater rotation resistance, thereby increasing the energy required to overcome the rotation obstacle.

[0102] The effect of branch distribution on the secondary transition temperature (T β The length, rigidity, distribution density and position of the side chains will affect the flexibility and rigidity of the polymer by changing the relative mobility of the polymer chain segments and the interaction between molecules, thereby affecting the secondary transition temperature (T β ).

[0103] Table 2 shows the side chain descriptors and R 2 Analysis of relationships.

[0104] Table 2

[0105]

[0106] The results show that highly electronegative atoms enhance the interaction between molecules through strong electronic effects, thereby restricting the movement of chain segments and leading to a higher secondary transition temperature (T β ) increases. The carbonyl group on the main chain will reduce the rotational freedom of the side chain through the steric hindrance effect, so that the side chain cannot rotate or move freely as in the case without the carbonyl group. In addition, the carbonyl group on the side chain further reduces the rotational freedom of the side chain due to its strong polar effect, increasing the rigidity of the polymer.

[0107] The distribution of the side chains has a direct relationship with the secondary transition temperature. When there are multiple single bonds in the side chains, the free rotation of these single bonds will increase the rotational freedom of the side chains, allowing the polymer segments to begin to transform or deform at lower temperatures. However, if the side chain structure contains double bonds or ring structures, the freedom of single bond rotation will be more restricted. Double bonds or ring structures will increase the rigidity of the polymer chain and inhibit its rotational freedom, thus increasing the secondary transition temperature (T β ) increased.

[0108] Further, training the secondary transition temperature prediction model includes:

[0109] Integrate the key structural features with the experimentally measured secondary transition temperature data to form an input data set and an output data set for model training, wherein the input data is the molecular features extracted by the molecular fragmentation method, and the output data is the secondary transition temperature value predicted by the model;

[0110] A machine learning algorithm is selected, and modeling is performed according to the characteristics of the data. The secondary transition temperature prediction model is trained through a training set, and tested through a test set, and the prediction accuracy of the model is evaluated. The training set and the test set are feature values ​​extracted by different methods and the corresponding secondary transition temperature data.

[0111] Specifically, in this embodiment, 80% of the preprocessed data is randomly selected as a training set and 20% as a test set;

[0112] The extracted molecular features were compared with the experimentally measured secondary transition temperatures (T β ) data to form the input and output data sets for model training, where the input data are the molecular features extracted by feature extraction technology, and the output data are the secondary transition temperature (T β )value;

[0113] Select appropriate machine learning algorithms, such as regression analysis methods, decision trees, random forests, gradient boosting, etc., and select appropriate algorithms for modeling based on the characteristics of the data;

[0114] The prediction model is obtained through training, which can predict the secondary transition temperature (T β ) and can be used to determine the secondary transition temperature (T β ) value prediction;

[0115] The model is validated, the model prediction results are verified based on experimental data, the prediction accuracy of the model is evaluated, and the main characteristics of the model are analyzed to ensure the reliability and accuracy of the model.

[0116] The evaluation parameters used to test the secondary transition temperature prediction model are:

[0117]

[0118] In the formula, R 2 is the coefficient of determination; y i is the true value, is the predicted value, is the average value of the true value, n is the total number of samples, and i is the sample index value.

[0119] Furthermore, based on the model prediction results, the rotation mechanism of the secondary transition temperature is analyzed and explained, including:

[0120] The key features and secondary transition temperature (T β ) and identify the secondary transition temperature (T β) have a significant impact on structural features, such as the rotational freedom of single bonds, the flexibility of molecular segments, the polarity of covalent bonds, the size of the ring and the influence of highly electronegative atoms;

[0121] Combined with the rotation mechanism of the secondary transition temperature, the model analysis results are compared with the existing literature on the secondary transition temperature (T β ) rotation mechanism, to explain and verify how each structural feature regulates the secondary transition temperature (T) by affecting molecular motion, chain segment flexibility and intermolecular interactions. β );

[0122] According to the model analysis results, the influence of the secondary transition temperature (T β ), such as the mechanism of action affected by the main chain and side chain, and the effect of intramolecular interactions on the secondary transition temperature (T β ) to optimize the secondary transition temperature (T β ) provide theoretical guidance.

[0123] The method of the present invention uses descriptors of different methods to analyze the secondary transition temperature (T β ) is a rotational nature. The prediction method developed by the present invention can be applied to the design and optimization of polymer materials, especially in the case of scarce data, and can effectively predict the secondary transition temperature (T β ), and provide theoretical support for the performance regulation of materials. At the same time, this method can also be widely used in the development of new polymer materials, the screening of functional polymers and their performance evaluation in engineering applications.

[0124] Given the secondary transition temperature (T β ) research data is scarce. By combining regression models with multiple feature extraction techniques, a variety of solutions are provided for the research of small amounts of data, effectively overcoming the problem of insufficient data in traditional experimental methods. Through this innovative analysis framework, the present invention not only improves scientific research efficiency, but also provides a new method for the secondary transition temperature (T β ) provides a new idea and theoretical support for the in-depth study of the secondary transition temperature (T β )’s behavior and application potential in polymer materials have opened up new research paths and promoted technological innovation in the field of materials science.

[0125] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for predicting secondary transition temperature of polymer materials based on machine learning, characterized in that: include: obtaining a polymer material to be tested; Construct a prediction model for secondary transition temperature; Inputting the polymer material to be tested into the secondary transition temperature prediction model for prediction to obtain a prediction result; The secondary transition temperature prediction model is obtained by training with a training set and testing with a test set.

2. The method for predicting secondary transition temperature of polymer materials based on machine learning according to claim 1, characterized in that: Constructing the secondary transition temperature prediction model includes: Collecting the chemical structure formula of the target polymer and the corresponding secondary transition temperature data, dividing the secondary transition temperature data to obtain divided data; Converting the chemical structure of the target polymer into a SMILES string, parsing the SMILES string, and extracting key structural features in the target polymer molecule; The secondary transition temperature prediction model is established based on the key structural features and the divided data.

3. The method for predicting secondary transition temperature of polymer materials based on machine learning according to claim 2, characterized in that: Classifying the secondary transition temperature data includes: The secondary transition temperature data are divided into data affected by the main chain and data affected by the side chain according to the rotation mechanism of the secondary transition temperature.

4. The method for predicting secondary transition temperature of polymer materials based on machine learning according to claim 2, characterized in that: Converting the chemical structure of the target polymer into a SMILES string includes: The molecular structure diagram of the target polymer is drawn by CHEMDRAW software, and the two-dimensional structure of the target polymer is graphically represented; The molecular structure diagram is converted into a simplified molecular input linear representation, namely the SMILES string, by CHEMDRAW software, and the quality control of the SMILES string is performed.

5. The method for predicting secondary transition temperature of polymer materials based on machine learning according to claim 4, characterized in that: The SMILES string is parsed to extract key structural features in the polymer molecule including: Decomposing the given SMILES string into basic chemical structural units by molecular fragmentation method, extracting heteroatoms, chemical bond types and ring structure characteristics in the basic chemical structural units, and calculating the physicochemical characteristics of different structural units, wherein the physicochemical characteristics include but are not limited to rotational freedom, polarity, and flexibility; Standardizing the heteroatoms, chemical bond types, and ring structure features, and performing data cleaning and outlier removal; The extracted features are converted and processed according to different processing methods to generate feature sets corresponding to each method, and the different feature sets are integrated into comprehensive features of the target polymer, that is, key structural features in the target polymer molecule.

6. The method for predicting secondary transition temperature of polymer materials based on machine learning according to claim 5, characterized in that: The key structural features in the target polymer molecule include: Chemical descriptors: total number of high electronegativity, number of fused rings, number of six-membered rings, number of five-membered rings, number of carbonyl groups, total number of single bonds, number of single bonds on the main chain, number of hydrogen bonds; Based on the molecular weight at both ends of a single bond: the sum of the molecular weight differences of the single bonds with side chain groups at both ends, the sum of the molecular weights of the single bonds with side chain groups at both ends, the sum of the molecular weight differences of the single bonds without side chain groups on the main chain, and the sum of the molecular weights of the groups on the side chains.

7. The method for predicting secondary transition temperature of polymer materials based on machine learning according to claim 6, characterized in that: The different processing methods include molecular fragmentation method, condensed ring assignment method, benzene ring assignment method and atom normalization plus condensed ring / benzene ring assignment method.

8. The method for predicting secondary transition temperature of polymer materials based on machine learning according to claim 2, characterized in that: Training the secondary transition temperature prediction model comprises: Integrate the key structural features with the experimentally measured secondary transition temperature data to form an input data set and an output data set for model training, wherein the input data is the molecular features extracted by the molecular fragmentation method, and the output data is the secondary transition temperature value predicted by the model; A machine learning algorithm is selected, modeling is performed according to the characteristics of the data, the secondary transition temperature prediction model is trained through a training set, and tested through a test set, and the prediction accuracy of the model is evaluated.

9. The method for predicting secondary transition temperature of polymer materials based on machine learning according to claim 8, characterized in that: The evaluation parameters used to test the secondary transition temperature prediction model are: In the formula, R 2 is the coefficient of determination; y i is the true value, is the predicted value, is the average value of the true value, n is the total number of samples, and i is the sample index value.

Images