Method, system and medium for analyzing, predicting and screening oxygen release performance of material based on data driving
Through a data-driven method, using text mining and machine learning models, the material structure-oxygen release performance data set is constructed, which solves the problem of inefficiency in trial and error methods in oxygen carrier material design, and realizes efficient and accurate material screening, promoting Progress in low-carbon energy utilization technology.
Patent Information
- Application Number
- CN202411269065.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-11
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-09-11
AI Technical Summary
When studying the oxygen release properties of oxygen carrier materials, the prior art relies on trial and error methods and calculation screening methods, resulting in the process of designing high-performance materials being complex and inefficient, and it is impossible to quickly screen out high-performance materials.
Using a data-driven method, the data set of material structure-oxygen release performance is reduced by text mining of publicly published literature, and the regression prediction and high-throughput screening of material oxygen release performance is used to reduce experimental dependence. and error improves the accuracy and efficiency of screening.
It realizes low-cost and high-reliability material screening, reduces the selection range of oxygen-releasing materials, improves research efficiency and accuracy, saves resources and energy, and is in line with the concept of green chemistry and sustainable development.
Smart Images

Figure CN119049589A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of chemical chaining oxygen decoupled combustion and chemical chaining air separation, and in particular to a data-driven material oxygen release performance analysis prediction and screening method, system and medium. Background Art
[0002] The massive emission of carbon dioxide has caused severe climate problems such as global warming, and has posed new challenges to the development of low-carbon energy utilization technologies. Chemical chaining oxygen decoupled combustion technology and chemical chaining air separation technology are both clean, efficient and low-energy low-carbon energy utilization technologies. Their principles are that metal oxides release oxygen at high temperatures. Therefore, the oxygen release performance of metal oxide materials as oxygen carriers is the most important research content, and the lack of oxygen release performance of materials is one of the main reasons limiting commercial applications.
[0003] At present, the main research methods for the oxygen release performance of oxygen carrier materials are: (1) experimental trial and error method, using XRD and other methods to characterize the material structure, and using thermogravimetric experiments to characterize the oxygen release performance and cycle stability. The focus of experimental research is on the modification and efficiency enhancement of materials. By changing the material preparation method, crystal structure, microscopic morphology, etc., the oxygen release performance, anti-sintering and other macroscopic performance of oxygen carrier materials can be improved; (2) computational screening method, using density functional theory, molecular simulation and other computational methods to explore the specific mechanism of the chemical chain reaction process and the oxygen release potential of the material from a microscopic perspective, calculate the lattice oxygen activity and adsorption energy, and reveal the atomic-level reaction mechanism, reaction path, rate-controlling step and other information. However, both experimental and computational research can only be considered as trial and error research, which can only provide limited empirical guidance. The huge element combination and structural variety of multi-metal oxides means that the design and development of high-performance oxygen carriers must break through the research model based on trial and error. How to quickly screen high-performance oxygen carriers is still a huge challenge. Summary of the invention
[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a data-driven material oxygen release performance analysis prediction and screening method, system, and medium. Compared with the traditional trial-and-error research method, it has low cost and high reliability, greatly reduces the screening range of oxygen-releasing materials, and helps to discover high-performance oxygen-releasing materials.
[0005] During the process of conception, the applicant believes that the data-driven research paradigm has broad application prospects for objects with incomplete basic principles and lack of original prior knowledge, but with large amounts of real data available. Using the data-driven paradigm to study the oxygen release performance of materials, it is possible to collect the structure and oxygen release data of metal oxide materials to construct a data set, train a machine learning model, and then realize the analysis, prediction and screening of the oxygen release performance of materials.
[0006] The purpose of the present invention can be achieved by the following technical solutions:
[0007] The present invention provides a data-driven material oxygen release performance analysis prediction and screening method, comprising the following steps:
[0008] S1: Text mining was performed on the publicly available literature on oxygen-releasing materials to extract data related to structure and oxygen-releasing performance, and a “material structure-oxygen-releasing performance” dataset was constructed;
[0009] S2: Based on the machine learning model, the X-ray diffraction data corresponding to the material structure information in the data set is used as input, and the oxygen release performance data corresponding to the material performance information in the data set is used as output to construct a regression prediction model for the material oxygen release performance;
[0010] S3: Use the open source crystal database to obtain the material's structure file, convert it into XRD data, input it into the trained model, output the regression prediction of the material's oxygen release performance, and achieve high-throughput screening of materials.
[0011] Furthermore, in S1, the data set construction process includes:
[0012] S1-1: Based on keywords, search and collect publicly published documents in the literature index database;
[0013] S1-2: Perform text mining on the collected literature data to extract data related to material structure and performance;
[0014] S1-3: The atlas information is digitally extracted, and the “material structure-oxygen release performance” dataset is constructed using the material name, structure data, performance data, image serial number, and digital object unique identifier as the dataset identifier.
[0015] Further, in S1-1, the keywords include but are not limited to: oxygen release, chemical chain, oxygen decoupling;
[0016] In S1-1, the literature index database includes but is not limited to CNKI, Web of Science, Elsevier, etc.;
[0017] In S1-2, the data related to the material structure include but are not limited to XRD spectrum, specific surface area, and elemental composition;
[0018] In S1-2, the data related to the material properties include but are not limited to thermogravimetric curve, oxygen release amount, and oxygen release rate.
[0019] In S1-2, the specific method of text mining includes: based on the XRD spectrum, thermogravimetric curve and other spectrum data that can uniformly characterize the structure and performance of the material, according to the keywords that appear frequently in the spectrum title, use the text mining tool to take the figure number information of each spectrum as the extraction target. Finally, the digital object unique identifier and figure number of the publicly published literature are used to realize the positioning extraction of each spectrum information, and to realize the extraction of material structure and performance related data from the publicly published literature data. Text mining tools include but are not limited to: manual retrieval, natural language processing, large language model, artificial intelligence, etc.;
[0020] Further, in S1-3, the process of digital extraction includes:
[0021] Use image processing tools to digitally extract XRD patterns and perform normalization to unify the data range of diffraction intensity, and expand or delete according to the diffraction angle range to achieve structuring of XRD data;
[0022] After digitizing the thermogravimetric curve and other oxygen release performance-related graphs, the oxygen release amount information and oxygen release rate information are extracted, the starting oxygen release point and the ending oxygen release point are defined, the oxygen release amount is calculated, and the differential thermogravimetric curve is obtained by differentiating the thermogravimetric curve to determine the optimal oxygen release point and the maximum oxygen release rate.
[0023] Furthermore, in the process of structuring the XRD spectrum, the following steps are included: in view of the different diffraction intensity ranges and diffraction angle ranges of XRD spectra in different literatures, on the one hand, in the image processing tool, the diffraction intensity of each individual XRD spectrum is normalized, and the data range of the diffraction intensity is unified to be between 0 and 1; on the other hand, based on the diffraction angle range given in the spectrum, it is expanded or deleted, and the initial angle diffraction intensity or the end angle diffraction intensity of the original spectrum is used as a reference, and the diffraction angle range is unified to 10 to 85°, with an angle step of 0.05°, to achieve structuring of the XRD data.
[0024] The process of determining the oxygen release data, the optimal oxygen release point and the maximum oxygen release rate from the thermogravimetric curve includes: defining the inflection point where the thermogravimetric curve shows obvious weight loss as the starting oxygen release point, defining the point where the thermogravimetric curve gradually smoothes as the ending oxygen release point, and the difference in mass percentage between the two points is the oxygen release amount. Differentiating the thermogravimetric curve to obtain the derivative thermogravimetric curve, the highest point of the derivative thermogravimetric curve is the optimal oxygen release point, which corresponds to the maximum oxygen release rate.
[0025] Furthermore, the machine learning model includes one or more of a convolutional neural network, an artificial neural network, a random forest, and a Transformer, and the machine learning model is used to achieve regression prediction of the oxygen release performance of the material.
[0026] Further, in S3, the high-throughput screening comprises the following steps:
[0027] S3-1: Batch obtain crystal information files of potential oxygen-releasing metal oxides from the open source crystal database and convert them into XRD data that meet the model input requirements;
[0028] S3-2: Input the structured XRD data into the trained material oxygen release performance regression prediction model, output the regression prediction of oxygen release performance, and realize high-throughput screening of oxygen-releasing materials.
[0029] Furthermore, the open source crystal database includes but is not limited to one or more of Materials Project Database, Open Quantum Materials Database, Inorganic Crystal Structure Database, and Powder Diffraction File Database.
[0030] The second aspect of the present invention provides a data-driven material oxygen release performance analysis prediction and screening system, comprising:
[0031] The data processing module conducts text mining on the publicly published literature on oxygen-releasing materials, extracts data related to structure and oxygen-releasing performance, and constructs a “material structure-oxygen-releasing performance” data set;
[0032] A model building module, based on a machine learning model, takes the X-ray diffraction data corresponding to the material structure information in the data set as input, and takes the oxygen release performance data corresponding to the material performance information in the data set as output, to build a regression prediction model for the material oxygen release performance;
[0033] The material screening module uses an open source crystal database to obtain the material's structure file, converts it into XRD data, inputs it into a trained model, and outputs a regression prediction of the material's oxygen release performance to achieve high-throughput screening of materials.
[0034] A third aspect of the present invention provides a storage medium comprising computer executable instructions, which, when executed by a computer processor, is used to perform the above-mentioned data-driven material oxygen release performance analysis prediction and screening method.
[0035] Compared with the prior art, the present invention has the following technical advantages:
[0036] 1) The present invention uses advanced text mining technology and digital processing methods to efficiently extract key data on material structure and oxygen release performance from a large number of publicly published literature, effectively solving the problems of difficult and inefficient data acquisition in the traditional material screening process. Using this data-driven analysis and prediction method, materials with potential high oxygen release performance can be quickly identified at a relatively low cost, greatly improving the efficiency and economy of the research.
[0037] 2) By constructing a mapping relationship between material structure data and oxygen release performance based on machine learning, the present invention significantly improves the accuracy and reliability of oxygen-releasing material screening. This method reduces the dependence on experimental conditions, avoids errors caused by experimental operations and environmental factors, and ensures the stability and repeatability of screening results. Compared with traditional experimental tests, the method of the present invention is more objective and accurate, providing a powerful tool for researchers in the field of materials science.
[0038] 3) The high-throughput screening mechanism of the present invention greatly reduces the screening range of oxygen-releasing materials, allowing researchers to focus resources on in-depth research and experimental verification of the materials that perform best in the predicted results. This method not only saves valuable time and human resources, but also reduces the consumption of materials and energy, which is in line with the concepts of green chemistry and sustainable development. By quickly identifying and optimizing oxygen-releasing materials, the present invention provides an important scientific basis and technical support for promoting the advancement of energy conversion and storage technologies. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 A workflow diagram of a method for analyzing, predicting and screening oxygen release performance of materials based on data-driven methods according to an embodiment of the present invention;
[0040] Figure 2 A schematic diagram of the structure of a system for analyzing, predicting and screening material oxygen release performance based on data-driven methods according to an embodiment of the present invention;
[0041] Figure 3 A comparison chart of the predicted value and the actual value of the oxygen release performance of the material based on the random forest model provided according to an embodiment of the present invention;
[0042] Figure 4 A comparison chart of the predicted value and the actual value of the oxygen release performance of the material based on the artificial neural network model provided according to an embodiment of the present invention;
[0043] Figure 5 This is a comparison chart between the predicted value and the actual value of the material oxygen release performance based on the convolutional neural network model provided according to an embodiment of the present invention.
[0044] Figure 6This is a comparison chart between the predicted value and the actual value of the oxygen release performance of the material based on the Transformer model provided according to an embodiment of the present invention.
[0045] In the figure: data processing module-10, model building module-20, material screening module-30. DETAILED DESCRIPTION
[0046] On the whole, the present invention provides a method and system for analyzing, predicting and screening oxygen-release performance of materials based on data-driven, wherein the specific method is: text mining of publicly published literature data on oxygen-releasing materials, obtaining structural information of the materials and data related to oxygen-release performance, and constructing an integrated data set of "material structure-oxygen-release performance"; based on the data set, constructing a machine learning model, inputting material structural characteristics to obtain regression predictions for oxygen-release performance; obtaining structural data of possible oxygen-releasing materials based on an open source crystal database, and inputting it into a trained and optimized model to achieve high-throughput screening of oxygen-releasing materials. The present invention mines data from publicly published literature, and constructs a mapping relationship between material structural data and oxygen-release performance through a machine learning model, which is low-cost and highly reliable, greatly reduces the screening range of oxygen-releasing materials, and avoids the time-consuming and labor-intensive problems of traditional material experimental testing.
[0047] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. Component models, material names, connection structures, control methods, algorithms and other features not clearly described in this technical solution are all considered to be common technical features disclosed in the prior art.
[0048] Example 1
[0049] This embodiment discloses a method for analyzing, predicting and screening oxygen release performance of materials based on data-driven, reference Figure 1 As shown, the specific steps include:
[0050] S1. Text mining was performed on the publicly published literature on the study of oxygen-releasing materials to extract data related to structure and oxygen-releasing performance. The atlas data was digitally extracted to construct an integrated data set of "material structure-oxygen-releasing performance".
[0051] S2: Based on the machine learning model, the X-ray diffraction data corresponding to the material structure information in the data set is used as input, and the oxygen release performance data corresponding to the material performance information in the data set is used as output to construct a regression prediction model for the material oxygen release performance;
[0052] S3 uses an open source crystal database to obtain the material's structure file, converts it into XRD data, inputs it into a trained model, and outputs a regression prediction of the material's oxygen release performance to achieve high-throughput screening of materials.
[0053] Specifically,
[0054] In step S1, the extraction of material structure and performance-related data based on text mining to construct a data set includes:
[0055] First, based on keywords such as oxygen release, chemical chain, and oxygen decoupling, publicly published literature was searched and collected in literature index databases, including but not limited to CNKI, Web of Science, Elsevier, etc.
[0056] Secondly, text mining is performed on the data of publicly published literature to extract data related to the structure and performance of the materials in the publicly published literature. Structural data include but are not limited to XRD spectra, specific surface area, elemental composition, etc., and performance data include but are not limited to thermogravimetric curves, oxygen release amount, oxygen release rate, etc. Text mining methods include but are not limited to manual retrieval, natural language processing, large language models, artificial intelligence and other methods.
[0057] Then, the spectral information related to the structure and performance is digitally extracted, including but not limited to XRD spectrum, thermogravimetric analysis curve, N 2 Adsorption curves, etc.
[0058] Finally, an integrated “material structure-oxygen release performance” dataset containing 451 sets of data was constructed using the material names, structural data, performance data, image serial numbers, and digital object unique identifiers in publicly published literature as dataset identifiers.
[0059] In step S1, the specific process and details of the text mining method include:
[0060] Use text mining tools to extract the structure and performance data of the materials studied in the literature. The structural data include but are not limited to XRD spectra, specific surface area, elemental composition and other information, and the performance data include but are not limited to thermogravimetric curves, oxygen release amount, oxygen release rate and other information. In view of the problem that there are wide differences in the descriptions of structure and performance-related data in the publicly published literature, especially the lack of a unified description of oxygen release performance-related data, the main objects of text mining information extraction are focused on XRD spectra widely used to characterize structures and thermogravimetric curves, micro-thermogravimetric curves and other spectrum information that can uniformly characterize the oxygen release behavior of materials. According to the keywords that appear frequently in the spectrum title, the figure number information of each spectrum is used as the extraction target using text mining tools. Finally, the digital object unique identifier and figure number of the publicly published literature are used to realize the positioning extraction of each spectrum information, and the structure and performance-related data of the material are extracted from the publicly published literature data. Text mining tools include but are not limited to: manual retrieval, natural language processing, large language model, artificial intelligence, etc.
[0061] In step S1, the method of digital extraction and data structure processing includes:
[0062] Image processing tools are used to realize the digital extraction of spectrum information. First, the XRD data is structured. In view of the different diffraction intensity ranges and diffraction angle ranges of XRD spectra in different literatures, on the one hand, the diffraction intensity of each individual XRD spectrum is normalized in the image processing tool, and the data range of the diffraction intensity is unified between 0 and 1; on the other hand, based on the diffraction angle range given in the spectrum, it is expanded or deleted, and the initial angle diffraction intensity or the end angle diffraction intensity of the original spectrum is used as a reference to unify the diffraction angle range to 10-85°, with an angle step of 0.05°, to achieve the structuring of XRD data.
[0063] Then, after the thermogravimetric curve and other performance-related graphs are digitized, information such as oxygen release amount and oxygen release rate are further extracted. The inflection point where the thermogravimetric curve shows obvious weight loss is defined as the starting oxygen release point, and the point where the thermogravimetric curve gradually becomes smooth is defined as the ending oxygen release point. The difference in mass percentage between the two points is the oxygen release amount. The differential thermogravimetric curve can be obtained by differentiating the thermogravimetric curve. The highest point of the differential thermogravimetric curve is the optimal oxygen release point, which corresponds to the maximum oxygen release rate.
[0064] In step S2, the oxygen release performance analysis and prediction model based on the machine learning algorithm is constructed, including:
[0065] With structured XRD data as input and oxygen release performance data as output, a machine learning model is constructed to achieve regression prediction of material oxygen release performance. Machine learning models include but are not limited to convolutional neural networks, artificial neural networks, random forests, Transformer, etc.
[0066] In this step, the structural feature XRD data in the constructed “material structure-oxygen release performance” integrated data set was used as the model input, and the material oxygen release data was used as the model output. Four machine learning models were constructed, such as Figures 3 to 6 As shown, they are random forest models ( Figure 3 ), artificial neural network model ( Figure 4 ), convolutional neural network model ( Figure 5 )、Transformer model( Figure 6 ), all the data in the constructed data set are divided into training set and test set in a ratio of 9:1. The training set is used to train the model, and the test set is used to verify the model prediction effect. After the model is trained, the model is evaluated based on the mean square error.
[0067] The structural characteristic XRD data of the material were input into four different machine learning models to obtain the predicted value of the oxygen release of the material. In this step, the effect of training the random forest model was that the mean square error of the training set was 0.2677 and the mean square error of the test set was 1.3912. The comparison between the oxygen release of the material predicted by the model and the true value is shown in the figure below. Figure 3 As shown in the figure, the effect of the trained artificial neural network model is that the mean square error of the training set is 0.0881, and the mean square error of the test set is 1.1714. The comparison between the oxygen release of the material predicted by the model and the actual value is shown in the figure Figure 4 As shown in the figure, the effect of the trained convolutional neural network model is that the mean square error of the training set is 0.0685, and the mean square error of the test set is 0.7396. The comparison between the oxygen release of the material predicted by the model and the actual value is shown in the figure Figure 5 The effect of the trained Transformer model is that the mean square error of the training set is 0.0559, and the mean square error of the test set is 0.9516. The comparison between the oxygen release of the material predicted by the model and the actual value is shown in the figure below. Figure 6 shown.
[0068] In step S3, the high-throughput screening of materials based on the open source crystal database includes:
[0069] First, batches of crystal information files of possible oxygen-releasing metal oxides are obtained from open source crystal databases and converted into XRD data that meet the model input requirements. Open source crystal databases include but are not limited to one or more of Materials Project Database, Open Quantum Materials Database, Inorganic Crystal Structure Database, and Powder Diffraction File Database.
[0070] Secondly, the structured XRD data is input into the trained model to output the regression prediction of oxygen release performance, thus achieving high-throughput screening of oxygen-releasing materials.
[0071] In an embodiment of the present invention, text mining is performed on publicly available literature data to extract data related to the structure and performance of the material in the publicly available literature, and after digital image extraction and data structuring, an integrated data set of "material structure-oxygen release performance" is constructed; based on the constructed data set, the structural characteristic XRD data is used as an input variable to predict the oxygen release of the material with high accuracy; the structure file is obtained from an open source crystal database and converted into XRD data that conforms to the model input, and a prediction of the material oxygen release data is output. The model realizes high-throughput screening of high-performance oxygen-releasing materials, greatly reduces the screening range of oxygen-releasing materials, and avoids the time-consuming and labor-intensive problems of traditional material experimental testing.
[0072] The present invention also discloses a data-driven material oxygen release performance analysis prediction and screening system, referring to Figure 2 As shown, specifically including:
[0073] Data processing module 10, performs text mining on publicly published literature on oxygen-releasing materials, extracts data related to structure and oxygen-releasing performance, and constructs a "material structure-oxygen-releasing performance" data set;
[0074] A model building module 20, based on a machine learning model, takes the X-ray diffraction data corresponding to the material structure information in the data set as input, and takes the oxygen release performance data corresponding to the material performance information in the data set as output, to build a material oxygen release performance regression prediction model;
[0075] The material screening module 30 uses an open source crystal database to obtain the material's structure file, converts it into XRD data, inputs it into a trained model, outputs a regression prediction of the material's oxygen release performance, and realizes high-throughput screening of materials.
[0076] The modules described above as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.
[0077] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The above integrated modules may be implemented in the form of hardware or in the form of software functional units.
[0078] This embodiment also proposes a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable the computer to execute the above-mentioned data-driven material oxygen release performance analysis prediction and screening method. The storage medium can be an electronic medium, a magnetic medium, an optical medium, an electromagnetic medium, an infrared medium or a semiconductor system or a propagation medium. The storage medium can also include semiconductor or solid-state memory, a magnetic tape, a removable computer disk, a random access memory (RAM), a read-only memory (ROM), a hard disk and an optical disk. The optical disk can include a compact disk-read only memory (CD-ROM), a compact disk-read and write (CD-RW) and a DVD.
[0079] The above description of the embodiments is to facilitate the understanding and use of the invention by those skilled in the art. It is obvious that those skilled in the art can easily make various modifications to these embodiments and apply the general principles described herein to other embodiments without creative work. Therefore, the present invention is not limited to the above embodiments, and improvements and modifications made by those skilled in the art based on the disclosure of the present invention without departing from the scope of the present invention should be within the scope of protection of the present invention.
Claims
1. A data-driven material oxygen release performance analysis, prediction and screening method, characterized in that: The following steps are involved: S1: Text mining was performed on the published literature on oxygen-releasing materials to extract data related to structure and oxygen-releasing performance, and a "material structure-oxygen-releasing performance" dataset was constructed; S2: Based on a machine learning model, using the X-ray diffraction data corresponding to the material structure information in the data set as input and the oxygen release performance data corresponding to the material performance information in the data set as output, a regression prediction model for the material oxygen release performance is constructed; S3: Use open source crystal databases to obtain the material's structural file, convert it into XRD data, input it into the trained model, and output the regression prediction of the material's oxygen release performance to achieve high-throughput screening of materials.
2. The data-driven material oxygen release performance analysis, prediction and screening method according to claim 1, characterized in that: In S1, the dataset construction process includes: S1-1: Search and collect published literature in literature index databases based on keywords; S1-2: Perform text mining on the collected literature data to extract data related to material structure and performance; S1-3: Digitally extract the atlas information and construct a "material structure-oxygen release performance" dataset using the material name, structure data, performance data, image serial number, and digital object unique identifier as the dataset identifier.
3. The data-driven material oxygen release performance analysis, prediction and screening method according to claim 2, characterized in that: In S1-1, the keywords include: oxygen release, chemical chaining, oxygen decoupling; In S1-1, the literature index databases include HowNet, Web of Science, and Elsevier; In S1-2, the material structure-related data include XRD patterns, specific surface area, and elemental composition; In S1-2, the data related to the material properties include thermogravimetric curve, oxygen release amount, and oxygen release rate; In S1-2, the specific process of text mining includes: Based on XRD patterns, thermogravimetric curves and other spectral data that can uniformly characterize material structure and properties, text mining tools are used to extract the image number information of each spectrum based on the keywords that appear frequently in the spectral titles. Finally, the unique digital object identifiers and figure serial numbers of the published literature are used to locate and extract the information of each atlas, and to extract the structure and performance related data of the material from the published literature data; The text mining tools include: manual retrieval, natural language processing, and large language models.
4. The data-driven material oxygen release performance analysis, prediction and screening method according to claim 2, characterized in that: In S1-3, the digital extraction process includes: Use image processing tools to digitally extract XRD patterns and perform normalization to unify the data range of diffraction intensity, and expand or delete according to the diffraction angle range to achieve structuring of XRD data; After digitizing the thermogravimetric curve and other oxygen release performance-related graphs, the oxygen release amount and oxygen release rate information are extracted, the starting oxygen release point and the ending oxygen release point are defined, the oxygen release amount is calculated, and the differential thermogravimetric curve is obtained by performing a differential operation on the thermogravimetric curve to determine the optimal oxygen release point and the maximum oxygen release rate.
5. The data-driven material oxygen release performance analysis, prediction and screening method according to claim 4, characterized in that: The process of structural processing of XRD data and extraction of thermogravimetric curve data includes: In view of the different diffraction intensity ranges and diffraction angle ranges selected for XRD patterns in different literatures, on the one hand, the diffraction intensity of each individual XRD pattern was normalized in the image processing tool to unify the diffraction intensity data range between 0 and 1. On the other hand, based on the diffraction angle range given in the pattern, it was expanded or deleted. The initial angle diffraction intensity or the end angle diffraction intensity of the original pattern was used as a benchmark to unify the diffraction angle range to 10-85° with an angle step of 0.05°, thus achieving structuring of XRD data. The process of determining the oxygen release amount, the optimal oxygen release point and the maximum oxygen release rate from the thermogravimetric curve includes: defining the inflection point where the thermogravimetric curve shows obvious weight loss as the starting oxygen release point, and defining the point where the thermogravimetric curve gradually smoothes as the ending oxygen release point. The difference in mass percentage between the two points is the oxygen release amount. Differentiating the thermogravimetric curve obtains the derivative thermogravimetric curve. The highest point of the derivative thermogravimetric curve is the optimal oxygen release point, which corresponds to the maximum oxygen release rate.
6. The data-driven material oxygen release performance analysis, prediction and screening method according to claim 1, characterized in that: The machine learning model includes one or more of a convolutional neural network, an artificial neural network, a random forest, and a Transformer, and the machine learning model is used to achieve regression prediction of the oxygen release performance of the material.
7. The data-driven material oxygen release performance analysis, prediction and screening method according to claim 1, characterized in that: In S3, the high-throughput screening comprises the following steps: S3-1: Batch obtain crystal information files of potential oxygen-releasing metal oxides from open source crystal databases and convert them into XRD data that meet the model input requirements; S3-2: Input structured XRD data into the trained material oxygen release performance regression prediction model, output the regression prediction of oxygen release performance, and achieve high-throughput screening of oxygen-releasing materials.
8. The data-driven material oxygen release performance analysis, prediction and screening method according to claim 7, characterized in that: The open source crystal database includes one or more of Materials Project Database, Open Quantum Materials Database, Inorganic Crystal Structure Database, and Powder Diffraction File Database.
9. A data-driven material oxygen release performance analysis, prediction and screening system, characterized in that: include: The data processing module conducts text mining on published literature on oxygen-releasing materials, extracts data related to structure and oxygen-releasing performance, and constructs a "material structure-oxygen-releasing performance" dataset; A model building module, based on a machine learning model, takes the X-ray diffraction data corresponding to the material structure information in the data set as input, and takes the oxygen release performance data corresponding to the material performance information in the data set as output, to build a material oxygen release performance regression prediction model; The material screening module uses an open source crystal database to obtain the material's structural file, converts it into XRD data, inputs it into a trained model, and outputs a regression prediction of the material's oxygen release performance to achieve high-throughput screening of materials.
10. A storage medium containing computer-executable instructions, characterized in that: When the computer processor executes the computer executable instruction storage medium, the computer executable instruction storage medium is used to perform the data-driven material oxygen release performance analysis, prediction and screening method according to any one of claims 1 to 8.
Citation Information
Patent Citations
Gasoline engine fuel injection quantity control method based on three-way catalyst oxygen storage state monitoring result
CN115217659A
High-throughput screening method, system, equipment and medium for high-temperature-resistant soft magnetic material
CN117954012A
Method and server for adaptive paper search using machine learning
KR102059309B1
Cited By
Rapid prediction method and system for thermal weight loss property of chemical-looping oxygen carrier material based on data driving, medium and program product
CN120072134A
Data-driven rapid prediction method, system, medium and program product for thermal gravimetric properties of chemical looping oxygen carrier materials
CN120072134B
A cementing material production energy consumption evaluation method based on internet of things collection
CN120317735B
