CN116564455B - Method and device for screening anode material of ion battery, equipment and medium - Google Patents

Method and device for screening anode material of ion battery, equipment and medium Download PDF

Info

Publication number
CN116564455B
CN116564455B CN202310820212.3A CN202310820212A CN116564455B CN 116564455 B CN116564455 B CN 116564455B CN 202310820212 A CN202310820212 A CN 202310820212A CN 116564455 B CN116564455 B CN 116564455B
Authority
CN
China
Prior art keywords
screening
preset
data
potassium
topological quantum
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202310820212.3A
Other languages
Chinese (zh)
Other versions
CN116564455A (en
Inventor
郄瑜
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Suzhou Inspur Intelligent Technology Co Ltd
Original Assignee
Suzhou Inspur Intelligent Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Suzhou Inspur Intelligent Technology Co Ltd filed Critical Suzhou Inspur Intelligent Technology Co Ltd
Priority to CN202310820212.3A priority Critical patent/CN116564455B/en
Publication of CN116564455A publication Critical patent/CN116564455A/en
Application granted granted Critical
Publication of CN116564455B publication Critical patent/CN116564455B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
    • G16C60/00—Computational materials science, i.e. ICT specially adapted for investigating the physical or chemical properties of materials or phenomena associated with their design, synthesis, processing, characterisation or utilisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/26—Composites
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y02E60/10—Energy storage using batteries

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Mathematical Physics (AREA)
  • Molecular Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Biology (AREA)
  • Medical Informatics (AREA)
  • Geometry (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)

Abstract

The invention relates to the technical field of battery material screening, in particular to a method, a device, a battery, equipment and a medium for screening an anode material of an ion battery. The method comprises the following steps: collecting a plurality of topological quantum material data to be predicted containing preset metals and performing pretreatment operation to obtain samples to be screened; predicting a sample to be screened by utilizing a pre-trained atomic line graph neural network to obtain a predicted ion migration barrier corresponding to the sample to be screened; performing first screening on the sample to be screened based on the predicted ion migration barrier; and (3) carrying out second screening on the topological quantum material containing the preset metal after the first screening based on the first principle to obtain an alternative material of the anode of the ion battery containing the preset metal. The scheme of the invention greatly reduces the data calculation amount, accelerates the material research and design process, and obviously reduces the experiment cost and the time cost.

Description

Method and device for screening anode material of ion battery, equipment and medium
Technical Field
The invention relates to the technical field of battery material screening, in particular to a method, a device, a battery, equipment and a medium for screening an anode material of an ion battery.
Background
With the continuous development of society, the demand for sustainable energy in daily life and industrial activities is continuously increasing, and secondary metal ion batteries, particularly lithium ion batteries, play an important role in the development of sustainable energy. The global reserve of lithium resources is limited, and with the development of new energy automobiles, the demand for batteries is greatly increased, and the resource bottleneck is gradually developed. Potassium ion batteries (KIBs) are paid attention to by researchers by virtue of high abundance of potassium element in the crust, rapid diffusion rate of potassium ion in electrolyte solution, and the like. As a key component of the secondary metal ion battery, the positive electrode material has obvious influence on the electrochemical properties of the battery, such as energy density, cycle stability, ion conductivity and the like. However, the size of potassium ions is larger than that of lithium ions and sodium ions, and a positive electrode material system of lithium ions or sodium ions cannot be simply and directly applied to a positive electrode material of a potassium ion battery. Therefore, the research of the anode materials of the ion batteries with different metal materials has important significance.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a method, apparatus, battery, device and medium for screening an anode material of an ion battery.
According to a first aspect of the present invention, there is provided a method of screening an anode material of an ion battery, the method comprising:
collecting a plurality of topological quantum material data to be predicted containing preset metals and performing pretreatment operation to obtain samples to be screened;
predicting a sample to be screened by utilizing a pre-trained atomic line graph neural network to obtain a predicted ion migration barrier corresponding to the sample to be screened;
performing first screening on the sample to be screened based on the predicted ion migration barrier to obtain a topological quantum material containing the preset metal after the first screening;
and (3) carrying out second screening on the topological quantum material containing the preset metal after the first screening based on the first principle to obtain an alternative material of the anode of the ion battery containing the preset metal.
In some embodiments, the method further comprises:
collecting crystal structure data containing preset metals and corresponding preset metal ion migration barrier data, and performing pretreatment operation to obtain training samples;
and training the atomic line graph neural network by using the training sample to obtain a pre-trained atomic line graph neural network.
In some embodiments, the preprocessing operation includes screening data according to fractional occupancy structure.
In some embodiments, the step of collecting a plurality of topological quantum material data to be predicted containing a preset metal and performing a preprocessing operation to obtain a sample to be screened includes:
and removing the data with the fraction occupation structure in the topological quantum material data to be predicted so as to obtain a sample to be screened.
In some embodiments, the step of collecting the crystal structure data containing the predetermined metal and the corresponding predetermined metal ion migration barrier data and performing a preprocessing operation to obtain the training sample includes:
counting the number distribution of atoms in each topological quantum material unit cell which contains preset metal and is to be predicted in the sample to be screened;
determining a minimum and a maximum screening value for the unit cell atomic number based on the intra-unit atomic number distribution;
and screening the crystal structure data containing the preset metal and the corresponding preset metal ion migration barrier data based on the maximum screening value and the minimum screening value to obtain a training sample.
In some embodiments, the step of determining a minimum and a maximum screening value for the unit cell atomic number based on the intra-cell atomic number distribution comprises:
acquiring the minimum atomic number and the maximum atomic number in the atomic number distribution in a unit cell;
The minimum number of atoms is taken as the minimum screening value, and the maximum number of atoms is taken as the maximum screening value.
In some embodiments, the step of determining a minimum and a maximum screening value for the unit cell atomic number based on the intra-cell atomic number distribution comprises:
acquiring the minimum atomic number and the maximum atomic number in the atomic number distribution in a unit cell;
responding to the minimum atomic number equal to zero or the minimum atomic number minus a preset number less than or equal to zero, and taking zero as a minimum screening value;
responding to the fact that the minimum atomic number is not equal to zero and the difference value between the minimum atomic number and the preset number is taken as a minimum screening value when the preset number is larger than zero and the minimum atomic number is subtracted from the minimum atomic number;
and increasing the maximum atomic number by the preset number to serve as a maximum screening value.
In some embodiments, the step of screening the preset metal-containing crystal structure data and the corresponding preset metal ion migration barrier data based on the maximum screening value and the minimum screening value to obtain a training sample includes:
counting the number of atoms in each unit cell of the collected crystal structure data containing the preset metal;
And taking the crystal structure data containing the preset metal, of which the atomic number in the unit cell is more than or equal to the minimum screening value and less than or equal to the maximum screening value, and the corresponding preset metal ion migration barrier data as training samples.
In some embodiments, the step of collecting the crystal structure data containing the predetermined metal and the corresponding predetermined metal ion migration barrier data and performing a preprocessing operation to obtain the training sample includes:
and removing the data with the fractional occupation structure in the crystal structure data containing the preset metal, and taking the obtained crystal structure data containing the preset metal and the corresponding preset metal ion migration barrier data as training samples.
In some embodiments, the step of training the atomic line graph neural network using the training sample to obtain a pre-trained atomic line graph neural network includes:
taking the crystal structure data containing the preset metal in the training sample as the input of the atomic diagram neural network, and taking the corresponding preset metal ion migration barrier data in the training sample as the expected output of the atomic diagram neural network for training;
evaluating the atomic line graph neural network according to the determination coefficient and the average absolute error until the atomic line graph neural network converges;
And taking the atomic line graph neural network at the time of convergence as the pre-trained atomic line graph neural network.
In some embodiments, the step of evaluating the atomic line graph neural network according to the decision coefficient and the mean absolute error until the atomic line graph neural network converges includes:
responding to all data in the training sample to complete one training, and respectively calculating a decision coefficient and an average absolute error according to a formula I and a formula II;
formula one;
a second formula;
wherein, in the formula I and the formula IIRepresenting the decision coefficient->Mean absolute error +.>Representing the size of the predicted dataset, +.>Representing the true value of the preset metal ion migration barrier data, < + >>Representing a predicted value of preset metal ion migration barrier data;
and responding to the difference value between the determining coefficient and the first value is smaller than or equal to a first preset threshold value, and the difference value between the average absolute error and zero is smaller than or equal to a second preset threshold value, and confirming convergence of the atomic line graph neural network.
In some embodiments, the step of predicting the sample to be screened by using the pre-trained atomic line graph neural network to obtain a predicted ion migration barrier corresponding to the sample to be screened includes:
Inputting each topological quantum material data in the sample to be screened into the pre-trained atomic line graph neural network;
and taking the output of the pre-trained atomic line graph neural network as a predicted ion migration barrier corresponding to each topological quantum material data in the sample to be screened.
In some embodiments, the step of screening the sample to be screened for the first time based on the predicted ion migration barrier to obtain the topological quantum material containing the preset metal after the first screening includes:
selecting all topological quantum materials containing preset metals, of which the predicted ion migration potential barriers are smaller than preset energy barriers, corresponding to the topological quantum material data in the sample to be screened as target materials;
and taking the target material as the topological quantum material containing the preset metal after the first screening.
In some embodiments, after the step of selecting, as the target material, all topological quantum materials containing the preset metal, where the predicted ion migration barrier corresponding to the topological quantum material data in the sample to be screened is smaller than the preset energy barrier, the method further includes:
ordering the target materials based on the order of predicting the ion migration barrier from small to large;
Taking the first preset number of topological quantum materials containing preset metals in the sequence as the topological quantum materials containing the preset metals after the first screening.
In some embodiments, the step of performing the second screening on the topological quantum material containing the preset metal after the first screening based on the first principle calculation to obtain an alternative material of the positive electrode of the ion battery containing the preset metal includes:
calculating to obtain the theoretical capacity of each topological quantum material containing the preset metal after first screening based on a first sexual principle;
selecting a topological quantum material containing preset metal, wherein the theoretical capacity obtained through calculation is larger than or equal to the preset theoretical capacity.
In some embodiments, the step of performing the second screening on the topological quantum material containing the preset metal after the first screening based on the first principle calculation to obtain an alternative material of the positive electrode of the ion battery containing the preset metal further includes:
calculating to obtain the stability index of each topological quantum material containing the preset metal after first screening based on a first sexual principle;
selecting the topological quantum material containing the preset metal, wherein the calculated stability index is smaller than or equal to the preset stability index.
In some embodiments, the step of performing the second screening on the topological quantum material containing the preset metal after the first screening based on the first principle calculation to obtain an alternative material of the positive electrode of the ion battery containing the preset metal further includes:
for the topological quantum material containing the preset metal, which simultaneously meets the theoretical capacity more than or equal to the preset theoretical capacity and the stability index less than or equal to the preset stability index, calculating to obtain the reversible capacity based on the first sex principle;
selecting the topological quantum material containing the preset metal, wherein the calculated reversible capacity of the topological quantum material is smaller than or equal to the preset reversible capacity.
In some embodiments, the step of collecting a plurality of topological quantum material data to be predicted containing a preset metal and performing a preprocessing operation to obtain a sample to be screened includes:
and collecting and summarizing topological quantum material data containing preset metals in the first database to obtain the topological quantum material data to be predicted.
In some embodiments, the step of collecting the crystal structure data containing the predetermined metal and the corresponding predetermined metal ion migration barrier data and performing a preprocessing operation to obtain the training sample includes:
crawling preset metal ion migration barrier data from a second database by utilizing a crawler tool;
Extracting a first number from the crawled preset metal ion migration barrier data;
and downloading the corresponding file containing the crystal structure information from the third database by utilizing a crawler tool again based on the first number.
In some embodiments, the first database is a literature library, the preset metal ion migration barrier data in the second database is calculated by an empirical valence method, and the third database is an inorganic crystal structure database.
In some embodiments, the predetermined metal is selected from any one of potassium, sodium, lithium, aluminum, magnesium, calcium.
According to a second aspect of the present invention, there is provided an ion battery cathode material screening apparatus, the apparatus comprising:
the collecting and preprocessing module is used for collecting a plurality of to-be-predicted topological quantum material data containing preset metals and preprocessing the data to obtain samples to be screened;
the prediction module is used for predicting a sample to be screened by utilizing a pre-trained atomic line graph neural network so as to obtain a predicted ion migration barrier corresponding to the sample to be screened;
the first screening module is used for carrying out first screening on the sample to be screened based on the predicted ion migration barrier so as to obtain a topological quantum material containing preset metal after the first screening;
And the second screening module is used for calculating to carry out second screening on the topological quantum material containing the preset metal after the first screening based on the first principle so as to obtain an alternative material of the positive electrode of the ion battery containing the preset metal.
According to a third aspect of the present invention, there is also provided a battery, the positive electrode material of which is obtained by screening by the above-described method.
According to a fourth aspect of the present invention, there is also provided an electronic device including:
at least one processor; and
and the memory stores a computer program which can be run on a processor, and the processor executes the screening method of the anode material of the ion battery when executing the program.
According to a fifth aspect of the present invention, there is also provided a computer readable storage medium storing a computer program which when executed by a processor performs the foregoing ion battery cathode material screening method.
According to the method for screening the anode material of the ion battery, the pre-trained atomic line graph neural network is utilized to predict the sample to be screened to obtain the predicted ion migration barrier, the predicted ion migration barrier is utilized to screen the sample to be screened, the screened sample to be screened is screened again based on the first principle calculation, and finally the candidate material of the anode of the ion battery containing the preset metal is obtained.
In addition, the invention also provides an ion battery anode material screening device, a battery, an electronic device and a computer readable storage medium, which can also achieve the technical effects, and are not repeated here.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are necessary for the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention and that other embodiments may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for screening an anode material of an ion battery according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of a screening method of a battery positive electrode material applied to potassium ions according to an embodiment of the present invention;
FIG. 3 is a graph showing an exemplary partial potassium ion migration barrier data for training according to one embodiment of the present invention;
FIG. 4 is a diagram of an example of topological quantum material data in a document to be predicted, provided in one embodiment of the present invention;
FIG. 5 shows the distribution of the number of atoms in a unit cell of a quantum material containing a potassium topology in a sample to be screened, which is applied to the screening of a potassium ion battery according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of an atomic line graph neural network ALIGNN for converting a crystal structure into a graph according to an embodiment of the present invention;
FIG. 7 is a schematic diagram of an atomic line graph neural network ALIGNN layer according to an embodiment of the present invention;
FIG. 8 is a graph showing the predicted migration barrier distribution of a quantum material containing potassium topology according to one embodiment of the present invention;
fig. 9 is a schematic structural diagram of an apparatus for screening an anode material of an ion battery according to another embodiment of the present invention;
FIG. 10 is an internal block diagram of an electronic device in accordance with another embodiment of the present invention;
fig. 11 is a block diagram of a computer readable storage medium according to another embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the following embodiments of the present invention will be described in further detail with reference to the accompanying drawings.
It should be noted that, in the embodiments of the present invention, all the expressions "first" and "second" are used to distinguish two entities with the same name but different entities or different parameters, and it is noted that the "first" and "second" are only used for convenience of expression, and should not be construed as limiting the embodiments of the present invention, and the following embodiments are not described one by one.
In one embodiment, referring to fig. 1, the present invention provides a method 100 for screening a positive electrode material of an ion battery, specifically, the method includes the following steps:
step 101, collecting a plurality of to-be-predicted topological quantum material data containing preset metals and performing pretreatment operation to obtain a sample to be screened;
in this embodiment, the preset metal refers to a specific metal in the metal ion battery, such as lithium in a lithium ion battery (Li-ion), or potassium in a potassium ion battery (KIBs), the collected data is derived from a database of materials in various existing forms, the preprocessing refers to pre-cleaning data which are obviously unavailable or defective after collecting the data and before using the data, the preprocessing mode can be adopted to reject according to a certain characteristic, certain data can be also adopted to reject according to experience, and the sample to be predicted refers to materials with known topology only.
Step 102, predicting a sample to be screened by utilizing a pre-trained atomic line graph neural network to obtain a predicted ion migration barrier corresponding to the sample to be screened;
in this embodiment, the prediction refers to taking each data in the sample to be predicted as input of the pre-trained atomic line graph neural network in turn, and taking the output result of the pre-trained atomic line graph neural network as the predicted ion migration barrier corresponding to each data.
Step 103, carrying out first screening on the sample to be screened based on the predicted ion migration barrier so as to obtain a topological quantum material containing preset metal after the first screening;
in this embodiment, the first screening means selecting using a specific range or value of the migration barrier, for example, the first screening may be performed by retaining samples with migration barriers within a specific range, or the first screening may be performed by rejecting samples with migration barriers exceeding a specific value.
And 104, performing second screening on the topological quantum material containing the preset metal after the first screening based on the first principle calculation to obtain an alternative material of the positive electrode of the ion battery containing the preset metal.
In this embodiment, the first principle computing (First Principle Calculation) is a computing method based on quantum mechanics theory, and only needs several known parameters including atomic fine structure constant, electron mass and charge quantity, atomic nuclear mass and charge quantity, planck constant and light speed, so that the schrodinger equation can be directly solved after a plurality of approximations according to the principle of interaction between atomic nuclei and electrons and the basic motion rule thereof, and almost all ground state properties of the material can be obtained. The first principle calculation has higher requirements on calculation resources and algorithm development, can provide high-precision results, and provides valuable guidance for developing new materials. Generalized first principles calculations include two broad classes, ab-initio based on Hartree-Fock self-consistent field calculations, and Density Functional Theory (DFT) calculations, where DFT calculations calculate the properties of a material by establishing a functional relationship of the material's charge density with total energy. The first principle of principle calculation can obtain the properties of electron density, energy band structure, theoretical capacity, stability, theoretical reversible capacity and the like of the electrode material, and can be combined according to one or more characteristics to screen the required material. The second screening in this embodiment refers to screening materials by using the calculation result of the first principle, and the second screening method may be performed according to various parameters designed by calculation of the first principle, for example, screening with three indexes of preset theoretical capacity, preset reversible capacity and preset stability.
According to the method for screening the anode material of the ion battery, the pre-trained atomic line graph neural network is utilized to predict the sample to be screened to obtain the predicted ion migration barrier, the predicted ion migration barrier is utilized to screen the sample to be screened, the screened sample to be screened is screened again based on the first principle calculation, and finally the candidate material of the anode of the ion battery containing the preset metal is obtained.
In some embodiments, the method further comprises:
collecting crystal structure data containing preset metals and corresponding preset metal ion migration barrier data, and performing pretreatment operation to obtain training samples; the training sample refers to a set formed by data with known crystal structures and migration barriers of materials.
And training the atomic line graph neural network by using the training sample to obtain a pre-trained atomic line graph neural network.
In this embodiment, during training of the atomic map neural network, the crystal structure data is input as a network, and the migration barrier corresponding to the crystal structure data is expected to be output as a model.
In some embodiments, the preprocessing operation includes screening data according to fractional occupancy structure.
In some embodiments, the foregoing step 101 of collecting a plurality of topological quantum material data to be predicted containing a preset metal and performing a preprocessing operation to obtain a sample to be screened includes:
and removing the data with the fraction occupation structure in the data of the plurality of topological quantum materials to be predicted containing the preset metal to obtain a sample to be screened.
In some embodiments, the step of collecting the crystal structure data containing the predetermined metal and the corresponding predetermined metal ion transfer barrier data and performing a preprocessing operation to obtain the training sample includes:
counting the number distribution of atoms in each topological quantum material unit cell which contains preset metal and is to be predicted in the sample to be screened;
determining a minimum and a maximum screening value for the unit cell atomic number based on the intra-unit atomic number distribution;
and screening the crystal structure data containing the preset metal and the corresponding preset metal ion migration barrier data based on the maximum screening value and the minimum screening value to obtain a training sample.
In some embodiments, the foregoing step of determining a minimum and a maximum screening of unit cell atomic numbers based on the intra-unit cell atomic number distribution comprises:
acquiring the minimum atomic number and the maximum atomic number in the atomic number distribution in a unit cell;
the minimum number of atoms is taken as the minimum screening value, and the maximum number of atoms is taken as the maximum screening value.
In some embodiments, the foregoing step of determining a minimum and a maximum screening of unit cell atomic numbers based on the intra-unit cell atomic number distribution comprises:
acquiring the minimum atomic number and the maximum atomic number in the atomic number distribution in a unit cell;
responding to the minimum atomic number equal to zero or the minimum atomic number minus a preset number less than or equal to zero, and taking zero as a minimum screening value;
responding to the fact that the minimum atomic number is not equal to zero and the difference value between the minimum atomic number and the preset number is taken as a minimum screening value when the preset number is larger than zero and the minimum atomic number is subtracted from the minimum atomic number;
and increasing the maximum atomic number by the preset number to serve as a maximum screening value.
In some embodiments, the step of screening the preset metal-containing crystal structure data and the corresponding preset metal ion migration barrier data based on the maximum screening value and the minimum screening value to obtain a training sample includes:
Counting the number of atoms in each unit cell of the collected crystal structure data containing the preset metal;
and taking the crystal structure data containing the preset metal, of which the atomic number in the unit cell is more than or equal to the minimum screening value and less than or equal to the maximum screening value, and the corresponding preset metal ion migration barrier data as training samples.
In some embodiments, the step of collecting the crystal structure data containing the predetermined metal and the corresponding predetermined metal ion transfer barrier data and performing a preprocessing operation to obtain the training sample includes:
and removing the data with the fractional occupation structure in the crystal structure data containing the preset metal, and taking the obtained crystal structure data containing the preset metal and the corresponding preset metal ion migration barrier data as training samples.
In some embodiments, the foregoing step trains an atomic line graph neural network using the training sample to obtain a pre-trained atomic line graph neural network, comprising:
taking the crystal structure data containing the preset metal in the training sample as the input of the atomic diagram neural network, and taking the corresponding preset metal ion migration barrier data in the training sample as the expected output of the atomic diagram neural network for training;
Evaluating the atomic line graph neural network according to the determination coefficient and the average absolute error until the atomic line graph neural network converges;
and taking the atomic line graph neural network at the time of convergence as the pre-trained atomic line graph neural network.
In some embodiments, the step of evaluating the atomic line map neural network according to the decision coefficient and the mean absolute error until the atomic line map neural network converges includes:
responding to all data in the training sample to complete one training, and respectively calculating a decision coefficient and an average absolute error according to a formula I and a formula II;
formula one;
a second formula;
wherein, in the formula I and the formula IIRepresenting the decision coefficient->Mean absolute error +.>Representing the size of the predicted dataset, +.>Representing the true value of the preset metal ion migration barrier data, < + >>Representing a predicted value of preset metal ion migration barrier data;
and responding to the difference value between the determining coefficient and the first value is smaller than or equal to a first preset threshold value, and the difference value between the average absolute error and zero is smaller than or equal to a second preset threshold value, and confirming convergence of the atomic line graph neural network.
In some embodiments, the step 102 of predicting the sample to be screened by using the pre-trained atomic line graph neural network to obtain a predicted ion migration barrier corresponding to the sample to be screened includes:
Inputting each topological quantum material data in the sample to be screened into the pre-trained atomic line graph neural network;
and taking the output of the pre-trained atomic line graph neural network as a predicted ion migration barrier corresponding to each topological quantum material data in the sample to be screened.
In some embodiments, the step 103 performs first screening on the sample to be screened based on the predicted ion migration barrier to obtain a topological quantum material containing the preset metal after the first screening, including:
selecting all topological quantum materials containing preset metals, of which the predicted ion migration potential barriers are smaller than preset energy barriers, corresponding to the topological quantum material data in the sample to be screened as target materials;
and taking the target material as the topological quantum material containing the preset metal after the first screening.
In some embodiments, the foregoing step selects, as the target material, all topological quantum materials containing a predetermined metal having a predicted ion migration barrier smaller than a predetermined energy barrier corresponding to the topological quantum material data in the sample to be screened, and then further includes:
ordering the target materials based on the order of predicting the ion migration barrier from small to large;
Taking the first preset number of topological quantum materials containing preset metals in the sequence as the topological quantum materials containing the preset metals after the first screening.
In some embodiments, the foregoing step 104 calculates, based on the first principle, that the second screening of the topological quantum material containing the preset metal after the first screening is performed to obtain an alternative material of the positive electrode of the ion battery containing the preset metal, including:
calculating to obtain the theoretical capacity of each topological quantum material containing the preset metal after first screening based on a first sexual principle;
selecting a topological quantum material containing preset metal, wherein the theoretical capacity obtained through calculation is larger than or equal to the preset theoretical capacity.
In some embodiments, the foregoing step 104 calculates, based on the first principle, that the second screening is performed on the topological quantum material containing the preset metal after the first screening to obtain an alternative material of the positive electrode of the ion battery containing the preset metal, and further includes:
calculating to obtain the stability index of each topological quantum material containing the preset metal after first screening based on a first sexual principle;
selecting the topological quantum material containing the preset metal, wherein the calculated stability index is smaller than or equal to the preset stability index.
In some embodiments, in step 104, the calculating based on the first principle performs the second screening on the topological quantum material containing the preset metal after the first screening to obtain an alternative material of the positive electrode of the ion battery containing the preset metal, and further includes:
for the topological quantum material containing the preset metal, which simultaneously meets the theoretical capacity more than or equal to the preset theoretical capacity and the stability index less than or equal to the preset stability index, calculating to obtain the reversible capacity based on the first sex principle;
selecting the topological quantum material containing the preset metal, wherein the calculated reversible capacity of the topological quantum material is smaller than or equal to the preset reversible capacity.
In some embodiments, the foregoing step 101 of collecting a plurality of topological quantum material data to be predicted containing a preset metal and performing a preprocessing operation to obtain a sample to be screened includes:
and collecting and summarizing topological quantum material data containing preset metals in the first database to obtain the topological quantum material data to be predicted.
In some embodiments, the step of collecting the crystal structure data containing the predetermined metal and the corresponding predetermined metal ion transfer barrier data and performing a preprocessing operation to obtain the training sample includes:
crawling preset metal ion migration barrier data from a second database by utilizing a crawler tool;
Extracting a first number from the crawled preset metal ion migration barrier data;
downloading a corresponding file containing crystal structure information from a third database by utilizing a crawler tool again based on the first number;
in some embodiments, the first database is a literature library, the preset metal ion migration barrier data in the second database is calculated by an empirical valence method, and the third database is an inorganic crystal structure database.
In some embodiments, the predetermined metal is selected from any one of potassium, sodium, lithium, aluminum, magnesium, calcium.
In another embodiment, in order to facilitate understanding of the solution of the present invention, the following is a detailed description of an example of application to a potassium ion battery, and referring to fig. 2, the present embodiment provides a method for screening a positive electrode material of a battery for a potassium ion battery, which specifically includes the following four parts:
a first part: data preparation
(1) Acquiring a training data set;
training a model for predicting the potassium ion migration barrier requires constructing a dataset corresponding to the crystal structure data and the ion migration barrier. Potassium ion migration barrier data are derived from databases (reference links http:// e01.Iphy. Ac. Cn/bmd /) provided by the national academy of sciences physical research, and are calculated using empirical valence bond methods. Because the data on the web page can only be queried and cannot be directly downloaded, the data crawling is performed by using a crawler tool python library Selenium. The Selenium can simulate the behavior of a user on a web page, automatically open a browser, input a website, click a button, scroll through a page, etc., and extract the required data from the web page. Finally, 2773 potassium ion migration barrier data of the potassium-containing materials are obtained through writing scripts, and part of the potassium ion migration barrier data are shown in fig. 3. Based on the ICSD number of the obtained migration barrier data, cif file and crystal structure information of all migration barrier data were downloaded from the Inorganic Crystal Structure Database (ICSD) again using the Selenium crawler tool.
(2) Acquiring a prediction data set;
the dataset of topological quantum materials in the sample to be predicted is derived from the annex data in the journal of natural science in 2019 (vergniry, m.g., elcro, l., felser, c.et al A complete catalogue of high-quality topological materials, nature 566, 480-485 (2019), https:// doi.org/10.1038/s 41586-019-0954-4), which provides data as shown in fig. 4. Since these data are provided in the format of PDF files, the data formats are not uniform and the data amount is large, for which this embodiment collects 7385 topological quantum material data contained in the appendix by writing a python script.
(3) Removing the crystal structure data occupied by the existence fraction;
the migration barrier data of 2773 crystal structures obtained through preliminary crawling has atomic fraction occupation, the data of the type are complicated to represent by a machine learning method, and a model trained by a processed data set is found after a few methods for processing fraction occupation are tried, so that performance can be reduced. In addition, the fraction occupancy in the data set of the topological quantum material containing potassium to be predicted is very low, about 3.7%. Therefore, the structural data having fractional occupancy in the migration barrier data set is directly removed in the present embodiment. The same operation is adopted for the data set of the topological quantum material containing potassium, and 178 topological quantum materials containing potassium are obtained after treatment.
(4) Removing the crystal structure data with the atomic number greater than 130 in the unit cell;
when the number of atoms contained in a single unit cell is too large, a crystal diagram generated by the graphic neural network model is large, so that the occupation of a video memory is large during training, the situation of overflow of the video memory is easy to occur, the selection range of the super parameter Batch is limited, and meanwhile, the time for training the model is also increased.
In order to reduce the influence on model training and also to enable the training dataset to cover the interval of the atomic number in a unit cell of the quantum material dataset containing potassium topology to be predicted, firstly counting the atomic number in 178 candidate unit cells of the quantum material containing potassium topology, as shown in fig. 5, the atomic number in all unit cells of the quantum material containing potassium topology is in the interval of 0-120, and the atomic number in most unit cells is below 50. And removing the structure with the atomic number of more than 130 in the migration barrier training set unit cells according to the distribution condition of the atomic number in the unit cells of the potassium-containing topological quantum material, and finally obtaining 1269 migration barrier data serving as training samples.
A second part: model network design
A Graph Neural Network (GNN) is a machine learning algorithm used to process graph data. Many graph neural network models are currently proposed for use in predicting material properties, such as Crystal Graph Convolutional Neural Network (CGCNN), atomic line graph neural network (align), matErials graph network (MEGNet), and modified crystal graph convolutional neural network (iccnn). In this embodiment, an atomic line graph neural network (align) model is adopted, and the align model is developed based on a CGCNN model, which has a good performance in predicting many properties of crystals.
As shown in FIG. 6, the detailed way in which the ALIGNN model converts a crystal structure into a pictorial representation, two diagrams are used to encode the crystal structure: the original atomic map G and the corresponding line map L (G) (reference CHOUDHARY K, DECOST B. Atomistic Line Graph Neural Network for improved materials property predictions [ J ]. Npj Computational Materials, 2021, 7 (1): 185.). Nodes and edges in G represent atoms and bonds, respectively, while nodes and edges in L (G) represent bonds and corners, respectively. Edges in G and nodes in L (G) represent the same entity and share the same feature during network updates. Each node in atomic diagram G is assigned 9 input node features, respectively, cycle number, group number, electronegativity, covalent radius, valence electron, first ionization energy, electron affinity energy, block, and atomic volume, according to its atomic species. The initial edge feature in atomic diagram G is the interatomic bond length developed using Radial Basis Functions (RBFs); the initial edge feature in line graph L (G) is the key angle developed using the radial basis function.
The cif document containing crystal structure information is read, each atom in a unit cell is taken as the center, adjacent atoms in the range of the truncated radius (super parameter: the truncated radius r) are searched, N atoms (super parameter: the maximum neighbor number) nearest to the atoms are found according to the distance sequence between the atoms from the center, then edges of the center atom and the N atoms nearest to the center atom are established, the reaction is the edges between nodes in the graph structure, and after all the atoms in the unit cell find the nearest N neighbors, the atomic graph is established. And generating a corresponding line graph on the basis of the atomic diagram. The vertices in the atomic diagram represent information of the elemental signature including cycle number, group number, electronegativity, covalent radius, valence electrons, first ionization energy, electron affinity, block, and atomic volume. These features are thermally encoded with each element corresponding to a feature vector of length 92, the edges of the atomic map representing the distances between the atoms. The vertices of the line graph correspond to the edges of the atomic map, i.e. the same distance between atoms as the edges of the atomic map, the edges of the line graph representing the angular information of the bond angles represented by two adjacent edges, in such a way that the crystal structure is transformed into a representation of the map.
Referring to fig. 7, the align uses an edge-gated graph convolution method to perform information transfer, and converts 92-dimensional vectors representing vertices of element information into vectors with a length of 64 dimensions through an embedding layer; the scalar quantity of the edge of the atomic diagram representing the key length information and the scalar quantity of the edge of the line diagram of the key angle information are first expanded into a vector of 80 dimensions and a vector of 40 dimensions, respectively, and then converted into a vector of 64 dimensions by the embedding layer. Then, the GCN layer and the ALIGNN layer are respectively sent in, wherein the GCN layer and the ALIGNN layer adopt a gating graph convolution mode, and the difference between the GCN layer and the ALIGNN layer is that the GCN layer only carries out primary graph convolution, and the ALIGNN layer contains both primary graph convolution and line graph convolution. The convolution manner of the align layer is: the information of the edges and the vertexes of the atomic diagram is updated firstly, the updated information of the edges of the atomic diagram is used as the vertex information of the diagram, and then the information of the edges and the vertexes of the diagram is updated. And after the combination of the plurality of GCN layers and the plurality of ALIGNN layers, the vertex information of the atomic diagram is read out in an aggregation way to be migration barrier data to be fitted.
Third section: model evaluation
The model obtained by training must be verified by its accuracy in predicting the properties of the training set. The present embodiment will determine coefficients And mean absolute error>As an index for evaluating the quality of the model. Assume +.>Is used for the prediction set of (1),representing the true value of the preset metal ion migration barrier data, < + >>Representing the predicted value of the preset metal ion migration barrier data, determining the coefficient +.>And the average absolute error are respectively defined as:
/>
wherein the coefficient is determinedThe closer to 1, the average absolute error +.>The closer to 0 indicates better model performance.
Fourth part: screening of topological quantum potassium ion battery anode material
As shown in fig. 2 again, in addition to the migration barrier, the reversible capacity of the battery positive electrode material is also an important factor affecting the overall electrochemical performance of the secondary battery. Firstly, a trained atomic line graph neural network is applied to primarily screen to obtain potassium-containing topological quantum materials with low migration potential barriers, and on the basis, reversible capacity is further calculated through a first sex principle to finely screen the candidate potassium-containing topological quantum materials. The trained ALIGNN model is used for predicting the migration barrier of the potassium-containing topological quantum material, the distribution of the migration barrier is shown in figure 8, and the migration barrier of 21 potassium-containing topological quantum materials in figure 8 is lower than 1eV. The lowest 10 materials are selected, theoretical capacity and reversible capacity are obtained through first sexual principle calculation, and further screening is carried out, and the method comprises the following three steps:
Step one, removing candidate potassium-containing topological quantum materials with lower theoretical capacity. At present, the reversible capacity of most positive electrode materials of potassium ion batteries is below 150 mAh/g, so that candidate potassium-containing topological quantum materials with theoretical capacity lower than 150 mAh/g are directly excluded.
And step two, calculating the stability index energy above hull of the candidate topological quantum material without potassium removal. In general, materials with energy above hull below 0.1 eV/atom are considered metastable. Thus, potassium-containing topology quantum materials with energy above hull higher than 0.1 eV/atom were screened out without potassium removal.
And thirdly, for the candidate potassium-containing topological quantum materials with the potassium content of energy above hull lower than 0.1 eV/atom under the condition of no potassium removal obtained by screening, further calculating the formation energy and energy above hull of different potassium-containing concentrations in the potassium removal process of the candidate materials. The pymatgen was used to construct structures with unequal symmetry at different concentrations, and then the first sexual principle calculation software was used to calculate the theoretical energy for these configurations. And drawing an energy diagram, finding the configuration with the lowest energy at each concentration, calculating energy above hull of the configurations, determining the maximum potassium removal amount according to the standard of energy above hull below 0.1 eV/atom, calculating to obtain reversible capacity, and screening the potassium-containing topological quantum material with the reversible capacity of more than 150 mAh/g.
The battery anode material screening method for the potassium ion battery has the following beneficial technical effects: 2773 potassium ion migration potential energy data are collected from the database, and an atomic line graph neural network (ALIGNN) model capable of predicting potassium ion migration potential barriers is trained according to the data set; 178 potassium-containing topological quantum materials are crawled according to literature materials, and an align model calculated and trained by combining a first sexual principle can accelerate the search and research and development of the potassium ion positive electrode materials, an atomic line diagram neural network model with excellent performance and strong portability is obtained by training, a flow frame for screening the topological quantum potassium ion battery positive electrode materials by calculating the atomic line diagram neural network and the first sexual principle is designed, the potassium ion battery positive electrode materials with low potassium ion migration potential barriers and high reversible capacity are rapidly screened by applying the frame, the material research and design process is accelerated, and the experimental cost and the time cost are reduced.
In some embodiments, referring to fig. 9, the present invention further provides an apparatus 200 for screening an anode material of an ion battery, the apparatus comprising:
the collecting and preprocessing module 201 is configured to collect a plurality of to-be-predicted topological quantum material data containing preset metals and perform preprocessing operation to obtain to-be-screened samples;
The prediction module 202 is configured to predict a sample to be screened by using a pre-trained atomic line graph neural network, so as to obtain a predicted ion migration barrier corresponding to the sample to be screened;
the first screening module 203 is configured to perform a first screening on a sample to be screened based on the predicted ion migration barrier, so as to obtain a topology quantum material containing a preset metal after the first screening;
the second screening module 204 is configured to calculate, based on the first principle, to perform a second screening on the topological quantum material containing the preset metal after the first screening, so as to obtain an alternative material of the positive electrode of the ion battery containing the preset metal.
According to the device for screening the anode material of the ion battery, the pre-trained atomic line graph neural network is utilized to predict the sample to be screened to obtain the predicted ion migration barrier, the predicted ion migration barrier is utilized to screen the sample to be screened, the screened sample to be screened is screened again based on the first principle calculation, and finally the candidate material of the anode of the ion battery containing the preset metal is obtained.
It should be noted that, the specific limitation of the screening device for the positive electrode material of the ion battery may be referred to the limitation of the screening method for the positive electrode material of the ion battery, which is not repeated herein. All or part of each module in the screening device for the anode material of the ion battery can be realized by software, hardware and a combination thereof. The above modules may be embedded in hardware or independent of a processor in the electronic device, or may be stored in software in a memory in the electronic device, so that the processor may call and execute operations corresponding to the above modules.
According to still another aspect of the present invention, there is provided a battery, wherein the positive electrode material of the battery is obtained by screening by the method described in the above embodiment, and the battery may be a potassium ion battery, a sodium ion battery, a lithium ion battery, an aluminum ion battery, a magnesium ion battery, or the like.
According to another aspect of the present invention, there is provided an electronic device, which may be a server, and an internal structure thereof is shown in fig. 10. The electronic device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic device includes a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the electronic device is for storing data. The network interface of the electronic device is used for communicating with an external terminal through a network connection. The computer program when executed by a processor realizes the above screening method of the positive electrode material of the ion battery, specifically, the method comprises the following steps:
Collecting a plurality of topological quantum material data to be predicted containing preset metals and performing pretreatment operation to obtain samples to be screened;
predicting a sample to be screened by utilizing a pre-trained atomic line graph neural network to obtain a predicted ion migration barrier corresponding to the sample to be screened;
performing first screening on the sample to be screened based on the predicted ion migration barrier to obtain a topological quantum material containing the preset metal after the first screening;
and (3) carrying out second screening on the topological quantum material containing the preset metal after the first screening based on the first principle to obtain an alternative material of the anode of the ion battery containing the preset metal.
According to still another aspect of the present invention, a computer readable storage medium is provided, as shown in fig. 11, on which a computer program is stored, the computer program, when executed by a processor, implements the above-mentioned method for screening an anode material of an ion battery, specifically, includes performing the following steps:
collecting a plurality of topological quantum material data to be predicted containing preset metals and performing pretreatment operation to obtain samples to be screened;
predicting a sample to be screened by utilizing a pre-trained atomic line graph neural network to obtain a predicted ion migration barrier corresponding to the sample to be screened;
Performing first screening on the sample to be screened based on the predicted ion migration barrier to obtain a topological quantum material containing the preset metal after the first screening;
and (3) carrying out second screening on the topological quantum material containing the preset metal after the first screening based on the first principle to obtain an alternative material of the anode of the ion battery containing the preset metal.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of protection of the present application is to be determined by the appended claims.

Claims (19)

1. A method for screening an anode material of an ion battery, the method comprising:
collecting crystal structure data containing preset metals and corresponding preset metal ion migration barrier data, and performing pretreatment operation to obtain a training sample, wherein the preset metals are selected from potassium, the training sample is obtained and comprises potassium ion migration barrier data of potassium-containing materials collected from a database, then the structure data with fractional occupation in the migration barrier data set is directly removed, and the structure with the atomic number in unit cells larger than a first threshold in the migration barrier data set is removed to obtain the training sample;
Collecting a plurality of topological quantum material data to be predicted containing preset metals, and performing preprocessing operation to obtain a sample to be screened, wherein obtaining the sample to be screened comprises collecting the topological quantum material data from a literature, and then directly removing structural data with fractional occupation in a potassium-containing topological quantum material data set to obtain the potassium-containing topological quantum material data to be predicted;
predicting a sample to be screened by utilizing a pre-trained atomic line graph neural network to obtain a predicted ion migration barrier corresponding to the sample to be screened;
carrying out first screening on a sample to be screened based on the predicted ion migration barrier to obtain a topological quantum material containing preset metal after the first screening, wherein the obtaining of the topological quantum material containing the preset metal after the first screening comprises obtaining potassium-containing topological quantum material data with a bottom migration barrier;
performing second screening on the topological quantum material containing the preset metal after the first screening based on the first sexual principle calculation to obtain an alternative material of the positive electrode of the ion battery containing the preset metal, wherein the alternative material comprises the following components: excluding candidate potassium-containing topology quantum materials having theoretical capacities below a second threshold; screening and excluding potassium-containing topological quantum materials with stability indexes energy above hull higher than a third threshold under the condition of not removing potassium; for candidate potassium-containing topological quantum materials which are obtained through screening and have theoretical capacity higher than a second threshold value and under the condition that potassium is not removed, energy above hull is lower than a third threshold value, forming energy of different potassium-containing concentrations in the potassium removing process of the candidate potassium-containing topological quantum materials and energy above hull are further calculated, symmetrical unequal structures under different concentrations are constructed by adopting pymatgen, then theoretical energy of the structures is calculated by using first principle calculation software, an energy map is drawn and formed, the structure with the lowest energy under each concentration is found, energy above hull of the structures is calculated, the maximum potassium removing amount is determined according to the standard that energy above hull is lower than the third threshold value, reversible capacity is calculated, and the potassium-containing topological quantum materials with reversible capacity larger than the second threshold value are selected.
2. The method of claim 1, further comprising:
and training an atomic line graph neural network by using the training sample to obtain the pre-trained atomic line graph neural network.
3. The method according to claim 1, wherein the step of collecting the crystal structure data containing the predetermined metal and the corresponding predetermined metal ion migration barrier data and performing a preprocessing operation to obtain the training sample comprises:
counting the number distribution of atoms in each topological quantum material unit cell which contains preset metal and is to be predicted in the sample to be screened;
determining a minimum and a maximum screening value for the unit cell atomic number based on the intra-unit atomic number distribution;
and screening the crystal structure data containing the preset metal and the corresponding preset metal ion migration barrier data based on the maximum screening value and the minimum screening value to obtain a training sample.
4. The method according to claim 3, wherein the step of determining a minimum and a maximum screening value of the number of unit cell atoms based on the distribution of the number of unit cell atoms comprises:
Acquiring the minimum atomic number and the maximum atomic number in the atomic number distribution in a unit cell;
the minimum number of atoms is taken as the minimum screening value, and the maximum number of atoms is taken as the maximum screening value.
5. The method according to claim 3, wherein the step of determining a minimum and a maximum screening value of the number of unit cell atoms based on the distribution of the number of unit cell atoms comprises:
acquiring the minimum atomic number and the maximum atomic number in the atomic number distribution in a unit cell;
responding to the minimum atomic number equal to zero or the minimum atomic number minus a preset number less than or equal to zero, and taking zero as a minimum screening value;
responding to the condition that the minimum atomic number is not equal to zero and the difference value between the minimum atomic number and the preset number is taken as the minimum screening value when the preset number is larger than zero;
and increasing the maximum atomic number by the preset number to serve as the maximum screening value.
6. The method according to claim 3, wherein the step of screening the preset metal-containing crystal structure data and the corresponding preset metal ion migration barrier data based on the maximum screening value and the minimum screening value to obtain training samples comprises:
Counting the number of atoms in each unit cell of the collected crystal structure data containing the preset metal;
and taking the crystal structure data containing the preset metal, of which the atomic number in the unit cell is more than or equal to the minimum screening value and less than or equal to the maximum screening value, and the corresponding preset metal ion migration barrier data as training samples.
7. The method of claim 2, wherein the training the atomic map neural network using the training sample to obtain a pre-trained atomic map neural network comprises:
taking the crystal structure data containing the preset metal in the training sample as the input of the atomic diagram neural network, and taking the corresponding preset metal ion migration barrier data in the training sample as the expected output of the atomic diagram neural network for training;
evaluating the atomic line graph neural network according to the determination coefficient and the average absolute error until the atomic line graph neural network converges;
and taking the atomic line graph neural network at the time of convergence as the pre-trained atomic line graph neural network.
8. The method according to claim 7, wherein the step of evaluating the atomic map neural network based on the determination coefficient and the mean absolute error until the atomic map neural network converges comprises:
Responding to all data in the training sample to complete one training, and respectively calculating a decision coefficient and an average absolute error according to a formula I and a formula II;
formula one;
a second formula;
wherein, in the formula I and the formula IIRepresenting the decision coefficient->Mean absolute error +.>Representing the size of the predicted dataset, +.>Representing the true value of the preset metal ion migration barrier data, < + >>Representing a predicted value of preset metal ion migration barrier data;
and responding to the difference value between the determining coefficient and the first value is smaller than or equal to a first preset threshold value, and the difference value between the average absolute error and zero is smaller than or equal to a second preset threshold value, and confirming convergence of the atomic line graph neural network.
9. The method according to claim 1, wherein the step of predicting a sample to be screened using a pre-trained atomic line graph neural network to obtain a predicted ion migration barrier corresponding to the sample to be screened comprises:
inputting each topological quantum material data in the sample to be screened into the pre-trained atomic line graph neural network;
and taking the output of the pre-trained atomic line graph neural network as a predicted ion migration barrier corresponding to each topological quantum material data in the sample to be screened.
10. The method for screening an anode material of an ion battery according to claim 1, wherein the step of screening the sample to be screened for the first time based on the predicted ion migration barrier to obtain a topological quantum material containing a predetermined metal after the first screening comprises:
selecting all topological quantum materials containing preset metals, of which the predicted ion migration potential barriers are smaller than preset energy barriers, corresponding to the topological quantum material data in the sample to be screened as target materials;
and taking the target material as the topological quantum material containing the preset metal after the first screening.
11. The method for screening an anode material of an ion battery according to claim 10, wherein after the step of selecting, as the target material, all topological quantum materials containing a predetermined metal having a predicted ion migration barrier smaller than a predetermined energy barrier corresponding to the topological quantum material data in the sample to be screened, further comprises:
ordering the target materials based on the order of predicting the ion migration barrier from small to large;
taking the first preset number of topological quantum materials containing preset metals in the sequence as the topological quantum materials containing the preset metals after the first screening.
12. The method for screening an anode material of an ion battery according to claim 2, wherein the step of collecting a plurality of topological quantum material data to be predicted containing a predetermined metal and performing a preprocessing operation to obtain a sample to be screened comprises:
and collecting and summarizing topological quantum material data containing preset metals in the first database to obtain the topological quantum material data to be predicted.
13. The method according to claim 12, wherein the step of collecting the crystal structure data containing the predetermined metal and the corresponding predetermined metal ion migration barrier data and performing the preprocessing operation to obtain the training sample comprises:
crawling preset metal ion migration barrier data from a second database by utilizing a crawler tool;
extracting a first number from the crawled preset metal ion migration barrier data;
and downloading the corresponding file containing the crystal structure information from the third database by utilizing a crawler tool again based on the first number.
14. The method according to claim 13, wherein the first database is a literature library, the preset metal ion migration barrier data in the second database is calculated by an empirical valence method, and the third database is an inorganic crystal structure database.
15. The method according to claim 1, wherein the predetermined metal is selected from any one of sodium, lithium, aluminum, magnesium, and calcium.
16. An ion battery positive electrode material screening apparatus, characterized in that the apparatus comprises:
the system comprises a collection preprocessing module, a storage preprocessing module and a storage preprocessing module, wherein the collection preprocessing module is used for collecting crystal structure data containing preset metals and corresponding preset metal ion migration barrier data and performing preprocessing operation to obtain a training sample, the preset metals are selected from potassium, the obtaining of the training sample comprises collecting potassium ion migration barrier data of potassium-containing materials from a database, then directly removing the structure data with fraction occupation in the migration barrier data set and removing structures with atomic numbers larger than a first threshold value in unit cells in the migration barrier data set to obtain the training sample; the method comprises the steps of collecting a plurality of topological quantum material data to be predicted containing preset metals, and performing preprocessing operation to obtain a sample to be screened, wherein the sample to be screened comprises collecting the topological quantum material data from a literature, and then directly removing structural data with fractional occupation in a potassium-containing topological quantum material data set to obtain the potassium-containing topological quantum material data to be predicted;
The prediction module is used for predicting a sample to be screened by utilizing a pre-trained atomic line graph neural network so as to obtain a predicted ion migration barrier corresponding to the sample to be screened;
the first screening module is used for carrying out first screening on the sample to be screened based on the predicted ion migration barrier so as to obtain a topological quantum material containing the preset metal after the first screening, and the obtained topological quantum material containing the preset metal after the first screening contains potassium-containing topological quantum material data with a bottom migration barrier;
the second screening module is configured to calculate, based on the first principle, to perform a second screening on the topological quantum material containing the preset metal after the first screening, so as to obtain an alternative material of the positive electrode of the ion battery containing the preset metal, where the second screening module includes: excluding candidate potassium-containing topology quantum materials having theoretical capacities below a second threshold; screening and excluding potassium-containing topological quantum materials with stability indexes energy above hull higher than a third threshold under the condition of not removing potassium; for candidate potassium-containing topological quantum materials with theoretical capacity higher than a second threshold and lower than a third threshold under the condition that potassium is not removed from energy above hull, further calculating formation energy of different potassium-containing concentrations in the potassium removing process of the candidate potassium-containing topological quantum materials and energy above hull, constructing structures with unequal symmetry at different concentrations by adopting pymatgen, calculating theoretical energy of the structures by using first principle calculation software, drawing formation energy patterns, finding the structure with the lowest energy at each concentration, calculating energy above hull of the structures, determining the maximum potassium removal amount according to the standard that energy above hull is lower than the third threshold, calculating to obtain reversible capacity, and screening the potassium-containing topological quantum materials with reversible capacity higher than the second threshold.
17. A battery, characterized in that the positive electrode material of the battery is obtained by screening by the method according to any one of claims 1 to 15.
18. An electronic device, comprising:
at least one processor; and
a memory storing a computer program executable in the processor, the processor performing the method of any of claims 1-15 when the program is executed.
19. A computer readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, performs the method of any one of claims 1-15.
CN202310820212.3A 2023-07-05 2023-07-05 Method and device for screening anode material of ion battery, equipment and medium Active CN116564455B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310820212.3A CN116564455B (en) 2023-07-05 2023-07-05 Method and device for screening anode material of ion battery, equipment and medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310820212.3A CN116564455B (en) 2023-07-05 2023-07-05 Method and device for screening anode material of ion battery, equipment and medium

Publications (2)

Publication Number Publication Date
CN116564455A CN116564455A (en) 2023-08-08
CN116564455B true CN116564455B (en) 2023-09-29

Family

ID=87502174

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310820212.3A Active CN116564455B (en) 2023-07-05 2023-07-05 Method and device for screening anode material of ion battery, equipment and medium

Country Status (1)

Country Link
CN (1) CN116564455B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113362914A (en) * 2021-05-18 2021-09-07 沈阳大学 Method for screening stable and efficient organic-inorganic hybrid perovskite type light absorption material
CN116264316A (en) * 2021-12-15 2023-06-16 香港城市大学 Organic lithium-iodine battery based on two-electron transfer mechanism and manufacturing method and application thereof
CN116364212A (en) * 2021-12-24 2023-06-30 中国科学院物理研究所 Screening optimization method and material for machine learning model acceleration fast ion conductor material

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113362914A (en) * 2021-05-18 2021-09-07 沈阳大学 Method for screening stable and efficient organic-inorganic hybrid perovskite type light absorption material
CN116264316A (en) * 2021-12-15 2023-06-16 香港城市大学 Organic lithium-iodine battery based on two-electron transfer mechanism and manufacturing method and application thereof
CN116364212A (en) * 2021-12-24 2023-06-30 中国科学院物理研究所 Screening optimization method and material for machine learning model acceleration fast ion conductor material

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
机器学习在热电材料领域中的应用;盛晔等;《硅酸盐学报》;第第51卷卷(第第2期期);第499-509页 *

Also Published As

Publication number Publication date
CN116564455A (en) 2023-08-08

Similar Documents

Publication Publication Date Title
He et al. High-throughput screening platform for solid electrolytes combining hierarchical ion-transport prediction algorithms
Allam et al. Application of DFT-based machine learning for developing molecular electrode materials in Li-ion batteries
Fischer et al. Predicting crystal structure by merging data mining with quantum mechanics
He et al. CAVD, towards better characterization of void space for ionic transport analysis
WO2024098653A1 (en) Automated sampling and analysis system and method for preparation of lithium hexafluorophosphate
CN111178641B (en) Short-term power load prediction method based on feature extraction and multi-core RSVR (reactive resource reservation Rate) combined model
CN116609668A (en) Lithium ion battery health state and residual life prediction method
CN116564455B (en) Method and device for screening anode material of ion battery, equipment and medium
CN117132591B (en) Battery data processing method and system based on multi-mode information
Zheng et al. Artificial intelligence-driven rechargeable batteries in multiple fields of development and application towards energy storage
Wang et al. IonML: A physically inspired machine learning platform to directed design superionic conductors
Wan et al. Atom substitution of the solid-state electrolyte Li10GeP2S12 for stabilized all-solid-state lithium metal batteries
CN116486938B (en) Method and device for predicting formation of double perovskite compound
CN114239799A (en) Efficient target detection method, device, medium and system
Allen et al. A segregated approach for modeling the electrochemistry in the 3-D microstructure of Li-ion batteries and its acceleration using block preconditioners
CN109616153B (en) Method for identifying key protein by adopting improved HITS algorithm
Duquesnoy et al. Toward high-performance energy and power battery cells with machine learning-based optimization of electrode manufacturing
CN116525038A (en) Optimal synthesis method, system, equipment and medium for hydrogel
Fernandez et al. NMR, x-ray, and Mössbauer results for amorphous Li-Si alloys using density functional tight-binding method
CN116364212A (en) Screening optimization method and material for machine learning model acceleration fast ion conductor material
CN115408929A (en) Data processing method and device for tunnel driving construction prediction
Golov et al. Enhancing first-principles simulations of complex solid-state ion conductors using topological analysis of procrystal electron density
Monareng et al. Machine learning models for predicting density of sodium-ion battery materials
Kench et al. TauFactor 2: A GPU accelerated python tool for microstructural analysis
CN112685690A (en) Large-scale feature selection method based on rough hypercube under cloud platform

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant