Screening method and device of catalyst for carbon dioxide reduction reaction, electronic equipment and medium

Abstract

The invention relates to a screening method and device for catalysts for carbon dioxide reduction reactions, electronic equipment and a medium. The method comprises the steps that a screening task is obtained, and reaction process information of all carbon dioxide reduction reactions and geometric structure information of all catalysts to be selected are determined; inputting the reaction process information of each carbon dioxide reduction reaction and the geometric structure information of each to-be-selected catalyst into a prediction model for calculation to obtain the adsorption energy of each to-be-selected catalyst-target intermediate combination; and determining the activity and selectivity of the corresponding to-be-selected catalysts, screening out a target catalyst from the plurality of to-be-selected catalysts according to the activity and selectivity of each to-be-selected catalyst, and taking the target catalyst as a screening result of the screening task. The screening speed of the novel high-performance catalyst suitable for the carbon dioxide reduction reaction can be remarkably increased, and a high-selectivity and high-efficiency solution is provided for industrial application.

Description

Technical Field

[0001] The present disclosure relates to the technical field of electrochemical reduction catalyst screening, and in particular to a screening method and device, electronic equipment and medium for a catalyst used in a carbon dioxide reduction reaction. Background Art

[0002] Carbon dioxide (CO2) electrochemical reduction reaction (CO2RR) is considered to be one of the effective ways to combat climate change. Currently, the industrial application of CO2RR is limited by multiple factors, among which the development of catalysts is a key factor, which directly affects the current density, Faradaic efficiency, energy efficiency and stability, thus having an important impact on the operating cost.

[0003] However, developing single-atom catalysts with excellent selectivity and activity still faces huge challenges in traditional computational and experimental methods, such as low efficiency and high computational cost when screening a large number of candidate catalysts. Summary of the invention

[0004] In view of this, the present disclosure proposes a screening method and device, electronic equipment and medium for catalysts for carbon dioxide reduction reactions, which can significantly accelerate the screening speed of new high-performance catalysts suitable for carbon dioxide reduction reactions and provide highly selective and efficient solutions for industrial applications.

[0005] According to one aspect of the present disclosure, a method for screening catalysts for a carbon dioxide reduction reaction is provided, comprising: obtaining a screening task, the screening task being used to indicate a plurality of candidate catalysts and a target reduction reaction, the target reduction reaction including at least one carbon dioxide reduction reaction, and determining reaction process information of each of the carbon dioxide reduction reactions and geometric structure information of each of the candidate catalysts; inputting the reaction process information of each of the carbon dioxide reduction reactions and the geometric structure information of each of the candidate catalysts into a prediction model for calculation to obtain adsorption energies of each candidate catalyst-target intermediate combination, each of the candidate catalyst-target intermediate combinations being formed based on a candidate catalyst and a target intermediate, and all target intermediates being determined based on the target reduction reaction, wherein the prediction model is determined based on a graph neural network; determining the activity and selectivity of the corresponding candidate catalyst according to the adsorption energies of all candidate catalyst-target intermediate combinations corresponding to each of the candidate catalysts, and screening out a target catalyst from the plurality of candidate catalysts according to the activity and selectivity of each of the candidate catalysts, and using the target catalyst as the screening result of the screening task.

[0006] In this way, the screening method provided by the embodiment of the present disclosure can perform intelligent high-throughput screening based on the graph neural network model and active learning strategy, and is specifically used for the efficient screening and design of catalysts used in CO2 reduction reactions. This screening method not only overcomes the defects of high computational cost and low efficiency in traditional screening methods, but also realizes automatic iterative screening of high-potential catalysts by introducing a prediction model. Through this screening method, several highly selective and highly active catalysts can be screened out, providing the possibility of industrial application for the efficient synthesis of high value-added products such as methanol. In addition, this screening method is versatile, automated and efficient, and is not only applicable to the field of CO2 reduction, but can also be extended to the screening and optimization of other electrochemical catalysts, and has broad industrial prospects and technical value.

[0007] In a possible implementation, the method further includes: acquiring a training sample set; and iteratively training an initial prediction model using the training sample set to obtain a trained prediction model.

[0008] In one possible implementation, obtaining a training sample set includes: obtaining geometric structure information of multiple sample catalysts and reaction process information of sample reduction reactions, determining each sample catalyst-sample intermediate combination and the corresponding geometric structure information, each of the sample catalyst-sample intermediate combinations is formed based on a sample catalyst and a sample intermediate, and all sample intermediates are determined based on the sample reduction reaction, wherein the sample reduction reaction includes at least one carbon dioxide reduction reaction; calculating the adsorption energy of each of the sample catalyst-sample intermediate combinations based on density functional theory; determining multiple training samples, each of the training samples corresponds to a different sample catalyst-sample intermediate combination, and each of the training samples includes the geometric structure information and adsorption energy of the corresponding sample catalyst-sample intermediate combination; and forming the training sample set based on all the training samples.

[0009] In this way, for the sample reduction reaction, all possible sample catalyst-sample intermediate combinations and their geometric structure information and adsorption energy are determined. The model trained based on this information can more accurately predict the interaction between catalysts and intermediates, thereby improving the accuracy of the prediction. And because the model training covers a variety of different catalyst and intermediate combinations, this helps to improve the generalization ability of the prediction model, so that it can adapt to carbon dioxide reduction reactions under different conditions.

[0010] In one possible implementation, the initial prediction model is iteratively trained using the training sample set to obtain a trained prediction model, including: in each iterative training process, selecting a subset used for this iterative training from the training sample set; and using the subset to train the current prediction model to obtain a trained prediction model.

[0011] In this way, through iterative training, the prediction model can gradually learn a wider range of sample features, thereby improving the generalization ability on unseen data; only part of the training samples are used for training in each iteration, which helps to optimize the use of computing resources, especially when processing large-scale data sets. Compared with processing the entire data set at one time, it can reduce some of the computing burden, thereby accelerating the training process of the prediction model. In addition, selecting a new subset for training in each iteration can make the prediction model more flexible and able to adapt to changes in data distribution; by continuously learning from different subsets, the prediction model can resist the influence of outliers and noise, and enhance the robustness of the prediction model.

[0012] In a possible implementation, selecting a subset from the training sample set for this iterative training includes: when it is determined that the training is currently in the first iterative training, randomly selecting some training samples from the training sample set to obtain a subset of the training sample set.

[0013] In one possible implementation, a subset for this iterative training is selected from the training sample set, including: when it is determined that the current training is not the first iterative training, the prediction model obtained from the last iterative training is used to predict the adsorption energy of the sample catalyst-sample intermediate combination corresponding to all the training samples in the training sample set, and the uncertainty corresponding to each of the training samples is determined based on the prediction results, and the subset is formed based on all the training samples whose uncertainties meet the preset conditions.

[0014] In this way, model training based on training samples with higher uncertainty in prediction results can enable the prediction model to focus more on learning those data points that are difficult to predict, thereby potentially improving the overall prediction accuracy of the prediction model; each iteration selects a new subset based on the current state of the model, dynamically adapting to the learning progress of the prediction model, which helps the prediction model better capture the complexity of the data and reduce the risk of overfitting; with limited computing resources, this training method allows computing resources to be concentrated on those training samples that are most in need of learning, thereby achieving optimal allocation of resources; and by continuously adjusting the subsets used for training, the prediction model can be exposed to more diverse data, which helps to enhance the robustness of the prediction model when facing new data.

[0015] In one possible implementation, the activity and selectivity of the corresponding catalyst to be selected are determined based on the adsorption energies of all the catalyst-target intermediate combinations corresponding to each of the catalysts to be selected, and the target catalyst is screened out from the multiple catalysts to be selected based on the activity and selectivity of each of the catalysts to be selected, including: determining the containment potential of the corresponding catalyst to be selected based on the adsorption energies of all the catalyst-target intermediate combinations corresponding to each of the catalysts to be selected; determining the activity and selectivity of the corresponding catalyst to be selected based on the containment potential of each of the catalysts to be selected; and screening out the target catalyst from the multiple catalysts to be selected according to preset activity screening conditions and selectivity screening conditions, the target catalyst including at least one catalyst to be selected.

[0016] In this way, catalysts with excellent selectivity and activity can be screened out.

[0017] According to another aspect of the present disclosure, a screening device for a catalyst for a carbon dioxide reduction reaction is provided, comprising: an acquisition module, for acquiring a screening task, the screening task being used to indicate a plurality of candidate catalysts and a target reduction reaction, the target reduction reaction comprising at least one carbon dioxide reduction reaction, and determining reaction process information of each of the carbon dioxide reduction reactions and geometric structure information of each of the candidate catalysts; a calculation module, for inputting the reaction process information of each of the carbon dioxide reduction reactions and the geometric structure information of each of the candidate catalysts into a prediction model for calculation, and obtaining the adsorption energy of each candidate catalyst-target intermediate combination, each of the candidate catalyst-target intermediate combinations being formed based on a candidate catalyst and a target intermediate, and all target intermediates being determined based on the target reduction reaction, wherein the prediction model is determined based on a graph neural network; a screening module, for determining the activity and selectivity of the corresponding candidate catalyst according to the adsorption energies of all candidate catalyst-target intermediate combinations corresponding to each of the candidate catalysts, and screening out a target catalyst from the plurality of candidate catalysts according to the activity and selectivity of each of the candidate catalysts, and taking the target catalyst as the screening result of the screening task.

[0018] In this way, the screening device provided by the embodiment of the present disclosure can perform intelligent high-throughput screening based on the graph neural network model and active learning strategy, and is specifically used for the efficient screening and design of catalysts used in CO2 reduction reactions. This screening device not only overcomes the defects of high computational cost and low efficiency in traditional screening devices, but also realizes automatic iterative screening of high-potential catalysts by introducing a prediction model. Through this screening device, several highly selective and highly active catalysts can be screened out, providing the possibility of industrial application for the efficient synthesis of high value-added products such as methanol. In addition, this screening device is versatile, automated and efficient, and is not only suitable for the field of CO2 reduction, but can also be extended to the screening and optimization of other electrochemical catalysts, and has broad industrial prospects and technical value.

[0019] In a possible implementation, the device further includes a training module, which is used to: obtain a training sample set; and iteratively train an initial prediction model using the training sample set to obtain a trained prediction model.

[0020] In one possible implementation, obtaining a training sample set includes: obtaining geometric structure information of multiple sample catalysts and reaction process information of sample reduction reactions, determining each sample catalyst-sample intermediate combination and the corresponding geometric structure information, each of the sample catalyst-sample intermediate combinations is formed based on a sample catalyst and a sample intermediate, and all sample intermediates are determined based on the sample reduction reaction, wherein the sample reduction reaction includes at least one carbon dioxide reduction reaction; calculating the adsorption energy of each of the sample catalyst-sample intermediate combinations based on density functional theory; determining multiple training samples, each of the training samples corresponds to a different sample catalyst-sample intermediate combination, and each of the training samples includes the geometric structure information and adsorption energy of the corresponding sample catalyst-sample intermediate combination; and forming the training sample set based on all the training samples.

[0021] In this way, for the sample reduction reaction, all possible sample catalyst-sample intermediate combinations and their geometric structure information and adsorption energy are determined. The model trained based on this information can more accurately predict the interaction between catalysts and intermediates, thereby improving the accuracy of the prediction. And because the model training covers a variety of different catalyst and intermediate combinations, this helps to improve the generalization ability of the prediction model, so that it can adapt to carbon dioxide reduction reactions under different conditions.

[0022] In one possible implementation, the initial prediction model is iteratively trained using the training sample set to obtain a trained prediction model, including: in each iterative training process, selecting a subset used for this iterative training from the training sample set; and using the subset to train the current prediction model to obtain a trained prediction model.

[0023] In this way, through iterative training, the prediction model can gradually learn a wider range of sample features, thereby improving the generalization ability on unseen data; only part of the training samples are used for training in each iteration, which helps to optimize the use of computing resources, especially when processing large-scale data sets. Compared with processing the entire data set at one time, it can reduce some of the computing burden, thereby accelerating the training process of the prediction model. In addition, selecting a new subset for training in each iteration can make the prediction model more flexible and able to adapt to changes in data distribution; by continuously learning from different subsets, the prediction model can resist the influence of outliers and noise, and enhance the robustness of the prediction model.

[0024] In a possible implementation, selecting a subset from the training sample set for this iterative training includes: when it is determined that the training is currently in the first iterative training, randomly selecting some training samples from the training sample set to obtain a subset of the training sample set.

[0025] In one possible implementation, a subset for this iterative training is selected from the training sample set, including: when it is determined that the current training is not the first iterative training, the prediction model obtained from the last iterative training is used to predict the adsorption energy of the sample catalyst-sample intermediate combination corresponding to all the training samples in the training sample set, and the uncertainty corresponding to each of the training samples is determined based on the prediction results, and the subset is formed based on all the training samples whose uncertainties meet the preset conditions.

[0026] In this way, model training based on training samples with higher uncertainty in prediction results can enable the prediction model to focus more on learning those data points that are difficult to predict, thereby potentially improving the overall prediction accuracy of the prediction model; each iteration selects a new subset based on the current state of the model, dynamically adapting to the learning progress of the prediction model, which helps the prediction model better capture the complexity of the data and reduce the risk of overfitting; with limited computing resources, this training method allows computing resources to be concentrated on those training samples that are most in need of learning, thereby achieving optimal allocation of resources; and by continuously adjusting the subsets used for training, the prediction model can be exposed to more diverse data, which helps to enhance the robustness of the prediction model when facing new data.

[0027] In one possible implementation, the activity and selectivity of the corresponding catalyst to be selected are determined based on the adsorption energies of all the catalyst-target intermediate combinations corresponding to each of the catalysts to be selected, and the target catalyst is screened out from the multiple catalysts to be selected based on the activity and selectivity of each of the catalysts to be selected, including: determining the containment potential of the corresponding catalyst to be selected based on the adsorption energies of all the catalyst-target intermediate combinations corresponding to each of the catalysts to be selected; determining the activity and selectivity of the corresponding catalyst to be selected based on the containment potential of each of the catalysts to be selected; and screening out the target catalyst from the multiple catalysts to be selected according to preset activity screening conditions and selectivity screening conditions, the target catalyst including at least one catalyst to be selected.

[0028] In this way, catalysts with excellent selectivity and activity can be screened out.

[0029] According to another aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0030] 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.

[0031] 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.

[0032] The present disclosure provides a screening method and device, electronic device and medium for a catalyst for a carbon dioxide reduction reaction. The screening task is obtained, the screening task is used to indicate a plurality of candidate catalysts and a target reduction reaction, the target reduction reaction includes at least one carbon dioxide reduction reaction, and the reaction process information of each carbon dioxide reduction reaction and the geometric structure information of each candidate catalyst are determined; the reaction process information of each carbon dioxide reduction reaction and the geometric structure information of each candidate catalyst are input into a prediction model for calculation, and the adsorption energy of each candidate catalyst-target intermediate combination is obtained, each candidate catalyst-target intermediate combination is formed based on a candidate catalyst and a target intermediate, and all target intermediates are based on The target reduction reaction is determined by the prediction model based on the graph neural network; the activity and selectivity of the corresponding candidate catalyst are determined according to the adsorption energy of all candidate catalyst-target intermediate combinations corresponding to each candidate catalyst, and the target catalyst is screened out from multiple candidate catalysts according to the activity and selectivity of each candidate catalyst, and the target catalyst is used as the screening result of the screening task. This can overcome the defects of high computational cost and low efficiency in traditional screening methods, and realize automatic iterative screening of high-potential catalysts by introducing the prediction model, and obtain several highly selective and highly active catalysts. This screening method is universal, automated and efficient, and has broad industrial prospects and technical value.

[0033] 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

[0034] 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.

[0035] Figure 1 A flow chart showing a method for screening a catalyst for a carbon dioxide reduction reaction provided by an embodiment of the present disclosure.

[0036] Figure 2 A schematic diagram showing the EquiformerV2 model and prediction performance test results provided by an embodiment of the present disclosure.

[0037] Figure 3 A schematic diagram showing the relationship between the metal provided by the embodiments of the present disclosure, the vacancy type in the catalyst material, and the containment charge.

[0038] Figure 4 A schematic diagram showing the model iteration results obtained by using the screening method provided in an embodiment of the present disclosure.

[0039] Figure 5A schematic diagram showing a method for screening catalysts for carbon dioxide reduction reactions provided by an embodiment of the present disclosure.

[0040] Figure 6 A schematic diagram showing a simulated catalytic space provided by an embodiment of the present disclosure is shown.

[0041] Figure 7 A block diagram of a device for screening catalysts for carbon dioxide reduction reaction provided by an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0042] 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.

[0043] 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.

[0044] 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 main purpose of the present disclosure.

[0045] In order to facilitate those skilled in the art to understand the technical solution provided by the embodiments of the present disclosure, the technical environment in which the technical solution is implemented is described below.

[0046] As global warming continues to intensify and extreme climate events occur more frequently, the urgency of reducing greenhouse gas emissions has become increasingly apparent. The current global temperature has risen by 1.2°C compared to pre-industrial levels. Countries have reached a consensus in the Paris Agreement to take immediate emission reduction measures to prevent irreversible damage to the ecological and socio-economic systems. In this context, the electrochemical reduction reaction of carbon dioxide (CO2) (CO2RR) is considered to be one of the effective ways to address climate change. CO2RR can convert CO2 into high-value-added chemicals and renewable fuels, providing a feasible path for industrial decarbonization and achieving carbon neutrality in the future. Although the reduction of CO2 to CO has made some progress in industry, there is still great potential for further conversion to other high-value-added products.

[0047] At present, the industrial application of CO2RR is limited by technical and economic factors, among which catalyst design, system design and electrolyzer design all affect its economic feasibility. In particular, the development of catalysts is a key factor, which directly affects the current density, Faraday efficiency, energy efficiency and stability, thus having an important impact on operating costs. However, highly selective and highly active CO2RR catalysts are scarce at this stage, especially in reducing CO2 to produce high value-added chemicals such as methanol.

[0048] However, in traditional computational and experimental methods, the development of catalysts with excellent selectivity and activity still faces huge challenges. First, traditional methods are inefficient and computationally expensive when screening a large number of candidate catalysts. Secondly, these studies mostly use features based on specific structural designs, such as alloy nanoparticles, metal oxides, and metal dual-site catalysts. Although this feature design based on human expertise improves physical understanding, it limits the transferability of the model between different systems and relies on expert redesign, resulting in usage restrictions. Therefore, the development of a design method that can efficiently explore a wide range of catalyst design space and target single-atom catalysts has become an urgent need in this field.

[0049] In order to solve the above technical problems, the present disclosure provides a method for screening catalysts for carbon dioxide reduction reaction. Figures 1 to 6 The method for screening a catalyst for a carbon dioxide reduction reaction provided in an embodiment of the present disclosure is schematically illustrated.

[0050] like Figure 1 As shown, the screening method may include the following steps S101 to S103.

[0051] Step S101, obtaining a screening task, where the screening task is used to indicate a plurality of candidate catalysts and a target reduction reaction, where the target reduction reaction includes at least one carbon dioxide reduction reaction, and determining reaction process information of each carbon dioxide reduction reaction and geometric structure information of each candidate catalyst.

[0052] The target reduction reaction is a type of reduction reaction of interest, and the number of specific carbon dioxide reduction reactions included therein is set according to actual needs.

[0053] In some embodiments, the selected catalyst can be a single atom catalyst (Single Atom Catalysts, SACs). Correspondingly, the screened target catalyst is also a single atom catalyst. Among them, single atom catalysts have shown great potential in CO2RR catalysis due to their high atomic utilization, low-coordinated metal atoms, unique electronic structure and strong metal-support interaction. Since single atom catalysts have the advantages of low cost, adjustable structure, high energy efficiency and low overpotential, they are regarded as ideal candidates for the next generation of CO2RR catalysts. The selected catalysts and the target catalysts mentioned below are essentially catalysts.

[0054] Step S102, input the reaction process information of each carbon dioxide reduction reaction and the geometric structure information of each candidate catalyst into the prediction model for calculation to obtain the adsorption energy of each candidate catalyst-target intermediate combination, each candidate catalyst-target intermediate combination is formed based on a candidate catalyst and a target intermediate, and all target intermediates are determined based on the target reduction reaction, wherein the prediction model is determined based on a graph neural network.

[0055] The prediction model is a machine learning model based on a graph neural network (GNN) for predicting adsorption energy. Unlike traditional machine learning models that rely on manual feature design, the prediction model provided by the embodiment of the present disclosure can directly model the interaction relationship between atoms by converting the geometric structure of the catalyst into a graph structure, thereby improving the adaptability to different catalyst types. In addition, the prediction model provided by the embodiment of the present disclosure is obtained through pre-training on a large-scale adsorption energy dataset and fine-tuning on a specific catalyst dataset, which enables the prediction model to accurately predict the adsorption energy of the catalyst and greatly improves the screening efficiency. The training method of the prediction model is described later.

[0056] Step S103, determining the activity and selectivity of the corresponding candidate catalyst according to the adsorption energies of all candidate catalyst-target intermediate combinations corresponding to each candidate catalyst, and screening out the target catalyst from multiple candidate catalysts according to the activity and selectivity of each candidate catalyst, and taking the target catalyst as the screening result of the screening task.

[0057] In order to better carry out the target reduction reaction, this screening method focuses on the two related indicators of catalyst activity and selectivity, and then screens out the ideal target catalyst based on activity and selectivity. The selectivity of a catalyst refers to the comparative result of the degree to which a catalyst promotes different reactions in a reaction system where multiple reactions can occur. The activity of a catalyst, or catalytic activity, refers to the ability of the catalyst to catalyze.

[0058] In this way, the screening method provided by the embodiment of the present disclosure can perform intelligent high-throughput screening based on the graph neural network model and active learning strategy, and is specifically used for the efficient screening and design of catalysts used in CO2 reduction reactions. This screening method not only overcomes the defects of high computational cost and low efficiency in traditional screening methods, but also realizes automatic iterative screening of high-potential catalysts by introducing a prediction model. Through this screening method, several highly selective and highly active catalysts can be screened out, providing the possibility of industrial application for the efficient synthesis of high value-added products such as methanol. In addition, this screening method is versatile, automated and efficient, and is not only applicable to the field of CO2 reduction, but can also be extended to the screening and optimization of other electrochemical catalysts, and has broad industrial prospects and technical value.

[0059] The screening method provided by the disclosed embodiment adopts an automatic iterative active learning framework to achieve efficient exploration and data optimization in the catalyst screening process. The active learning framework selects the candidate catalysts with the highest uncertainty and the greatest screening potential for a new round of model training through iterative data sampling and model updating. Unlike traditional screening methods, this active learning framework can gradually improve the model accuracy in the learning cycle, while avoiding the calculation of a large number of invalid samples, thereby optimizing resource utilization.

[0060] The screening method may also include a prediction model training process, which may include: obtaining a training sample set; iteratively training the initial prediction model using the training sample set to obtain a trained prediction model.

[0061] During the training process of the prediction model, obtaining a training sample set may include: obtaining geometric structure information of multiple sample catalysts and reaction process information of sample reduction reactions, determining each sample catalyst-sample intermediate combination and the corresponding geometric structure information, each sample catalyst-sample intermediate combination is formed based on a sample catalyst and a sample intermediate, and all sample intermediates are determined based on the sample reduction reaction, wherein the sample reduction reaction includes at least one carbon dioxide reduction reaction; calculating the adsorption energy of each sample catalyst-sample intermediate combination based on density functional theory (DFT); determining multiple training samples, each training sample corresponds to a different sample catalyst-sample intermediate combination, and each training sample includes the geometric structure information and adsorption energy of the corresponding sample catalyst-sample intermediate combination; forming a training sample set based on all training samples. In this way, for the sample reduction reaction, all possible sample catalyst-sample intermediate combinations and their geometric structure information and adsorption energy are determined. The model trained based on this information can more accurately predict the interaction between catalysts and intermediates, thereby improving the accuracy of the prediction. And because the model training covers a variety of different catalyst and intermediate combinations, this helps to improve the generalization ability of the prediction model, so that it can adapt to carbon dioxide reduction reactions under different conditions.

[0062] In some embodiments, a large number of candidate catalysts can be screened using a DFT computing platform. The computing platform can be based on a simulation software package such as the VASP software package and use a computational hydrogen electrode (CHE) model to estimate the activity and selectivity of the candidate catalyst. In this way, the activity of the catalyst and its selectivity for the CO2 reduction reaction (CO2RR) can be quickly judged to ensure the accuracy of the preliminary screening, and the candidate catalysts with better activity and selectivity can be selected as the data basis for the subsequent construction of the training sample set, which provides high-quality data for the subsequent training of the prediction model. In the DFT calculation, the Projected Augmented Wave (PAW) method and commonly used functionals such as the PBE (Perdew-Burke-Ernzerhof) functional are used to calculate the electron exchange energy, and the DFT-D3 correction is added to describe the medium and long-range interactions. All calculations take spin polarization into account to ensure the accuracy of the electronic structure. At the same time, in order to avoid the influence of periodic boundary conditions, 15 angstroms can be used. The vacuum layer isolates the adsorption surface, making the simulated environment of each candidate catalyst closer to the actual experimental conditions to obtain more realistic adsorption energy.

[0063] Traditional catalyst screening often faces problems such as low efficiency and high consumption of computing resources, but this screening method significantly improves the screening efficiency through a high-throughput DFT computing platform. The DFT computing platform is used to automatically process, analyze and record the adsorption energy of each candidate catalyst-sample intermediate combination to ensure the accuracy and consistency of the screening, providing reliable basic data for further analysis.

[0064] In the process of training the prediction model, the initial prediction model is iteratively trained using the training sample set to obtain the trained prediction model, which may include: in each iterative training process, a subset used for this iterative training is selected from the training sample set; the current prediction model is trained using the subset to obtain the trained prediction model. In this way, through iterative training, the prediction model can gradually learn a wider range of sample features, thereby improving the generalization ability on unseen data; each iteration is trained only through part of the training samples, which helps to optimize the use of computing resources, especially when processing large-scale data sets, compared with processing the entire data set at one time, it can reduce some of the computing burden, thereby accelerating the training process of the prediction model. In addition, selecting a new subset for training each iteration can make the prediction model more flexible and able to adapt to changes in data distribution; by continuously learning from different subsets, the prediction model can resist the influence of outliers and noise, and enhance the robustness of the prediction model.

[0065] Selecting a subset from the training sample set for this iteration training may include: randomly selecting some training samples from the training sample set during the first iteration training, and using some training samples as subsets. Selecting a subset from the training sample set for this iteration training may also include: in a non-first iteration training process, using the prediction model obtained from the last iteration training to predict the adsorption energy of the sample catalyst-sample intermediate combination corresponding to all training samples in the training sample set, and determining the uncertainty corresponding to each training sample based on the prediction result, and forming a subset based on all training samples whose uncertainties meet the preset conditions. In this way, model training based on training samples with higher uncertainty in prediction results can enable the prediction model to focus more on learning those data points that are difficult to predict, thereby potentially improving the overall prediction accuracy of the prediction model; each iteration selects a new subset based on the current state of the model, dynamically adapting to the learning progress of the prediction model, which helps the prediction model better capture the complexity of the data and reduce the risk of overfitting; with limited computing resources, this training method allows computing resources to be concentrated on those training samples that are most in need of learning, thereby achieving optimal allocation of resources; and by continuously adjusting the subsets used for training, the prediction model can be exposed to more diverse data, which helps to enhance the robustness of the prediction model when facing new data.

[0066] The screening method provided by the embodiment of the present disclosure adopts an active learning strategy to achieve efficient screening and data optimization through automatic iterative sampling and model updating. For example, the active learning framework starts from a complete training sample set including 100 training samples, first randomly selects a number of training samples, such as 8 training samples, and forms a subset used in the first iterative training of the prediction model based on these 8 training samples; then, using a five-fold cross-validation method, the subset is used to train and evaluate the initial prediction model to obtain a prediction model after the first iterative training; then, the prediction model is used to predict the adsorption energy of the sample catalyst-sample intermediate combination corresponding to the 100 training samples in the complete training sample set, and the prediction results (i.e., adsorption energy) of these 100 training samples are determined. The uncertainty corresponding to the adsorption energy of 100 training samples, in this example, can be estimated by variance and the uncertainty corresponding to 10 training samples meets the preset conditions, such as being higher than the preset threshold, then a subset used for the second iteration training is formed based on these 10 training samples; in each subsequent round of iteration (second, third...), the active learning framework automatically selects several training samples with the highest uncertainty to train the prediction model, thereby improving the prediction accuracy of the prediction model, and compared to processing the entire data set at one time, the computational burden of each iteration can be reduced, thereby accelerating the training process of the prediction model. The iterative process of the active learning framework provided in the disclosed embodiment significantly improves the screening efficiency. At the same time, the active learning framework saves the calculated and screened data in each round of iteration, so as to facilitate cross-validation of future experiments and simulation results. The active learning process of the disclosed embodiment can adapt to different catalyst screening needs in continuous optimization, and provides new ideas for high-throughput screening tasks in other fields.

[0067] In the data sampling process, the screening method provided by the embodiment of the present disclosure adopts a multi-dimensional mixed sampling strategy. Each round of iteration not only selects the sample catalyst-sample intermediate combination with excellent performance for verification, but also randomly extracts from the globally unexplored samples to ensure the diversity of the data set and the comprehensiveness of the model. After each round of sampling, new data points are automatically added to the training set, and the prediction model is updated to gradually improve the accuracy of the prediction model in adsorption energy prediction.

[0068] In the training process of the prediction model, the disclosed embodiment can realize automated high-throughput DFT calculation and data processing, greatly improving the speed and efficiency of catalyst screening. By combining the self-compiled code package, it is possible to automatically collect, clean and store each round of calculation data, reducing the errors caused by human intervention. In addition, this screening method performs multiple rounds of screening on the catalyst materials to be selected, providing a solid data foundation for machine learning prediction.

[0069] In the specific implementation, the EquiformerV2 model (such as Figure 2 As shown in a, the model can use the embedding layer to perform node embedding and edge embedding on the input geometric structure information to obtain the corresponding node features and edge features, and then input the node features and edge features into the attention module including the attention unit and the feedforward unit for calculation, and obtain the adsorption energy of the corresponding geometric structure information) as a prediction model. The model contains 31 million parameters and is pre-trained on the OC20+22 dataset. The OC20+22 dataset includes adsorption energy prediction tasks for a variety of material systems, with strong adaptability and good mobility. During the fine-tuning process, all parameters are adjusted by reducing the learning rate, so that the prediction model can focus on a specific CO2RR reaction system. Through this fine-tuning method, the strong adaptability of the original model in generalization tasks is retained, and the target reduction reaction of actual concern is accurately optimized, thereby enhancing the accuracy of adsorption energy prediction. The performance test results of the prediction model can be shown by Figure 2 The prediction error distribution (Prediction Error Distribution) and mean absolute error (Mean Absolute Error, MAE) shown in b change with the number of iterations (Round) and Figure 2 The statistic R calculated based on the predicted adsorption energy (Prediction Energy) and the DFT calculated adsorption energy (DFT Energy) is shown in c. 2 , Root Mean Square Error (RMSE), and MAE show that the fitting results of the prediction model are good and the error is low.

[0070] This screening method trains and verifies the prediction model through five-fold cross-validation to ensure the robustness and accuracy of the prediction results. In each round of iteration, the average value of the prediction results of the prediction model is used to predict the adsorption energy, and the variance is used to estimate the uncertainty of the prediction. This prediction model can not only quickly and accurately evaluate the activity of the selected catalyst, but also has strong scalability and can be applied to other similar electrochemical catalytic reaction systems. Finally, the prediction model can obtain the optimal training samples, which provides a basis for the sampling of the subsequent active learning process.

[0071] In this way, this screening method realizes the rapid screening of a large number of single-atom catalysts by combining high-throughput DFT calculations with graph neural network models, overcoming the problems of high computing resource consumption and low screening efficiency in traditional screening methods. By combining deep learning with active learning, an automatic iterative screening framework is designed to optimize the prediction of catalyst activity and selectivity, and select the most promising catalyst-intermediate combination for calculation based on uncertainty, thereby significantly improving the speed and accuracy of screening. In addition, the graph neural network model used in this screening method does not need to rely on manual feature design, and has high adaptability and versatility, so that the screening process can be easily extended to different material systems. At the same time, the active learning framework continuously updates and optimizes the screening strategy in multiple rounds of learning iterations, effectively improving the screening accuracy and resource utilization efficiency of the selected materials.

[0072] like Figure 3 The relationship between different metals (Metal), vacancy types (Vacancy Type) in different catalyst materials, and limiting potential (Limiting Prediction) determined based on adsorption energy can be used to determine the activity and selectivity of the catalyst, thereby achieving catalyst screening. In some embodiments, step S103 determines the activity and selectivity of the corresponding catalyst according to the adsorption energies of all the catalyst-target intermediate combinations corresponding to each catalyst, and screens out the target catalyst from multiple catalysts according to the activity and selectivity of each catalyst. It may include: determining the containment potential of the corresponding catalyst according to the adsorption energies of all the catalyst-target intermediate combinations corresponding to each catalyst, for example, the containment potential of catalyst A can be obtained according to the adsorption energies of all combinations corresponding to catalyst A (i.e., catalyst A-target intermediate 1 combination, catalyst A-target intermediate 2 combination...catalyst A-target intermediate N combination), wherein the method for determining the containment potential can be flexibly selected according to actual needs, and the present disclosure does not limit this; determining the activity and selectivity of the corresponding catalyst according to the containment potential of each catalyst; and screening out the target catalyst from multiple catalysts according to preset activity screening conditions and selectivity screening conditions, wherein the target catalyst includes at least one catalyst. The present screening method can be used to obtain Figure 4 The iterative results are shown. Figure 4 The horizontal axis represents the containment potential of each catalyst, and the vertical axis represents the selectivity of various catalysts. Through several rounds of iterative screening results, a catalyst with significantly improved performance over the existing CuN4 catalyst can be obtained, such as the catalyst Pd-C3P based on palladium (Pd), carbon (C), and phosphorus (P).

[0073] In some embodiments, in addition to adsorption energy, each candidate catalyst may also be screened based on its binding energy to obtain a target catalyst with a relatively balanced stability, activity, and selectivity.

[0074] The intelligent screening method based on density functional theory calculation, graph neural network and active learning strategy provided by the disclosed embodiment can achieve high-efficiency and high-precision catalyst screening, and overcome the adaptability problems faced by traditional methods when designing general models. Active learning technology provides a new idea for solving the above problems. By adopting an automatic iterative machine learning framework, it can not only reduce the prediction uncertainty of the prediction model in the screening, but also significantly accelerate the screening speed of new high-performance CO2RR catalysts, providing highly selective and efficient solutions for industrial applications. This screening method can overcome the efficiency and applicability limitations of traditional screening methods, and combine advanced deep learning models and DFT calculations to achieve efficient screening and evaluation of catalysts to promote the generation of high value-added chemicals and meet the needs of industrial decarbonization applications in the field of electrochemistry.

[0075] This screening method innovatively uses a graph neural network model, which is pre-trained and then fine-tuned for a specific adsorption energy database. Compared with models that rely on traditional manually designed features, graph neural networks show extremely high adaptability and versatility in processing different types of catalyst materials, so that the screening process no longer relies on the redesign of structural features. In this way, the prediction model provided by the embodiment of the present disclosure can not only quickly predict the adsorption energy and then determine the corresponding activity and selectivity of the catalyst, but also supports the screening of multiple types of catalyst materials, greatly improving the migration ability and versatility of the model.

[0076] The high-throughput screening process of this screening method combines large-scale DFT calculations and deep learning predictions, has a high degree of automation, simplifies the operation process, and makes it easier for non-professional users to get started. This screening method adopts a modular design, such as Figure 5 As shown, only the catalyst screening space and the catalyst reaction space need to be input to simulate the catalytic space (e.g. Figure 6), thereby constructing a training sample set through the DFT calculation platform in the active learning framework, and then performing graph neural network training and catalyst rapid prediction, and then combining density functional theory verification, that is, using the adsorption energy calculated by DFT to adjust the prediction model, and then performing active learning sampling when the prediction model has not converged, and obtaining the trained prediction model when it converges, thereby predicting the target catalyst, that is, the optimal selection result that meets the needs. In addition, through the continuous optimization of the prediction model, the screening process can adapt to new catalyst design requirements and has extremely high scalability and practicality. Through the organic combination of high-throughput DFT calculations, graph neural network models and active learning strategies, not only the speed and accuracy of CO2RR catalyst screening are significantly improved, but also a convenient and practical solution is provided for the screening and development of new catalysts. This screening method can be widely used in the theoretical design and screening of other reaction catalysts, and has high promotion value and industrialization prospects.

[0077] The embodiment of the present disclosure also provides a screening device for catalysts for carbon dioxide reduction reaction, including: an acquisition module, used to acquire a screening task, the screening task is used to indicate multiple candidate catalysts and a target reduction reaction, the target reduction reaction includes at least one carbon dioxide reduction reaction, and the reaction process information of each carbon dioxide reduction reaction and the geometric structure information of each candidate catalyst are determined; a calculation module, used to input the reaction process information of each carbon dioxide reduction reaction and the geometric structure information of each candidate catalyst into a prediction model for calculation, and obtain the adsorption energy of each candidate catalyst-target intermediate combination, each candidate catalyst-target intermediate combination is formed based on a candidate catalyst and a target intermediate, all target intermediates are determined based on the target reduction reaction, wherein the prediction model is determined based on a graph neural network; a screening module, used to determine the activity and selectivity of the corresponding candidate catalyst according to the adsorption energies of all candidate catalyst-target intermediate combinations corresponding to each candidate catalyst, and screen out the target catalyst from the multiple candidate catalysts according to the activity and selectivity of each candidate catalyst, and use the target catalyst as the screening result of the screening task.

[0078] In this way, the screening device provided by the embodiment of the present disclosure can perform intelligent high-throughput screening based on the graph neural network model and active learning strategy, and is specifically used for the efficient screening and design of catalysts used in CO2 reduction reactions. This screening device not only overcomes the defects of high computational cost and low efficiency in traditional screening devices, but also realizes automatic iterative screening of high-potential catalysts by introducing a prediction model. Through this screening device, several highly selective and highly active catalysts can be screened out, providing the possibility of industrial application for the efficient synthesis of high value-added products such as methanol. In addition, this screening device is versatile, automated and efficient, and is not only suitable for the field of CO2 reduction, but can also be extended to the screening and optimization of other electrochemical catalysts, and has broad industrial prospects and technical value.

[0079] In a possible implementation, the device further includes a training module, which is used to: obtain a training sample set; and iteratively train an initial prediction model using the training sample set to obtain a trained prediction model.

[0080] In a possible implementation, obtaining a training sample set includes: obtaining geometric structure information of multiple sample catalysts and reaction process information of sample reduction reactions, determining each sample catalyst-sample intermediate combination and the corresponding geometric structure information, each sample catalyst-sample intermediate combination is formed based on a sample catalyst and a sample intermediate, and all sample intermediates are determined based on the sample reduction reaction, wherein the sample reduction reaction includes at least one carbon dioxide reduction reaction; calculating the adsorption energy of each sample catalyst-sample intermediate combination based on density functional theory; determining multiple training samples, each of which corresponds to a different sample catalyst-sample intermediate combination, and each of which includes geometric structure information and adsorption energy of the corresponding sample catalyst-sample intermediate combination; and forming the training sample set based on all training samples. In this way, for the sample reduction reaction, all possible sample catalyst-sample intermediate combinations and their geometric structure information and adsorption energy are determined, and the model trained based on this information can more accurately predict the interaction between the catalyst and the intermediate, thereby improving the accuracy of the prediction, and because the model training covers a variety of different catalyst and intermediate combinations, this helps to improve the generalization ability of the prediction model, so that it can adapt to carbon dioxide reduction reactions under different conditions.

[0081] In a possible implementation, the initial prediction model is iteratively trained using the training sample set to obtain a trained prediction model, including: in each iterative training process, a subset used for this iterative training is selected from the training sample set; the current prediction model is trained using the subset to obtain a trained prediction model. In this way, through iterative training, the prediction model can gradually learn a wider range of sample features, thereby improving the generalization ability on unseen data; each iteration is trained only with part of the training samples, which helps to optimize the use of computing resources, especially when processing large-scale data sets, compared to processing the entire data set at one time, it can reduce part of the computing burden, thereby accelerating the training process of the prediction model. In addition, selecting a new subset for training each iteration can make the prediction model more flexible and able to adapt to changes in data distribution; by continuously learning from different subsets, the prediction model can resist the influence of outliers and noise, and enhance the robustness of the prediction model.

[0082] In a possible implementation, selecting a subset from the training sample set for this iterative training includes: when it is determined that the training is currently in the first iterative training, randomly selecting some training samples from the training sample set to obtain a subset of the training sample set.

[0083] In a possible implementation, the subset used for the current iterative training is selected from the training sample set, including: when it is determined that the current training is not the first iterative training, the prediction model obtained in the last iterative training is used to predict the adsorption energy of the sample catalyst-sample intermediate combination corresponding to all the training samples in the training sample set, and the uncertainty corresponding to each of the training samples is determined according to the prediction results, and the subset is formed based on all the training samples whose uncertainties meet the preset conditions. In this way, the model training is based on the training samples with high uncertainty in the prediction results, so that the prediction model can focus more on learning those data points that are difficult to predict, thereby possibly improving the overall prediction accuracy of the prediction model; each iteration selects a new subset based on the current state of the model, dynamically adapts to the learning progress of the prediction model, and helps the prediction model better capture the complexity of the data and reduce the risk of overfitting; under limited computing resources, this training method allows the computing resources to be concentrated on the training samples that need to be learned the most, thereby achieving optimal allocation of resources; and by continuously adjusting the subset used for training, the prediction model can be exposed to more diverse data, which helps to enhance the robustness of the prediction model when facing new data.

[0084] In a possible implementation, the activity and selectivity of the corresponding catalyst to be selected are determined according to the adsorption energies of all the catalyst-target intermediate combinations corresponding to each of the catalysts to be selected, and the target catalyst is screened out from the multiple catalysts to be selected according to the activity and selectivity of each of the catalysts to be selected, including: determining the containment potential of the corresponding catalyst to be selected according to the adsorption energies of all the catalyst-target intermediate combinations corresponding to each of the catalysts to be selected; determining the activity and selectivity of the corresponding catalyst to be selected according to the containment potential of each of the catalysts to be selected; and screening out the target catalyst from the multiple catalysts to be selected according to the preset activity screening conditions and selectivity screening conditions, wherein the target catalyst includes at least one catalyst to be selected. In this way, a catalyst with excellent selectivity and activity can be screened out.

[0085] 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.

[0086] 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.

[0087] 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.

[0088] 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.

[0089] Figure 7 The block diagram of the screening device for the catalyst of the carbon dioxide reduction reaction provided by the embodiment of the present disclosure is shown. For example, the device 1900 can be provided as a server or a terminal device. Figure 7 , 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.

[0090] 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.

[0091] 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.

[0092] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0093] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0094] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0095] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be customized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0096] Various aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0097] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0098] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0099] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of the module, program segment or instruction includes one or more executable instructions for realizing the specified logical function. In some alternative implementations, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of special hardware and computer instructions.

[0100] 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 screening a catalyst for a carbon dioxide reduction reaction, characterized in that: include: Acquire a screening task, the screening task being used to indicate a plurality of candidate catalysts and a target reduction reaction, the target reduction reaction including at least one carbon dioxide reduction reaction, and determine reaction process information of each of the carbon dioxide reduction reactions and geometric structure information of each of the candidate catalysts; Inputting the reaction process information of each of the carbon dioxide reduction reactions and the geometric structure information of each of the candidate catalysts into a prediction model for calculation, and obtaining the adsorption energy of each candidate catalyst-target intermediate combination, wherein each candidate catalyst-target intermediate combination is formed based on a candidate catalyst and a target intermediate, and all target intermediates are determined based on the target reduction reaction, wherein the prediction model is determined based on a graph neural network; The activity and selectivity of the corresponding catalyst are determined based on the adsorption energies of all the catalyst-target intermediate combinations corresponding to each catalyst, and the target catalyst is screened out from the multiple catalysts based on the activity and selectivity of each catalyst, and the target catalyst is used as the screening result of the screening task.

2. The method according to claim 1, characterized in that The method further comprises: Obtain a training sample set; The initial prediction model is iteratively trained using the training sample set to obtain a trained prediction model.

3. The method according to claim 2, characterized in that Get the training sample set, including: Acquire geometric structure information of a plurality of sample catalysts and reaction process information of a sample reduction reaction, determine each sample catalyst-sample intermediate combination and the corresponding geometric structure information, each of the sample catalyst-sample intermediate combinations is formed based on a sample catalyst and a sample intermediate, and all sample intermediates are determined based on the sample reduction reaction, wherein the sample reduction reaction includes at least one carbon dioxide reduction reaction; Calculating the adsorption energy of each sample catalyst-sample intermediate combination based on density functional theory; Determine a plurality of training samples, each of the training samples corresponds to a different sample catalyst-sample intermediate combination, and each of the training samples includes geometric structure information and adsorption energy of the corresponding sample catalyst-sample intermediate combination; The training sample set is formed based on all training samples.

4. The method according to claim 2 or 3, characterized in that: Iteratively training the initial prediction model using the training sample set to obtain a trained prediction model includes: in each iterative training process, Selecting a subset used for this iterative training from the training sample set; The subset is used to train the current prediction model to obtain a trained prediction model.

5. The method according to claim 4, characterized in that A subset used for this iterative training is selected from the training sample set, including: When it is determined that the training is currently in the first iteration, some training samples are randomly selected from the training sample set to obtain a subset of the training sample set.

6. The method according to claim 4, characterized in that A subset used for this iterative training is selected from the training sample set, including: When it is determined that the current training is not the first iteration, the prediction model obtained from the last iteration is used to predict the adsorption energy of the sample catalyst-sample intermediate combination corresponding to all the training samples in the training sample set, and the uncertainty corresponding to each training sample is determined based on the prediction result, and the subset is formed based on all the training samples whose uncertainties meet the preset conditions.

7. The method according to any one of claims 1 to 3, 5 and 6, characterized in that: Determining the activity and selectivity of the corresponding catalyst according to the adsorption energies of all the catalyst-target intermediate combinations corresponding to each catalyst, and screening out the target catalyst from the plurality of catalysts according to the activity and selectivity of each catalyst, including: Determining the containment potential of the corresponding candidate catalyst according to the adsorption energies of all candidate catalyst-target intermediate combinations corresponding to each candidate catalyst; Determining the activity and selectivity of the corresponding catalyst to be selected according to the containment potential of each catalyst to be selected; According to preset activity screening conditions and selectivity screening conditions, a target catalyst is screened out from the plurality of candidate catalysts, and the target catalyst includes at least one candidate catalyst.

8. A screening device for a catalyst for a carbon dioxide reduction reaction, characterized in that: include: an acquisition module, used to acquire a screening task, the screening task is used to indicate a plurality of candidate catalysts and a target reduction reaction, the target reduction reaction includes at least one carbon dioxide reduction reaction, and determine the reaction process information of each of the carbon dioxide reduction reactions and the geometric structure information of each of the candidate catalysts; A calculation module, for inputting the reaction process information of each of the carbon dioxide reduction reactions and the geometric structure information of each of the candidate catalysts into a prediction model for calculation, to obtain the adsorption energy of each candidate catalyst-target intermediate combination, wherein each candidate catalyst-target intermediate combination is formed based on a candidate catalyst and a target intermediate, and all target intermediates are determined based on the target reduction reaction, wherein the prediction model is determined based on a graph neural network; A screening module is used to determine the activity and selectivity of the corresponding candidate catalyst according to the adsorption energy of all candidate catalyst-target intermediate combinations corresponding to each candidate catalyst, and to screen out a target catalyst from the multiple candidate catalysts according to the activity and selectivity of each candidate catalyst, and to use the target catalyst as the screening result of the screening task.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to implement the method described in any one of claims 1 to 7 when executing the instructions stored in the memory.

10. 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 7 is implemented.

Images