Method and device for designing and screening polymer gas separation membrane and storage medium

The integration of high-throughput molecular dynamics simulation and machine learning optimizes polymer gas separation membranes by predicting performance using a hierarchical graph convolutional neural network, addressing limitations in traditional membranes and improving separation efficiency and selectivity.

CN119075681AActive Publication Date: 2024-12-06TSINGHUA UNIVERSITY

Patent Information

Application Number
CN202411230657.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-12-06
Estimated Expiration
2044-09-03

AI Technical Summary

Technical Problem

Traditional polymer gas separation membranes face limitations in separation efficiency and selectivity due to their inherent physical and chemical properties, and existing data in polymer material databases are limited by experimental conditions, making large-scale material screening and optimization difficult.

Method used

A method combining high-throughput molecular dynamics simulation with machine learning to rapidly screen and characterize polymer materials, using a membrane performance prediction model based on a hierarchical graph convolutional neural network to predict and optimize polymer gas separation membranes.

Benefits of technology

Enables efficient and accurate screening of polymer materials for gas separation membranes, improving separation performance by predicting and determining solubility and diffusion coefficients, thereby enhancing gas permeability and selectivity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119075681A_ABST
    Figure CN119075681A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of high-throughput molecular dynamics simulation, and discloses a method and a device for designing and screening a high-molecular gas separation membrane and a storage medium. Monomer information of a plurality of polymer monomers is obtained; generating a membrane search space based on the monomer information; performing performance prediction on the plurality of candidate simulation polymer gas separation membranes in the search space based on a pre-trained membrane performance prediction model; based on the performance prediction result, screening out at least one target simulation polymer gas separation membrane meeting the expected performance from the membrane search space; performing Monte Carlo simulation and molecular dynamics simulation on each target simulation polymer gas separation membrane based on a molecular dynamics simulation tool to obtain a solubility coefficient and a diffusion coefficient; based on the solubility coefficient and the diffusion coefficient, determining the performance index of the target simulation polymer gas separation membrane; a large number of non-existing simulated membrane materials can be formed for analysis, and the screening efficiency of a large number of high polymer materials is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the technical field of high-throughput molecular dynamics simulation, and in particular to a design and screening method, device and storage medium for a polymer gas separation membrane. Background Art

[0002] Polymer gas separation membranes are a type of thin film material used for gas separation and purification. They are made of polymers and can selectively pass certain gas molecules based on the size, shape and chemical properties of the gas molecules. The gas separation principle of polymer gas separation membranes is generally based on the dissolution-diffusion mechanism, that is, the gas molecules first dissolve on the surface of one side of the membrane, then diffuse inside the membrane, and finally release from the other side of the membrane. Polymer gas separation membranes are widely used in gas separation and purification application scenarios such as natural gas processing, air separation and pollution control due to their low energy consumption, simple operation and easy large-scale production.

[0003] Traditional polymer gas separation membranes have limitations in separation efficiency and selectivity due to their inherent physical and chemical properties. For example, traditional polymer gas separation membranes often sacrifice selectivity while improving permeability, which is determined by the microstructure of the material and the diffusion mechanism of gas molecules. Therefore, in order to improve the separation performance of membrane materials, it is necessary to optimize the structure and properties of traditional polymer gas separation membranes.

[0004] A typical design method for a polymer gas separation membrane includes: designing a polymer gas separation membrane based on the performance indicated by data in a polymer material database and literature.

[0005] However, most of the data in polymer material databases and literature are experimental values. These data are limited by the diversity of experimental conditions and methods, making it difficult to conduct large-scale material screening and optimization design. Summary of the invention

[0006] In view of this, the present disclosure proposes a design and screening method, device and storage medium for polymer gas separation membranes, which can overcome the limitations of traditional molecular dynamics simulation in high-throughput processing, and at the same time combine machine learning technology to achieve rapid screening and accurate characterization of a large number of polymer materials, thereby achieving rapid, efficient and accurate screening and characterization of a large number of polymer materials.

[0007] According to one aspect of the present disclosure, a method for designing and screening a polymer gas separation membrane is provided, the method comprising:

[0008] Obtain monomer information of multiple polymer monomers;

[0009] generating a membrane search space based on the monomer information, wherein the membrane search space includes a plurality of candidate simulated polymer gas separation membranes;

[0010] Based on a pre-trained membrane performance prediction model, a plurality of candidate simulated polymer gas separation membranes in the search space are predicted for performance, and a performance prediction result of each candidate simulated polymer gas separation membrane is obtained; the membrane performance prediction model is obtained by training based on a machine learning algorithm;

[0011] Screening out at least one target simulated polymer gas separation membrane that meets the desired performance from the membrane search space based on the performance prediction result;

[0012] Performing Monte Carlo simulation on each target simulated polymer gas separation membrane based on a molecular dynamics simulation tool to determine the solubility coefficient of the target simulated polymer gas separation membrane for different gases; and performing molecular dynamics simulation on each target simulated polymer gas separation membrane based on the molecular dynamics simulation tool to determine the diffusion coefficient of the target simulated polymer gas separation membrane for different gases;

[0013] Based on the solubility coefficient and the diffusion coefficient, the performance index of the target simulated polymer gas separation membrane is determined, and the performance index includes the gas permeability coefficient and / or the selectivity of the membrane.

[0014] The performance prediction of multiple candidate simulated polymer gas separation membranes in the search space is performed based on the pre-trained membrane performance prediction model to obtain a performance prediction result of each candidate simulated polymer gas separation membrane, including:

[0015] For each candidate simulated polymer gas separation membrane, multi-scale information of the candidate simulated polymer gas separation membrane is obtained; wherein the multi-scale information includes at least microscopic information and mesoscopic information; wherein the microscopic information is used to indicate the microscopic properties of the candidate simulated polymer gas separation membrane, and the mesoscopic information is used to indicate the mesoscopic properties of the candidate simulated polymer gas separation membrane;

[0016] The multi-scale information is input into the membrane performance prediction model to obtain the performance prediction result; wherein, the membrane performance prediction model is established based on a hierarchical graph convolutional neural network model.

[0017] In a possible implementation, the membrane performance prediction model includes, in order according to the information transmission direction:

[0018] An input layer, used for inputting the multi-scale information;

[0019] A convolutional layer, used for extracting features based on the multi-scale information;

[0020] An attention pooling layer, used to determine an attention score of the feature data output by the convolutional layer; and weighted pool the feature data based on the attention score to obtain processed feature data;

[0021] The output layer is used to generate the performance prediction result based on the processed feature data.

[0022] In one possible implementation, the attention pooling layer is specifically used to determine the attention score of the feature data based on a multi-head attention mechanism guided by a mask matrix.

[0023] In a possible implementation, the microscopic information includes: a SMILES string of at least one polymer monomer forming the candidate simulated polymer gas separation membrane, and / or, atom and bond information; wherein the atom and bond information is generated based on the SMILES string;

[0024] The mesoscopic information includes: the inter-monomer linking mode of the polymer monomers.

[0025] In a possible implementation, the multi-scale information further includes: macroscopic information, where the macroscopic information is used to indicate the macroscopic properties of the candidate simulated polymer gas separation membrane.

[0026] In a possible implementation, the macroscopic information includes: free volume fraction FFV and / or glass transition temperature Tg, wherein FFV refers to the proportion of space not occupied by molecules in the candidate simulated polymer gas separation membrane; the glass transition temperature Tg is the temperature at which the candidate simulated polymer gas separation membrane transforms from a glassy state to a rubbery state.

[0027] In a possible implementation, the method further includes: performing performance prediction on a plurality of candidate simulated polymer gas separation membranes in the search space based on the pre-trained membrane performance prediction model, and obtaining the performance prediction result of each candidate simulated polymer gas separation membrane before obtaining the performance prediction result of each candidate simulated polymer gas separation membrane:

[0028] Generate samples to simulate polymer gas separation membranes;

[0029] Obtaining sample microscopic information and sample mesoscopic information of the sample simulated polymer gas separation membrane;

[0030] Simulating the molecular dynamics properties of each candidate simulated polymer gas separation membrane based on the molecular dynamics simulation tool to obtain simulated performance parameters of the sample simulated polymer gas separation membrane;

[0031] The membrane performance prediction model is obtained by training a pre-created hierarchical graph convolutional neural network model based on the sample microscopic information, the sample mesoscopic information and the simulation performance parameters.

[0032] In a possible implementation, the obtaining of monomer information of a plurality of polymer monomers includes:

[0033] Receive the SMILES string of the polymer monomer;

[0034] and / or,

[0035] Receiving a polymer single-chain model; parsing the polymer single-chain model to obtain the monomer information.

[0036] According to another aspect of the present disclosure, a design and screening device for polymer gas separation membranes is provided, comprising: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0037] According to another aspect of the present disclosure, a non-volatile computer-readable storage medium is provided, on which computer program instructions are stored, wherein the computer program instructions implement the above method when executed by a processor.

[0038] According to another aspect of the present disclosure, a computer program product is provided, including a computer-readable code, or a non-volatile computer-readable storage medium carrying the computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above method.

[0039] By obtaining monomer information of multiple polymer monomers; generating a membrane search space based on the monomer information, the membrane search space includes multiple candidate simulated polymer gas separation membranes; predicting the performance of multiple candidate simulated polymer gas separation membranes in the search space based on a pre-trained membrane performance prediction model, and obtaining a performance prediction result for each candidate simulated polymer gas separation membrane; screening out at least one target simulated polymer gas separation membrane that meets the expected performance from the membrane search space based on the performance prediction result; performing Monte Carlo simulation on each target simulated polymer gas separation membrane based on a molecular dynamics simulation tool to determine the solubility coefficient of the target simulated polymer gas separation membrane for different gases; and performing molecular dynamics simulation on each target simulated polymer gas separation membrane based on a molecular dynamics simulation tool to determine the diffusion coefficient of the target simulated polymer gas separation membrane for different gases; determining the performance indicators of the target simulated polymer gas separation membrane based on the solubility coefficient and the diffusion coefficient, the performance indicators including the gas permeability coefficient and / or the selectivity of the membrane; generating a search space based on the monomer information, thereby forming a large number of non-existing simulated membrane materials for analysis, solving the problem that the data is limited by the diversity of experimental conditions and methods, making it difficult to conduct large-scale material screening and optimization design, and providing the possibility for designing new membrane materials. At the same time, the membrane performance prediction model is used to search for feasible target simulated polymer gas separation membranes in the search space, eliminating the need to manually screen the performance of a large number of simulated membrane materials in the search space, which can improve the efficiency of screening a large number of polymer materials.

[0040] Further features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the disclosure and, together with the description, serve to explain the principles of the disclosure.

[0042] Figure 1 A flow chart showing a method for designing and screening a polymer gas separation membrane according to an embodiment of the present disclosure;

[0043] Figure 2 A structural diagram showing a membrane performance prediction model according to an embodiment of the present disclosure;

[0044] Figure 3 A schematic diagram showing multi-scale information according to an embodiment of the present disclosure;

[0045] Figure 4 A schematic diagram showing microscopic information according to an embodiment of the present disclosure;

[0046] Figure 5 A block diagram showing a design and screening device for a polymer gas separation membrane according to an embodiment of the present disclosure;

[0047] Figure 6 A block diagram showing a design and screening device for a polymer gas separation membrane according to another embodiment of the present disclosure. DETAILED DESCRIPTION

[0048] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0049] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0050] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present disclosure.

[0051] First, several terms involved in this application are introduced.

[0052] Polymer chain: A long chain of molecules formed by many repeated monomer units (i.e. polymer monomers) connected by chemical bonds. These monomer units are linked together through polymerization reactions to form polymers with specific molecular weights and structures. The length of polymer chains can range from a few thousand atoms to millions of atoms, which gives polymers unique physical and chemical properties.

[0053] Polymer: refers to the basic molecular unit that constitutes the polymer chain. In the polymerization reaction, polymer monomers are linked together by covalent bonds to form long chains of polymer molecules (i.e. the above-mentioned polymer chain). Each polymer monomer has one or more functional groups that can react with other monomers. For example, polyethylene is a polymer chain formed by the polymerization of ethylene monomers. Each ethylene monomer contains a double bond. These double bonds are opened in the polymerization reaction and connected to the double bonds of other ethylene monomers to form long chains of polyethylene molecules.

[0054] Polymer single chain model: a mathematical model or computational model used to describe the physical and chemical properties of a single polymer chain. This model usually combines factors such as the conformation, flexibility, and interaction of the polymer chain to predict and explain the macroscopic properties of polymer materials.

[0055] Hierarchical Graph Convolutional Networks (HGCN) model: a deep learning model that combines a hierarchical structure with a graph convolutional neural network (GCN). HGCN is used to process graph data with a hierarchical or layered structure, for example, to process the multi-scale simulated polymer gas separation membrane structure data in this application.

[0056] Figure 1 A flow chart of a method for designing and screening a polymer gas separation membrane according to an embodiment of the present disclosure is shown. In this embodiment, the method is described by taking an electronic device with computing capabilities as an example. The electronic device may be a user terminal or a server. The user terminal may be a computer, a tablet computer, etc. This embodiment does not limit the implementation of the electronic device. Figure 1 As shown, the method comprises at least the following steps:

[0057] Step 101, obtaining monomer information of a plurality of polymer monomers.

[0058] Monomer information is used to describe polymer monomers. Schematically, obtaining monomer information of multiple polymer monomers includes: receiving a simplified molecular linear input specification (Simplified Molecular Input Line Entry System, SMILES) string of the polymer monomer; and / or receiving a polymer single chain model; parsing the polymer single chain model to obtain monomer information.

[0059] Optionally, the molecular dynamics simulation tool running in the electronic device provides a human-computer interaction interface to receive a SMILES string describing a polymer monomer or a polymer single chain model. In this embodiment, the molecular dynamics simulation tool can simulate various dynamic motion conditions of the polymer gas separation membrane when separating gases. Among them, the molecular dynamics simulation tool includes but is not limited to: Large-scale Atomic / Molecular MassivelyParallel Simulator (LAMMPS), Highly Optimized Object-oriented Many-particle Dynamics-blue (HOOMD-blue) and other open source molecular dynamics simulation software packages suitable for simulating dynamic systems composed of polymers and gases. This embodiment does not limit the implementation method of the molecular dynamics simulation tool.

[0060] Optionally, the polymer single-chain model is parsed to obtain monomer information, including: identifying at least one repeating unit in the polymer single-chain model to obtain at least one monomer; and generating SMILES strings describing each monomer to obtain monomer information.

[0061] Optionally, in other embodiments, the electronic device may also read monomer information of the polymer monomer from an existing material library.

[0062] Step 102 : generating a membrane search space based on monomer information, wherein the membrane search space includes a plurality of candidate simulated polymer gas separation membranes.

[0063] The membrane search space is used to provide a variety of possible simulated polymer gas separation membranes.

[0064] In one example, generating a membrane search space based on monomer information includes: randomly selecting at least one monomer information from a plurality of acquired monomer information, randomly arranging, and copolymerizing to form a candidate simulated polymer gas separation membrane. The monomer information constituting different candidate simulated polymer gas separation membranes is different, and / or the arrangement of the monomer information constituting different candidate simulated polymer gas separation membranes is different.

[0065] Schematically, the method of copolymerizing monomer information to generate candidate simulated polymer gas separation membranes includes: using RDKit and Open Babel toolkits to convert SMILES strings into simulated single-chain polymers to obtain candidate simulated polymer gas separation membranes.

[0066] In other embodiments, the method of generating the membrane search space based on the monomer information may also be: generating candidate simulated polymer gas separation membranes based on the monomer information and arrangement specified by the user. This embodiment does not limit the method of generating the membrane search space.

[0067] Optionally, after generating a plurality of candidate simulated polymer gas separation membranes, the rationality of the candidate simulated polymer gas separation membranes can also be determined, and the candidate simulated polymer gas separation membranes whose rationality does not meet the desired target can be filtered to obtain a membrane search space. The rationality of the candidate simulated polymer gas separation membrane can be represented by a value calculated by a Qe model or a Flory-Reaction Probability Factor model, and this embodiment does not limit the method for determining the rationality.

[0068] Step 103 , based on the pre-trained membrane performance prediction model, the performance of multiple candidate simulated polymer gas separation membranes in the search space is predicted to obtain a performance prediction result for each candidate simulated polymer gas separation membrane.

[0069] The membrane performance prediction model is obtained based on training of a machine learning algorithm, which includes but is not limited to a graph neural network model, a hierarchical graph convolutional neural network model, etc. This embodiment does not limit the implementation of the machine learning algorithm.

[0070] In one example, a membrane performance prediction model is established based on a hierarchical graph convolutional neural network model. Accordingly, based on the pre-trained membrane performance prediction model, a plurality of candidate simulated polymer gas separation membranes in the search space are predicted for performance, and a performance prediction result of each candidate simulated polymer gas separation membrane is obtained, including: for each candidate simulated polymer gas separation membrane, multi-scale information of the candidate simulated polymer gas separation membrane is obtained. The multi-scale information includes at least microscopic information and mesoscopic information; the multi-scale information is input into the membrane performance prediction model to obtain a performance prediction result.

[0071] The microscopic information is used to indicate the microscopic properties of the candidate simulated polymer gas separation membrane. Schematically, the microscopic information includes: a SMILES string of at least one polymer monomer forming the candidate simulated polymer gas separation membrane, and / or, atom and bond information. The atom and bond information is generated based on the SMILES string, that is, when the microscopic information includes the atom and bond information, the electronic device obtains the SMILES string of at least one polymer monomer of the candidate simulated polymer gas separation membrane, and generates the atom and bond information based on the SMILES string.

[0072] Illustratively, generating atom and bond information based on the SMILES string includes: parsing the SMILES string to obtain atom information and bond information, wherein the atom information is used to describe the atom type, and the bond information is used to indicate the connection relationship between atoms. Optionally, the atom information includes but is not limited to: the atomic symbol; the bond information includes but is not limited to: the bond type (such as a single bond, a double bond, a triple bond, etc.).

[0073] Mesoscopic information is used to indicate the mesoscopic properties of the candidate simulated polymer gas separation membrane. Schematically, the mesoscopic information includes: the inter-monomer linking mode of the polymer monomer. The inter-monomer linking mode refers to: the specific way in which the polymer monomers are connected to each other during the polymerization process. The inter-monomer linking mode describes which parts of the polymer monomers (such as functional groups) are involved in the connection and how they are arranged with each other. The inter-monomer linking mode can affect the geometric shape, spatial arrangement and final physical properties of the polymer. For example: the inter-monomer linking mode includes but is not limited to: head-to-tail linking (Head-to-Head Linkage), head-to-head linking (Head-to-Head Linkage), tail-to-tail linking (Tail-to-Tail Linkage), alternating linking (Alternate Linkage), branching linking (Branching Linkage), cross-linking (Cross-linking), cyclic linking (Cyclic Linkage), etc. This embodiment does not list the inter-monomer linking modes one by one. Optionally, the inter-monomer linking mode can be randomly generated, or it can also be received based on the human-computer interaction interface. This embodiment does not limit the acquisition method of the inter-monomer linking mode.

[0074] Optionally, the multi-scale information also includes: macroscopic information, which is used to indicate the macroscopic properties of the candidate simulated polymer gas separation membrane. Schematically, the macroscopic information includes: free volume fraction (Fraction of free volume, FFV) and / or glass transition temperature Tg. Among them, FFV refers to the proportion of space not occupied by molecules in the candidate simulated polymer gas separation membrane; the glass transition temperature Tg is the temperature at which the candidate simulated polymer gas separation membrane changes from a glass state to a rubber state.

[0075] In one example, FFV and Tg are obtained by performing molecular dynamics simulation on a candidate simulated polymer gas separation membrane using a molecular dynamics simulation tool.

[0076] The molecular dynamics simulation tool is used to perform molecular dynamics simulation on the candidate simulated polymer gas separation membrane, including: the molecular dynamics simulation of the candidate simulated polymer gas separation membrane is performed by using the optimized potentials for liquid simulations-all atom (OPLS-AA) force field for all-atom molecular dynamics (MD) simulation.

[0077] Since OPLS-AA is more effective and robust in simulating various molecular systems, its use in the polymer-gas simulation system in this embodiment can improve the simulation accuracy.

[0078] Specifically, taking the molecular dynamics simulation tool as the LAMMPS software package as an example, the simulation process includes the following steps 31 and 32:

[0079] Step 31, generating a periodic molecular box suitable for LAMMPS simulation;

[0080] Before starting the simulation through LAMMPS, define a three-dimensional space region, which is the simulation box (or simulation unit). Place the simulated molecules corresponding to the candidate simulated polymer gas separation membrane currently being simulated in this region for subsequent simulation calculations to obtain a molecular box.

[0081] When creating a simulation box, it is necessary to specify the boundary conditions of the simulation unit. Common boundary conditions include periodic boundary conditions and non-periodic boundary conditions. Periodic boundary conditions indicate that when a particle leaves one boundary of the simulation box, it will re-enter the simulation unit from the opposite boundary, so that an infinite system can be simulated. Non-periodic boundary conditions do not have such restrictions. Non-periodic boundary conditions indicate that the boundary of the simulation box will not be regarded as part of the continuous space, that is, the particles in the simulation box will not jump to the other side of the box when reaching the boundary, but will be affected by the boundary, such as being reflected or absorbed. Schematically, in this embodiment, the boundary condition of the periodic molecular box is an infinite period molecular box, that is, the boundary of the simulation box is regarded as extending infinitely, and the particles will immediately appear at the opposite boundary when they reach one side boundary, thereby simulating an infinite system.

[0082] Optionally, before step 31, the simulation system may be subjected to energy minimization and annealing processes to eliminate any unnatural atomic overlaps or high energy states through the energy minimization process, and the system temperature may be slowly reduced through the annealing process to ensure that the simulation system reaches a thermodynamic equilibrium state.

[0083] Step 32, execute the run file for molecular dynamics simulation, simulate the temperature change process through the heating sequence and cooling sequence in the run file; process the density of the candidate simulated polymer gas separation membrane recorded during the temperature change process through a customized analysis script, and determine the glass transition temperature Tg of the candidate simulated polymer gas separation membrane based on the density; and / or, process the ratio of the volume occupied by atoms and the volume of pores in the candidate simulated polymer gas separation membrane during the temperature change process through a customized analysis script to obtain FFV.

[0084] In this embodiment, the operation file includes a heating sequence and a cooling sequence for simulating a temperature change process during the molecular dynamics simulation.

[0085] Schematically, the temperature change process simulated by the heating sequence and the cooling sequence includes: starting from a preset initial temperature, gradually increasing to a preset maximum temperature according to a preset temperature interval. The maximum temperature is greater than the initial temperature, the initial temperature may be 200 Kelvin (K), and the maximum temperature may be 1000K; in other embodiments, the values ​​of the initial temperature and the maximum temperature may also be other values, and this embodiment does not limit the values ​​of the initial temperature and the maximum temperature.

[0086] During the glass transition, the volume of the candidate simulated polymer gas separation membrane will change significantly, and then, the density of the candidate simulated polymer gas separation membrane will change significantly. Based on this, after the temperature ramp process is completed, the recorded density is processed by the analysis script, and the script calculates the discrete derivative (or slope) of the density as a function of temperature, and determines the temperature with the maximum slope (or the maximum rate of density change) as the glass transition temperature Tg. In addition, the volume occupied by atoms and the volume of pores recorded by the analysis script are processed, and then the ratio of the volume occupied by atoms to the volume of pores is calculated to obtain FFV.

[0087] Optionally, the running file also includes other files required for the molecular dynamics simulation process, such as: topology files, coordinate files and parameter files of each polymer, etc., to indicate how to build a force field model of the molecules, determine the position and direction of the molecules, define the topological information and parameters of a single molecule type, etc. This embodiment does not limit the content included in the running file.

[0088] In this embodiment, reference Figure 2 , the pre-trained membrane performance prediction model includes, in order according to the information transmission direction: an input layer 210, a convolutional layer 220, an attention pooling layer 230 and an output layer 240.

[0089] The input layer 210 is used to input multi-scale information. In this embodiment, the electronic device labels the attributes of different scale information in the multi-scale information to record the source and level of each piece of information. The attributes of different scale information remain unchanged during the prediction process.

[0090] Taking multi-scale information including microscopic information, mesoscopic information and macroscopic information as an example, refer to Figure 3 The schematic diagram of the 3-layer input configuration of the input layer 210 shown in FIG. Figure 3 It can be seen that the information input by the input layer 210 includes three levels, and the scale information of the third level is microscopic information. Accordingly, the attributes of the third level indicate that the microscopic information belongs to the third level. The microscopic information includes a SMILES string of at least one polymer monomer forming a candidate simulated polymer gas separation membrane; then, the input layer 210 generates the atomic and bond information of each polymer monomer based on the SMILES string. Figure 4, the atom and bond information includes node features, node labels, edge features and edge weights. Among them, node features include: atomic charge number, atomic type (such as hydrogen mixture, aromaticity), etc.; node labels include: atomic number. Edge features include: bond type and conjugated system form; edge weight: steric effect.

[0091] When constructing atom and bond information, if the candidate simulated polymer gas separation membrane is generated by at least two different polymer monomer units, an additional edge can be added to indicate the connection mode of different types of polymer monomer units to indicate that each polymer monomer unit may also be connected to other monomers. In this way, when the membrane performance prediction model analyzes one of the polymer monomer units, it can take into account the existence of other monomers based on the additional edge, so that the end atoms can understand the local chemical environment. Among them, the polymer monomer unit refers to a unit composed of multiple repeated polymer monomers. The performance of this representation is better than using virtual atoms to represent the monomer ends. For example, the polymer chain corresponding to the candidate simulated polymer gas separation membrane is: AAABBABCCCBBAAA, where A, B and C represent different polymer monomer units. At this time, an edge can be connected between each of the polymer monomer units A, B and C to define the corresponding connection method, for example, an edge is connected between the polymer monomer units A and B to define that the polymer monomer units A and B are connected as a CC bond; an edge is connected between the polymer monomer units A and B to define that the polymer monomer units A and B are connected as a CN bond; an edge is connected between the polymer monomer units B and C to define that the polymer monomer units B and C are connected as a CN bond.

[0092] The input layer 210 of this embodiment supports the input of multiple SMILES strings, such as: "Monomer SMILES1. Monomer SMILES2. Monomer SMILES3. Monomer SMILES4", while the traditional neural network model only allows the input of a single SMILES string. At this time, it can only predict the performance of the candidate simulated polymer gas separation membrane formed by the polymerization of a single polymer monomer. This embodiment supports the prediction of the performance of the candidate simulated polymer gas separation membrane formed by the polymerization of multiple polymer monomers, expanding the application scenarios of the model.

[0093] The second-level attribute indicates that mesoscopic information belongs to the second level, and the mesoscopic information includes the inter-monomer linking mode of the polymer monomers corresponding to the atoms and bonds indicated by the microscopic information.

[0094] The first level of attribute indication macroscopic information belongs to the first level, and the macroscopic information includes Tg and FFV of the candidate simulated polymer gas separation membrane composed of the polymer monomer corresponding to the atoms and bonds indicated by the microscopic information.

[0095] Based on this, the input layer 210 receives the input information and associates the multi-scale information for performance prediction by the hierarchical graph convolutional neural network. The performance of the membrane material can be predicted by combining the different scale characteristics of the candidate simulated polymer gas separation membrane to improve the prediction accuracy.

[0096] Optionally, in order to process attributes of different scales, the input layer 210 includes a multilayer perceptron (MLP) to preprocess multi-scale information so that each attribute is subjected to a nonlinear transformation before being input into the convolutional layer to obtain a processed feature representation, and the processed feature representation is uniformly input into the same convolutional layer to ensure the integrity and coherence of the information.

[0097] Optionally, the input layer 210 is also connected to an embedding layer to map the input information received by the input layer 210 to a continuous vector space (or embedding space) to improve the model's ability to understand the input information, thereby improving the model performance.

[0098] Convolutional layer 220 is used to extract features based on multi-scale information. Figure 3 In this embodiment, the convolutional layer is a graph convolutional network (GCN) layer. The GCN layer updates the feature representation of the node by performing specific operations on the node features and the adjacency matrix.

[0099] The attention pooling layer 230 is used to determine the attention score of the feature data output by the convolution layer; based on the attention score, the feature data is weighted pooled to obtain processed feature data.

[0100] Optionally, the attention pooling layer is specifically used to: determine the attention score of the feature data based on a multi-head attention mechanism guided by a mask matrix. That is, for each head in the multi-head attention mechanism, after determining the attention score corresponding to the head, the mask matrix is ​​used to set the attention score between the preset feature position and other feature positions to a preset smaller value (such as negative infinity) to obtain an updated attention score; and the updated attention score is used to perform weighted pooling on the feature data to obtain processed feature data.

[0101] The output layer 240 is used to generate performance prediction results based on the processed feature data.

[0102] Optionally, the performance prediction result is determined based on the performance prediction requirement, for example, if the performance prediction requirement is to predict the Tg of the candidate simulated polymer gas separation membrane, then the membrane performance prediction model is at least trained with Tg as a label, and the performance prediction result includes a parameter indicating Tg; another example: if the performance prediction requirement is to predict the solubility of the candidate simulated polymer gas separation membrane, then the membrane performance prediction model is at least trained with solubility as a label, and the performance prediction result includes a parameter indicating solubility. In addition, the performance prediction result can be one or more. In the case of multiple performance prediction results, the membrane performance prediction model is a multi-task learning model.

[0103] The training process of the membrane performance prediction model is detailed in the following embodiment, which will not be described in detail in this embodiment.

[0104] Step 104 : Screen out at least one target simulated polymer gas separation membrane that meets the desired performance from the membrane search space based on the performance prediction result.

[0105] The expected performance is preset in the electronic device, and the electronic device compares the performance prediction result with the expected performance. If the performance prediction result matches the expected performance, the candidate simulated polymer gas separation membrane corresponding to the performance prediction result is determined as the target simulated polymer gas separation membrane.

[0106] In this embodiment, a large number of candidate simulated polymer gas separation membranes in the membrane search space are screened through the membrane performance prediction model, and the screened target simulated polymer gas separation membranes are designed. This can filter out some candidate simulated polymer gas separation membranes that cannot be designed, thereby improving the design efficiency of polymer gas separation membranes.

[0107] Step 105, performing Monte Carlo simulation on each target simulated polymer gas separation membrane based on the molecular dynamics simulation tool to determine the solubility coefficient of the target simulated polymer gas separation membrane for different gases; and performing molecular dynamics simulation on each target simulated polymer gas separation membrane based on the molecular dynamics simulation tool to determine the diffusion coefficient of the target simulated polymer gas separation membrane for different gases.

[0108] The relevant description of the molecular dynamics simulation tool is detailed in the above embodiment, and will not be repeated in this embodiment.

[0109] In one example, the solubility coefficient of the target simulated polymer gas separation membrane is first determined, and then the diffusion coefficient of the target simulated polymer gas separation membrane is determined.

[0110] A Monte Carlo simulation is performed on each target simulated polymer gas separation membrane to obtain the solubility coefficient of the target simulated polymer gas separation membrane for different gases, including the following steps 51-58:

[0111] Step 51, performing energy minimization and annealing process on the simulation system to obtain a stabilized system;

[0112] The relevant description of the energy minimization and annealing process is detailed in the above embodiment, and will not be repeated in this embodiment.

[0113] Step 52, performing a constant number of atoms, volume and temperature (NVT) simulation on the stabilized system;

[0114] Among them, NVT simulation refers to molecular dynamics simulation performed under constant conditions of atomic number (N), volume (V) and temperature (T).

[0115] For example, the electronic device performs a 0.5 nanosecond NVT simulation at 500K on the stabilized system. In other embodiments, the simulation duration and temperature may also be other values, and this embodiment does not limit the values ​​of the simulation duration and temperature.

[0116] Step 53, performing a constant number of atoms, pressure and temperature (NPT) simulation on the system after the NVT simulation;

[0117] Among them, NPT simulation refers to the molecular dynamics simulation performed under constant conditions of atomic number (N), pressure (P) and temperature (T).

[0118] For example, the electronic device performs NPT simulation of 0.5 nanoseconds, 500K and 1 bar on the system after NVT simulation. In other embodiments, the simulation time and temperature may also be other values, and this embodiment does not limit the values ​​of simulation time, temperature and pressure.

[0119] Step 54, performing x thermal annealing cycles on the system after NPT simulation, so that the temperature of the simulation system reaches a first preset temperature, maintaining the pressure at a preset pressure value during the annealing process, and performing NVT simulation and NPT simulation on the simulation system after annealing again;

[0120] The first preset temperature may be 250K. In other embodiments, the first preset temperature may also be other values. This embodiment does not limit the value of the first preset temperature. x is the number of cycles for reducing the temperature of the simulation system to 250K. For example, if the initial temperature of the simulation system is 500K and each thermal annealing cycle reduces the temperature by 50K, the value of x is 5. The preset pressure value may be 1 bar.

[0121] The duration of NVT simulation and NPT simulation may be the same as that in the above steps, for example, 0.5 nanoseconds. In other embodiments, the duration may be other values. This embodiment does not limit the duration of NVT simulation and NPT simulation.

[0122] Step 55, determining the solubility coefficients of the target simulated polymer gas separation membrane for different gases.

[0123] Under infinite dilution, the solubility coefficient can be expressed by the Henry constant of the target simulated polymer gas separation membrane. The Henry constant of the polymer gas separation membrane refers to the proportional constant between the solubility of the gas in the polymer membrane material and its partial pressure on the membrane surface under specific temperature and pressure conditions. The solubility coefficient is the limiting value of the ratio of the concentration of the gas in the target simulated polymer gas separation membrane to its partial pressure on the membrane surface. The solubility coefficient and Henry constant are numerically inversely related. The larger the Henry constant, the smaller the solubility coefficient, indicating that the solubility of the gas in the target simulated polymer gas separation membrane is lower.

[0124] Schematically, the solubility coefficient S corresponding to the i-th gas i It is expressed by the following formula:

[0125]

[0126] Among them, c i represents the concentration of the i-th gas in the target simulated polymer gas separation membrane, f i represents the partial pressure of the i-th gas in the membrane surface, the limit c i →0 represents the situation when the gas concentration approaches 0, that is, under the condition of infinite dilution. i is a positive integer.

[0127] Step 56, after completing the solubility calculation, introducing gas molecules into the molecular box containing the target simulated polymer gas separation membrane; and performing energy minimization and annealing processes on the simulated system filled with gas to make the molecules in the simulated system reach a stable or equilibrium state;

[0128] The gas molecules may be CO2 and / or N2, etc., or may be gases in the actual working environment of the target simulated polymer gas separation membrane. This embodiment does not limit the type of gas molecules.

[0129] Step 57, performing NVT simulation and NPT simulation at a second preset temperature.

[0130] The second preset temperature may be 300K. In other embodiments, the second preset temperature may also be other values. This embodiment does not limit the value of the second preset temperature.

[0131] For example, the electronic device first performs NVT simulation and maintains it for 1 nanosecond; then, performs NPT simulation at 300K and 1 atmosphere and maintains it for 2 nanoseconds. In other embodiments, the duration of the NVT simulation and NPT simulation in this step can also be other durations, as long as the duration allows the system to reach stability.

[0132] Step 58, running the simulation system for a preset maintenance time, and collecting and analyzing data within the preset maintenance time to obtain a mean square displacement (MSD), and determining the diffusion coefficient of the target simulated polymer gas separation membrane for different gases based on the mean square displacement.

[0133] The preset maintenance time is used to make the simulation system reach equilibrium and complete data collection and analysis. For example, if the preset maintenance time is 50 nanoseconds, the first 2 nanoseconds are used for equilibrium, and the remaining 48 nanoseconds are used for data collection and analysis. In actual implementation, the preset maintenance time may also be other values, and this embodiment does not limit the value of the preset maintenance time.

[0134] The diffusion coefficient is used to describe the rate of random movement of gas in the target simulated polymer gas separation membrane.

[0135] The diffusion coefficient D corresponding to the i-th gas i It can be expressed by the following formula:

[0136]

[0137] Where t represents time, r(t) represents the position vector of the particle of the i-th gas at time t; r(0) represents the position vector of the particle of the i-th gas at the initial time (t=0); |r(t)-r(0)| 2 represents the square of the distance between the particle's position at time t and the initial moment, and <> represents the statistical average, that is, the average of the results of all particles or multiple experiments.

[0138] According to the above formula, the diffusion coefficient is one sixth of the average square displacement of the particle over an infinite time. In practical applications, since it is impossible to measure an infinite time, the MSD is measured on a sufficiently long time scale and the diffusion coefficient is estimated by linear fitting. Ideally, the MSD is linearly related to time, that is, <|r(t)-r(0)| 2 >=6D i t, that is, in three-dimensional space, the MSD of the particle increases linearly with time, and the slope of MSD is the diffusion coefficient D i .

[0139] Step 106 : determining the performance index of the target simulated polymer gas separation membrane based on the solubility coefficient and the diffusion coefficient. The performance index includes the gas permeability coefficient and / or the selectivity of the membrane.

[0140] The permeability coefficient is used to describe the rate at which a gas passes through a target simulated polymer gas separation membrane. The permeability coefficient combines the solubility and diffusion capacity of the gas in the target simulated polymer gas separation membrane. For the i-th gas, the permeability coefficient Pi It can be expressed by the following formula:

[0141] P i =S i ·D i ;

[0142] Among them, S i represents the solubility coefficient corresponding to the i-th gas; D i represents the diffusion coefficient corresponding to the i-th gas.

[0143] Selectivity is used to describe the separation ability of the target simulated polymer gas separation membrane for different gases. Generally, selectivity is expressed by the ratio of the permeability coefficients of two gases.

[0144] Schematically, the selectivity α for gas i relative to gas j is i / j , the selectivity can be expressed by the following formula:

[0145] α i / j =P i / P j ;

[0146] Among them, P i represents the permeability coefficient corresponding to the i-th gas, P j represents the permeability coefficient corresponding to the j-th gas, where j is a positive integer different from the value of i.

[0147] In summary, the design and screening method of the polymer gas separation membrane provided in the present embodiment obtains monomer information of multiple polymer monomers; generates a membrane search space based on the monomer information, wherein the membrane search space includes multiple candidate simulated polymer gas separation membranes; performs performance prediction on multiple candidate simulated polymer gas separation membranes in the search space based on a pre-trained membrane performance prediction model to obtain a performance prediction result for each candidate simulated polymer gas separation membrane; screens out at least one target simulated polymer gas separation membrane that meets the expected performance from the membrane search space based on the performance prediction result; performs Monte Carlo simulation on each target simulated polymer gas separation membrane based on a molecular dynamics simulation tool to determine the target Simulate the solubility coefficient of polymer gas separation membrane for different gases; and perform molecular dynamics simulation on each target simulated polymer gas separation membrane based on molecular dynamics simulation tools to determine the diffusion coefficient of the target simulated polymer gas separation membrane for different gases; determine the performance indicators of the target simulated polymer gas separation membrane based on the solubility coefficient and diffusion coefficient, and the performance indicators include the gas permeability coefficient and / or membrane selectivity; a search space can be generated based on monomer information, thereby forming a large number of non-existing simulated membrane materials for analysis, solving the problem that data is limited by the diversity of experimental conditions and methods, making it difficult to conduct large-scale material screening and optimization design, and providing the possibility for designing new membrane materials. At the same time, using the membrane performance prediction model to search for feasible target simulated polymer gas separation membranes in the search space, there is no need to manually screen the performance of a large number of simulated membrane materials in the search space, which can improve the efficiency of screening a large number of polymer materials.

[0148] In addition, a membrane performance prediction model is established based on a hierarchical graph convolutional neural network model to input the multi-scale information of the simulated membrane material into the membrane performance prediction model for performance prediction, so that the membrane performance prediction model can perform performance prediction based on the comprehensive characteristics of the simulated membrane material, thereby improving the accuracy of performance prediction.

[0149] In addition, by setting an attention pooling layer in the membrane performance prediction model, the membrane performance prediction model can be dynamically weighted according to the importance of the features, effectively retaining more key information and filtering out redundant information, thereby improving the accuracy of performance prediction and improving computational efficiency.

[0150] In addition, by setting the attention pooling layer based on the mask matrix-guided multi-head attention mechanism to determine the attention score of the feature data, the impact and importance of attribute changes at different levels on performance prediction can be analyzed, allowing the model to dynamically adjust the importance of features, thereby more accurately capturing key features and improving model accuracy.

[0151] In addition, by manually constructing a polymer single chain model or a polymer monomer described by a SMILES string, non-professionals can also use the method provided in this embodiment to input monomer information, thereby improving the applicability of the method.

[0152] Based on the above embodiment, before step 103, that is, based on the pre-trained membrane performance prediction model, the performance of multiple candidate simulated polymer gas separation membranes in the search space is predicted, and before the performance prediction result of each candidate simulated polymer gas separation membrane is obtained, it also includes training to obtain the membrane performance prediction model.

[0153] The training process of the membrane performance prediction model includes the following steps:

[0154] Step 1: Generate a sample to simulate a polymer gas separation membrane.

[0155] Among them, the generation method of the sample simulated polymer gas separation membrane is the same as the generation method of the candidate simulated polymer gas separation membrane, that is, it is generated based on the acquired monomer information; or, the sample simulated polymer gas separation membrane is a simulation model corresponding to the existing polymer gas separation membrane. This embodiment does not limit the acquisition method of the sample simulated polymer gas separation membrane.

[0156] Step 2: Obtain sample microscopic information and sample mesoscopic information of the sample simulating polymer gas separation membrane.

[0157] The method for acquiring the sample microscopic information and the sample mesoscopic information is the same as the method for acquiring the microscopic information and the mesoscopic information of the candidate simulated polymer gas separation membrane, and the information types are consistent, so this embodiment will not be repeated here.

[0158] Step 3: Based on the molecular dynamics simulation tool, the molecular dynamics properties of each sample simulated polymer gas separation membrane are simulated to obtain the simulated performance parameters of the candidate simulated polymer gas separation membrane.

[0159] In this embodiment, the simulated performance parameters can be simulated by a molecular dynamics simulation tool to simulate the molecular dynamics properties of the sample simulated polymer gas separation membrane, and the simulated performance parameters are obtained by collecting data analysis during the simulation process. The simulated performance parameters are the same as the prediction target of the membrane performance prediction model. For example, the membrane performance prediction model is used to predict the Tg, solubility coefficient and diffusion coefficient of the candidate simulated polymer gas separation membrane, and the simulated performance parameters include the Tg, solubility coefficient and diffusion coefficient of the sample simulated polymer gas separation membrane.

[0160] Step 4: Train the pre-created hierarchical graph convolutional neural network model based on the sample microscopic information, sample mesoscopic information and simulation performance parameters to obtain a membrane performance prediction model.

[0161] Specifically, after the sample microscopic information and sample mesoscopic information are input into a pre-created hierarchical graph convolutional neural network model, the model's prediction results are obtained, and the prediction results are compared with the simulation performance parameters; based on the comparison results, the model parameters of the hierarchical graph convolutional neural network model are iteratively optimized to obtain a membrane performance prediction model.

[0162] Figure 5 A block diagram of a design and screening device for a polymer gas separation membrane according to an embodiment of the present disclosure is shown. Figure 5 As shown, the device includes the following modules: an information acquisition module 510, a space generation module 520, a performance prediction module 530, a membrane screening module 540, a motion simulation module 550 and a performance determination module 560.

[0163] An information acquisition module 510 is used to acquire monomer information of a plurality of polymer monomers;

[0164] A space generation module 520, for generating a membrane search space based on the monomer information, wherein the membrane search space includes a plurality of candidate simulated polymer gas separation membranes;

[0165] A performance prediction module 530 is used to perform performance prediction on multiple candidate simulated polymer gas separation membranes in the search space based on a pre-trained membrane performance prediction model to obtain a performance prediction result for each candidate simulated polymer gas separation membrane; the membrane performance prediction model is trained based on a machine learning algorithm;

[0166] A membrane screening module 540, for screening out at least one target simulated polymer gas separation membrane that meets the desired performance from the membrane search space based on the performance prediction result;

[0167] The motion simulation module 550 is used to perform Monte Carlo simulation on each target simulated polymer gas separation membrane based on a molecular dynamics simulation tool to determine the solubility coefficient of the target simulated polymer gas separation membrane for different gases; and perform molecular dynamics simulation on each target simulated polymer gas separation membrane based on the molecular dynamics simulation tool to determine the diffusion coefficient of the target simulated polymer gas separation membrane for different gases;

[0168] The performance determination module 560 is used to determine the performance index of the target simulated polymer gas separation membrane based on the solubility coefficient and the diffusion coefficient, and the performance index includes the gas permeability coefficient and / or the selectivity of the membrane.

[0169] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0170] The embodiment of the present disclosure also provides a computer-readable storage medium on which computer program instructions are stored, and the computer program instructions implement the above method when executed by a processor. The computer-readable storage medium can be a volatile or non-volatile computer-readable storage medium.

[0171] An embodiment of the present disclosure further proposes an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0172] The embodiments of the present disclosure also provide a computer program product, including a computer-readable code, or a non-volatile computer-readable storage medium carrying the computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above method.

[0173] Figure 6 1 is a block diagram of a device 1900 for designing and screening a polymer gas separation membrane according to an exemplary embodiment. For example, the device 1900 can be provided as a server or a terminal device. Figure 6 , the apparatus 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by a memory 1932 for storing instructions, such as an application, that can be executed by the processing component 1922. The application stored in the memory 1932 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 1922 is configured to execute instructions to perform the above method.

[0174] The device 1900 may also include a power supply component 1926 configured to perform power management of the device 1900, a wired or wireless network interface 1950 configured to connect the device 1900 to a network, and an input / output interface 1958 (I / O interface). The device 1900 may operate based on an operating system stored in the memory 1932, such as Windows Server 2000. TM , MacOS X TM , Unix TM ,Linux TM , FreeBSD TM or similar.

[0175] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions, which can be executed by the processing component 1922 of the device 1900 to perform the above method.

[0176] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method for designing and screening a polymer gas separation membrane, characterized in that: The method comprises: Obtain monomer information of multiple polymer monomers; generating a membrane search space based on the monomer information, wherein the membrane search space includes a plurality of candidate simulated polymer gas separation membranes; Based on a pre-trained membrane performance prediction model, a plurality of candidate simulated polymer gas separation membranes in the search space are predicted for performance, and a performance prediction result of each candidate simulated polymer gas separation membrane is obtained; the membrane performance prediction model is obtained by training based on a machine learning algorithm; Screening out at least one target simulated polymer gas separation membrane that meets the desired performance from the membrane search space based on the performance prediction result; Performing Monte Carlo simulation on each target simulated polymer gas separation membrane based on a molecular dynamics simulation tool to determine the solubility coefficient of the target simulated polymer gas separation membrane for different gases; and performing molecular dynamics simulation on each target simulated polymer gas separation membrane based on the molecular dynamics simulation tool to determine the diffusion coefficient of the target simulated polymer gas separation membrane for different gases; Based on the solubility coefficient and the diffusion coefficient, the performance index of the target simulated polymer gas separation membrane is determined, and the performance index includes the gas permeability coefficient and / or the selectivity of the membrane.

2. The method according to claim 1, characterized in that The performance prediction of multiple candidate simulated polymer gas separation membranes in the search space is performed based on the pre-trained membrane performance prediction model to obtain a performance prediction result of each candidate simulated polymer gas separation membrane, including: For each candidate simulated polymer gas separation membrane, multi-scale information of the candidate simulated polymer gas separation membrane is obtained; wherein the multi-scale information includes at least microscopic information and mesoscopic information; wherein the microscopic information is used to indicate the microscopic properties of the candidate simulated polymer gas separation membrane, and the mesoscopic information is used to indicate the mesoscopic properties of the candidate simulated polymer gas separation membrane; The multi-scale information is input into the membrane performance prediction model to obtain the performance prediction result; wherein, the membrane performance prediction model is established based on a hierarchical graph convolutional neural network model.

3. The method according to claim 2, characterized in that The membrane performance prediction model includes the following in order according to the information transmission direction: An input layer, used for inputting the multi-scale information; A convolutional layer, used for extracting features based on the multi-scale information; An attention pooling layer, used to determine an attention score of the feature data output by the convolutional layer; and weighted pool the feature data based on the attention score to obtain processed feature data; The output layer is used to generate the performance prediction result based on the processed feature data.

4. The method according to claim 3, characterized in that The attention pooling layer is specifically used to determine the attention score of the feature data based on a multi-head attention mechanism guided by a mask matrix.

5. The method according to claim 2, characterized in that: The microscopic information includes: a SMILES string of at least one polymer monomer forming the candidate simulated polymer gas separation membrane, and / or, atom and bond information; wherein the atom and bond information is generated based on the SMILES string; The mesoscopic information includes: the inter-monomer linking mode of the polymer monomers.

6. The method according to claim 2, characterized in that The multi-scale information further includes: macroscopic information, wherein the macroscopic information is used to indicate the macroscopic properties of the candidate simulated polymer gas separation membrane.

7. The method according to claim 6, characterized in that The macroscopic information includes: free volume fraction FFV and / or glass transition temperature Tg, wherein FFV refers to the proportion of space not occupied by molecules in the candidate simulated polymer gas separation membrane; the glass transition temperature Tg is the temperature at which the candidate simulated polymer gas separation membrane changes from a glass state to a rubber state.

8. The method according to claim 2, characterized in that: The method further comprises: performing performance prediction on a plurality of candidate simulated polymer gas separation membranes in the search space based on the pre-trained membrane performance prediction model to obtain a performance prediction result for each candidate simulated polymer gas separation membrane; Generate samples to simulate polymer gas separation membranes; Obtaining sample microscopic information and sample mesoscopic information of the sample simulated polymer gas separation membrane; Simulating the molecular dynamics properties of each candidate simulated polymer gas separation membrane based on the molecular dynamics simulation tool to obtain simulated performance parameters of the sample simulated polymer gas separation membrane; The membrane performance prediction model is obtained by training a pre-created hierarchical graph convolutional neural network model based on the sample microscopic information, the sample mesoscopic information and the simulation performance parameters.

9. The method according to any one of claims 1 to 8, characterized in that: The obtaining of monomer information of a plurality of polymer monomers comprises: Receive the SMILES string of the polymer monomer; and / or, Receiving a polymer single-chain model; parsing the polymer single-chain model to obtain the monomer information.

10. A design and screening device for polymer gas separation membrane, characterized in that: include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement the method described in any one of claims 1 to 9 when executing the instructions stored in the memory.

11. A non-volatile computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 9 is implemented.

Citation Information

Patent Citations

  • Membrane separation performance optimization method and device and computing equipment

    CN118471392A

  • Separation membrane simulation method, simulation device, program and computer-readable storage medium with the program recorded thereon, and separation membrane

    JP2007319808A

  • Machine learning-assisted rational design of separation membranes

    WO2023200979A1

Cited By

  • Gas separation membrane model construction method, performance test simulation method and electronic equipment

    CN119943174A

  • Model training method and device, medium and electronic equipment

    CN120636581A

  • Polymeric gas separation membrane designing and screening method and device, and storage medium

    WO2026051428A1