A processing system for predicting the electrochemical window of a lithium battery electrolyte

Abstract

An embodiment of the present invention relates to a processing system for predicting the electrochemical window of a lithium battery electrolyte. The system includes: a simulation preprocessing module, a data sharing area, a molecular dynamics simulation module, a property prediction module, a force field parameter optimization module, and an electrochemical window prediction module; the simulation preprocessing module is used to prepare a force field topology file and a three-dimensional structure file; the molecular dynamics simulation module performs simulations according to the force field topology file and the three-dimensional structure file; the property prediction module performs property calculations based on the simulation trajectory file, and when the property calculation results meet the error range, it transmits the simulation trajectory file to the electrochemical window prediction module, otherwise it transmits the property calculation results to the force field parameter optimization module; the force field parameter optimization module optimizes the force field topology file and transmits the optimization results to the molecular dynamics simulation module for re-simulation; the electrochemical window prediction module makes predictions based on the simulation trajectory file. The present invention can improve the analysis efficiency of the electrochemical window.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a processing system for predicting the electrochemical window of a lithium battery electrolyte. Background Art

[0002] A lithium battery consists of four key materials: a positive electrode, a negative electrode, an electrolyte, and a separator. Among them, the role of the electrolyte is to transport lithium ions between the positive and negative electrodes, which is an important guarantee for the lithium battery to obtain high energy density and good safety performance. The electrolyte is generally prepared by mixing high-purity electrolyte lithium salts, solvents, and / or additives and other raw materials in a certain proportion under certain conditions; the corresponding electrochemical windows (ECW) of electrolytes with different raw material components or the same components but different preparation ratios may be different. Conventionally, the electrochemical window of a lithium battery electrolyte is measured by physical experimental methods. However, there are many drawbacks in experimental means, such as long R & D cycle, inability to perform high-throughput operations, high costs for preparation, testing, and characterization, etc. Summary of the Invention

[0003] The purpose of the present invention is to provide a processing system for predicting the electrochemical window of a lithium battery electrolyte in view of the defects of the prior art, including: a simulation preprocessing module, a data sharing area, a molecular dynamics simulation module, a property prediction module, a force field parameter optimization module, and an electrochemical window prediction module; wherein, the simulation preprocessing module is used to prepare a force field topology file and a three-dimensional structure file for the electrolyte formula; the molecular dynamics simulation module performs molecular dynamics simulation according to the force field topology file and the three-dimensional structure file and transmits the simulation trajectory file to the property prediction module at the end of the simulation; the property prediction module calculates the specified property according to the simulation trajectory file and determines whether the error between the property calculation result and the preset expected value meets the preset error range. If it meets, the simulation trajectory file is transmitted to the electrochemical window prediction module for processing. If it does not meet, the property calculation result is transmitted to the force field parameter optimization module for processing; the force field parameter optimization module optimizes the force field parameters of the force field topology file by selecting a property-force field parameter optimization objective function matching the electrolyte formula and transmits the optimized force field topology file to the molecular dynamics simulation module for re-simulation; the electrochemical window prediction module first regards each frame of the simulation trajectory file as the three-dimensional structure of the electrolyte at a time point and performs cluster structure recognition on each three-dimensional structure of the electrolyte to obtain the corresponding cluster structure file, and then uses one or more selected electrochemical window prediction processing units to perform electrochemical window prediction based on each cluster structure file. Through the present invention, the problem that conventional test methods cannot perform high-throughput analysis can be solved, the analysis efficiency of the electrochemical window can be improved, and the artificial experimental cost of the analysis can be reduced.

[0004] To achieve the above object, an embodiment of the present invention provides a processing system for predicting the electrochemical window of a lithium battery electrolyte. The system includes: a simulation preprocessing module, a data sharing area, a molecular dynamics simulation module, a property prediction module, a force field parameter optimization module, and an electrochemical window prediction module; the data sharing area is respectively connected to the simulation preprocessing module, the molecular dynamics simulation module, the property prediction module, the force field parameter optimization module, and the electrochemical window prediction module;

[0005] The simulation preprocessing module is respectively connected to the molecular dynamics simulation module and the force field parameter optimization module; the simulation preprocessing module is configured to store the input first electrolyte formula, first property type, first property expected value, and first prediction method type into the data sharing area; and prepare a force field topology file and a three-dimensional structure file according to the first electrolyte formula to obtain a corresponding first initial topology file and a first electrolyte three-dimensional structure file; and send the first initial topology file and the first electrolyte three-dimensional structure file to the molecular dynamics simulation module, and send the first initial topology file to the force field parameter optimization module;

[0006] The molecular dynamics simulation module is respectively connected to the property prediction module and the force field parameter optimization module; the molecular dynamics simulation module is configured to, when receiving the first initial topology file and the first electrolyte three-dimensional structure file sent by the simulation preprocessing module, set up an electrolyte lithium salt simulation system according to the first electrolyte three-dimensional structure file and the first electrolyte formula in the data sharing area to obtain a corresponding current simulation system, and perform molecular dynamics simulation according to the first initial topology file in the current simulation system to obtain a corresponding first simulation trajectory file and send it to the property prediction module; the molecular dynamics simulation module is further configured to, when receiving a first optimized topology file sent by the force field parameter optimization module, perform molecular dynamics simulation according to the first optimized topology file in the current simulation system to obtain the corresponding first simulation trajectory file and send it to the property prediction module;

[0007] The property prediction module is respectively connected to the force field parameter optimization module and the electrochemical window prediction module; the property prediction module includes a plurality of first property prediction units, and each of the first property prediction units corresponds to a property type; the property prediction module is configured to, when receiving the first simulation trajectory file, obtain the corresponding first property type and the first property expected value from the data sharing area; and the first property prediction unit corresponding to the first property type performs property prediction on the first simulation trajectory file to obtain a corresponding first property prediction result; and calculates a prediction error based on the first property prediction result and the first property expected value to generate a corresponding first error; and identifies whether the first error exceeds a preset first error range; if it exceeds, the first property prediction result is sent to the force field parameter optimization module; if it does not exceed, the first simulation trajectory file is sent to the electrochemical window prediction module;

[0008] The force field parameter optimization module locally pre-sets a plurality of property-force field parameter optimization objective functions, and each of the property-force field parameter optimization objective functions corresponds to a property type; the property-force field parameter optimization objective function is used to estimate the property with the force field parameters of the force field topology file as the independent variable and the corresponding property type as the dependent variable, and iteratively optimize the force field parameters of the force field topology file in the direction of minimizing the error between the property estimated value and the property expected value, and output the force field topology file with the optimized force field parameters as the corresponding optimized topology file; the force field parameter optimization module is configured to save the first initial topology file as the corresponding first force field topology file when receiving it; the force field parameter optimization module is further configured to, when receiving the first property prediction result, use the first property prediction result as the corresponding first property initial value, obtain the corresponding first property type and the first property expected value from the data sharing area, use the property-force field parameter optimization objective function corresponding to the first property type as the corresponding first optimization objective function, and input the first force field topology file, the first property initial value and the first property expected value into the first optimization objective function for force field parameter optimization and send the optimized topology file output by the function as the corresponding first optimized topology file to the molecular dynamics simulation module, and replace the first force field topology file locally saved with the first optimized topology file;

[0009] The electrochemical window prediction module includes a cluster structure extraction unit and multiple first method prediction units, and each first method prediction unit corresponds to a prediction method type; the cluster structure extraction unit is respectively connected to each of the first method prediction units; the cluster structure extraction unit is configured to extract single-frame trajectory cluster structures from the first simulation trajectory file to obtain multiple first cluster structure files; and send each of the first cluster structure files to the first method prediction unit corresponding to the first prediction method type in the data sharing area; the first method prediction unit is configured to predict the electrochemical window based on each of the first cluster structure files and output a corresponding first prediction result.

[0010] Preferably, the first electrolyte formulation includes multiple first electrolyte components; the first electrolyte component includes a first component name, a first component type, a first component SMILES sequence, and a first component ratio; the first component type includes an electrolyte lithium salt type, a solvent type, and an additive type; the first component ratio is the formulation ratio of the current component in the current electrolyte;

[0011] The first property type at least includes density, dielectric constant, diffusion coefficient, conductivity, and transference number;

[0012] The first prediction method type includes multiple first type switches; the switch state of the first type switch includes an on state and an off state; each first type switch corresponds to a prediction method type; the prediction method type at least includes a HOMO / LUMO prediction method, an IP / EA prediction method, a thermodynamic cycle prediction method combined with reorganization energy and solvation energy correction, and a deep learning model prediction method;

[0013] The first simulation trajectory file includes multiple first single-frame trajectory files, and each first single-frame trajectory file corresponds to a sampling time point.

[0014] Preferably, when preparing the corresponding first initial topology file and the three-dimensional electrolyte structure file based on the first electrolyte formula by the simulation preprocessing module, the simulation preprocessing module is specifically configured to generate a corresponding three-dimensional structure file for each first component SMILES sequence in the first electrolyte formula based on a molecular three-dimensional structure generation software; optimize each three-dimensional structure file of the first component by using a configuration optimization tool based on the DFT theory to obtain a corresponding three-dimensional structure file of the second component; perform atomic electrostatic potential fitting based on each three-dimensional structure file of the second component to obtain a corresponding atomic charge file of the first component; generate a three-dimensional electrolyte structure based on all the three-dimensional structure files of the second component and the proportion of the first component by using an electrolyte three-dimensional structure generation software to obtain the corresponding first electrolyte three-dimensional structure file; construct a GAFF2 force field topology file based on the first electrolyte three-dimensional structure file and the atomic charge file of the first component, and use the constructed GAFF2 force field topology file as the corresponding first initial topology file.

[0015] Preferably, when setting up the electrolyte lithium salt simulation system according to the first electrolyte three-dimensional structure file and the first electrolyte formula in the data sharing area to obtain the corresponding current simulation system, the molecular dynamics simulation module is specifically configured to extract the first component name of the first electrolyte component whose first component type is the electrolyte lithium salt type in the first electrolyte formula as the corresponding current electrolyte lithium salt name; use the electrolyte lithium salt simulation system corresponding to the current electrolyte lithium salt name as the corresponding current simulation system; and construct the system structure of the current simulation system based on the first electrolyte three-dimensional structure file.

[0016] Preferably, when extracting the single-frame trajectory cluster structure of the first simulation trajectory file to obtain multiple first cluster structure files, the cluster structure extraction unit is specifically configured to use each first single-frame trajectory file of the first simulation trajectory file as a three-dimensional electrolyte sampling structure file at a sampling time point; and identify the cluster structure of each three-dimensional electrolyte sampling structure file to obtain the corresponding first cluster structure file; the first cluster structure file includes multiple three-dimensional first cluster structures.

[0017] Preferably, the cluster structure extraction unit is specifically configured to obtain the corresponding first prediction method type from the data sharing area when sending each of the first cluster structure files to the first method prediction unit corresponding to the first prediction method type of the data sharing area; traverse each of the first type switches of the first prediction method type; and when traversing, use the currently traversed first type switch as the corresponding current switch; and when the switch state of the current switch is the on state, send each of the first cluster structure files to the first method prediction unit corresponding to the prediction method type corresponding to the current switch.

[0018] An embodiment of the present invention provides a processing system for predicting the electrochemical window of a lithium battery electrolyte, including: a simulation preprocessing module, a data sharing area, a molecular dynamics simulation module, a property prediction module, a force field parameter optimization module, and an electrochemical window prediction module; wherein, the simulation preprocessing module is used to prepare a force field topology file and a three-dimensional structure file for the electrolyte formulation; the molecular dynamics simulation module performs molecular dynamics simulation according to the force field topology file and the three-dimensional structure file and transmits the simulation trajectory file to the property prediction module at the end of the simulation; the property prediction module performs specified property calculations according to the simulation trajectory file and determines whether the error between the property calculation result and the preset expected value meets the preset error range. If it meets, the simulation trajectory file is transmitted to the electrochemical window prediction module for processing. If it does not meet, the property calculation result is transmitted to the force field parameter optimization module for processing; the force field parameter optimization module selects a property-force field parameter optimization objective function matching the electrolyte formulation to optimize the force field parameters of the force field topology file, and transmits the optimized force field topology file to the molecular dynamics simulation module for re-simulation; the electrochemical window prediction module first regards each frame of the simulation trajectory file as the three-dimensional structure of the electrolyte at a time point and performs cluster structure identification on each three-dimensional structure of the electrolyte to obtain the corresponding cluster structure file, and then uses one or more selected electrochemical window prediction processing units to perform electrochemical window prediction based on each cluster structure file. By the present invention, the problem that the conventional test method cannot perform high-throughput analysis is solved, the analysis efficiency of the electrochemical window is improved, and the artificial experiment cost is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a module structure diagram of a processing system for predicting the electrochemical window of a lithium battery electrolyte provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0021] An embodiment of the present invention provides a processing system for predicting the electrochemical window of a lithium battery electrolyte, as Figure 1 shown in the module structure diagram of a processing system for predicting the electrochemical window of a lithium battery electrolyte provided by an embodiment of the present invention. The present processing system 1 mainly includes: a simulation preprocessing module 11, a data sharing area 12, a molecular dynamics simulation module 13, a property prediction module 14, a force field parameter optimization module 15, and an electrochemical window prediction module 16.

[0022] Here, the connection relationships of the various modules of the system in the embodiment of the present invention are: the simulation preprocessing module 11 is respectively connected to the molecular dynamics simulation module 13 and the force field parameter optimization module 15; the data sharing area 12 is respectively connected to the simulation preprocessing module 11, the molecular dynamics simulation module 13, the property prediction module 14, the force field parameter optimization module 15, and the electrochemical window prediction module 16; the molecular dynamics simulation module 13 is respectively connected to the property prediction module 14 and the force field parameter optimization module 15; the property prediction module 14 is respectively connected to the force field parameter optimization module 15 and the electrochemical window prediction module 16.

[0023] (1) Simulation preprocessing module 11:

[0024] The simulation preprocessing module 11 is used to store the input first electrolyte formula, first property type, first property expected value, and first prediction method type into the data sharing area 12; and prepare a force field topology file and a three-dimensional structure file according to the first electrolyte formula to obtain the corresponding first initial topology file and the first electrolyte three-dimensional structure file; and send the first initial topology file and the first electrolyte three-dimensional structure file to the molecular dynamics simulation module 13, and send the first initial topology file to the force field parameter optimization module 15.

[0025] Among them, the first electrolyte formulation includes multiple first electrolyte components; the first electrolyte component includes the first component name, the first component type, the first component SMILES (Simplified molecular input line entry system) sequence, and the proportion of the first component; the first component type includes electrolyte lithium salt type, solvent type, and additive type; the proportion of the first component is the formulation ratio of the current component in the current electrolyte; here, the SMILES sequence is a standard text sequence that clearly describes the molecular structure using ASCII strings, and the first component SMILES sequence is the molecular structure sequence of the electrolyte lithium salt, solvent, or additive component in the first electrolyte formulation. Two-dimensional structure optimization and three-dimensional structure simulation can be performed based on this sequence;

[0026] The first property type includes at least density, dielectric constant, diffusion coefficient, conductivity, and transference number; here, other electrochemical property types can also be added to the first property type according to specific implementation requirements;

[0027] The first prediction method type includes multiple first type switches; the switch state of the first type switch includes the on state and the off state; each first type switch corresponds to a prediction method type; the prediction method type includes at least HOMO / LUMO prediction method, IP / EA prediction method, thermodynamic cycle prediction method combined with reorganization energy and solvation energy correction, and deep learning model prediction method; here, the system of the embodiment of the present invention provides at least four prediction methods for the electrochemical window, namely HOMO / LUMO prediction method, IP / EA prediction method, thermodynamic cycle prediction method, and deep learning model prediction method. Each first type switch corresponds to one prediction method. If only one first type switch in the first prediction method type is in the on state, only one method will be called for prediction. If two or more first type switches are in the on state, multiple methods will be called for synchronous prediction.

[0028] In addition, in a specific implementation manner of the embodiment of the present invention, the simulation preprocessing module 11 is specifically used for preparing the force field topology file and the three-dimensional structure file according to the first electrolyte formulation to obtain the corresponding first initial topology file and the first electrolyte three-dimensional structure file:

[0029] Step A1, generating a corresponding first component three-dimensional structure file for each first component SMILES sequence in the first electrolyte formulation based on the molecular three-dimensional structure generation software;

[0030] Here, the molecular three-dimensional structure generation software used in the system of the embodiment of the present invention is a processing software that can perform three-dimensional structure simulation on the molecule based on the molecular SMILES sequence, such as Packmol software;

[0031] Step A2, and use a configuration optimization tool based on the DFT theory to optimize each three-dimensional structure file of the first component to obtain a corresponding three-dimensional structure file of the second component;

[0032] Here, the system of the embodiment of the present invention will use a structure optimization tool based on the Density Functional Theory (DFT) to optimize the three-dimensional molecular structure of each component in the first electrolyte formulation, that is, the three-dimensional structure file of the first component. The working principle of this structure optimization tool is as follows: Based on the DFT theory, an energy calculation function for calculating the energy of the current molecular structure is constructed, and a target function with the atomic structure parameters in the current molecular structure as independent variables and the energy of the current molecular structure as the dependent variable is constructed in the direction of making the function value of this energy calculation function reach the minimum energy threshold. Then, by solving this target function, the optimal set of atomic structure parameters of the current molecular structure is determined, and the atomic structure parameters of each atom in the current molecular structure, that is, the three-dimensional structure file of the first component, are adjusted using this optimal set of atomic structure parameters to obtain a corresponding optimized structure. Then, the optimized structure is converted into a three-dimensional structure file in the three-dimensional structure file format matching the current structure optimization tool, and the converted three-dimensional structure file is output as the corresponding three-dimensional structure file of the second component; the energy calculation function mentioned here is usually selected as the eigenenergy calculation function of the Schrödinger equation. Of course, other types of energy calculation functions can also be selected based on specific implementation requirements; constructing an energy calculation function based on the DFT theory belongs to the public technical knowledge in the field of electrochemistry and will not be further elaborated here; it should be noted that in the system of the embodiment of the present invention, the above-mentioned configuration optimization tool based on the DFT theory can be replaced by similar processing modules / interfaces / functions of other mature quantum chemistry software on the market, such as Gaussian software, ORCA software, etc.;

[0033] Step A3, and perform atomic electrostatic potential fitting based on each three-dimensional structure file of the second component to obtain a corresponding atomic charge file of the first component;

[0034] Here, the atomic attribute information of each atom in the three-dimensional structure file of the second component does not have charge information. Therefore, in the current step, the system of the embodiment of the present invention will obtain the atomic charge information of each atom, that is, the electrostatic potential charge (electrostatic potential change) information, based on the electrostatic potential fitting method; the electrostatic potential fitting method is a publicly known method for calculating electrostatic potential charges. Briefly speaking, it is to construct a wave function based on the three-dimensional structure file of the second component, and then perform electrostatic potential fitting calculation by the wave function to obtain the corresponding atomic charge file of the first component. Further elaboration will not be made here; it should be noted that when the system of the embodiment of the present invention performs atomic electrostatic potential fitting, in addition to the working components developed based on the conventional electrostatic potential fitting method, similar processing modules / interfaces / functions of other mature quantum chemistry software on the market can also be used for replacement, such as Multiwfn software, etc.;

[0035] Step A4, and use the electrolyte three-dimensional structure generation software to generate the corresponding first electrolyte three-dimensional structure file according to all the three-dimensional structure files of the second component and the proportion of the first component;

[0036] Here, the electrolyte three-dimensional structure generation software used by the system of the embodiment of the present invention is a processing software that can perform three-dimensional structure simulation on a compound based on the composition ratio of the compound and the three-dimensional structures of its components, such as Packmol software;

[0037] Step A5, and construct the GAFF2 force field topology file according to the first electrolyte three-dimensional structure file and the first component atomic charge file, and use the constructed GAFF2 force field topology file as the corresponding first initial topology file;

[0038] Here, the force field topology construction tool used in the system of the embodiment of the present invention is a tool that takes the input three-dimensional compound structure as the current ligand, applies the specified force field to the current ligand to output the corresponding ligand process file, and performs the conversion of the specified format topology file according to the obtained ligand process file; this tool can be composed of one or more common tools, and one combination method is composed of the common AmberTools and ACPYPE; it should be noted that when the force field topology construction tool is composed of AmberTools and ACPYPE, the processing process of the system of the embodiment of the present invention using the force field topology construction tool to construct the GAFF2 force field topology file according to the first electrolyte three-dimensional structure file is as follows: AmberTools takes the first electrolyte three-dimensional structure file as the current ligand and applies it to the current ligand with the GAFF2 force field as the specified force field to output the corresponding ligand process file (prmtop file and inpcrd file), and ACPYPE takes the GROMACS topology file as the specified format topology file and performs the conversion of the specified format topology file according to the obtained ligand process file to obtain a GAFF2 force field topology file that meets the GROMACS topology file format, that is, the first initial topology file.

[0039] (II) Data sharing area 12:

[0040] The data sharing area 12 is used to store shared data, including the first electrolyte formula, the first property type, the first property expected value, and the first prediction method type.

[0041] Here, in addition to being used to store the above multiple shared data, the data sharing area 12 of the system of the embodiment of the present invention can also store the input / output data of each module and the module processing process data based on actual application requirements.

[0042] (III) Molecular dynamics simulation module 13:

[0043] The molecular dynamics simulation module 13 is used to set up the electrolyte lithium salt simulation system according to the first electrolyte three-dimensional structure file and the first electrolyte formula in the data sharing area 12 to obtain the corresponding current simulation system when receiving the first initial topology file and the first electrolyte three-dimensional structure file sent by the simulation preprocessing module 11, and perform molecular dynamics simulation according to the first initial topology file under the current simulation system to obtain the corresponding first simulation trajectory file and send it to the property prediction module 14;

[0044] Among them, the first simulation trajectory file includes multiple first single-frame trajectory files, and each first single-frame trajectory file corresponds to a sampling time point.

[0045] Here, the molecular dynamics simulation module 13 of the system according to the embodiments of the present invention realizes molecular dynamics simulation by calling common molecular dynamics simulation software, such as GROMACS software, LAMMPS software, OPENMM software, etc.; the output of the molecular dynamics simulation software is all output in the form of a trajectory file, and this trajectory file, that is, the first simulation trajectory file, is composed of compound structure snapshots collected at multiple sampling time points, that is, the first single-frame trajectory files; each first single-frame trajectory file is actually a structure snapshot of a three-dimensional electrolyte structure at a certain sampling time point during the current electrolyte simulation process.

[0046] In another specific implementation manner of the embodiments of the present invention, the molecular dynamics simulation module 13 is specifically configured to, when setting the electrolyte lithium salt simulation system according to the first three-dimensional electrolyte structure file and the first electrolyte formula in the data sharing area 12 to obtain the corresponding current simulation system, extract the first component name of the first electrolyte component whose first component type is the electrolyte lithium salt type in the first electrolyte formula as the corresponding current electrolyte lithium salt name; and use the electrolyte lithium salt simulation system corresponding to the current electrolyte lithium salt name as the corresponding current simulation system; and construct the system structure of the current simulation system based on the first three-dimensional electrolyte structure file.

[0047] Here, the molecular dynamics simulation software of the system according to the embodiments of the present invention can provide multiple electrolyte lithium salt simulation systems for selection when simulating compounds such as electrolytes. The system according to the embodiments of the present invention needs to confirm the electrolyte lithium salt simulation system corresponding to the current simulation according to the name of the electrolyte lithium salt in the current electrolyte formula, that is, the first component name of the first electrolyte component whose first component type is the electrolyte lithium salt type in the first electrolyte formula, and record it as the current simulation system. After confirming the current simulation system, the corresponding system simulation structure can be set by the first three-dimensional electrolyte structure file.

[0048] The molecular dynamics simulation module 13 is further configured to, when receiving the first optimized topology file sent by the force field parameter optimization module 15, perform molecular dynamics simulation according to the first optimized topology file in the current simulation system to obtain the corresponding first simulation trajectory file and send it to the property prediction module 14.

[0049] Here, as will be shown in the following content, the property prediction module 14 of the system according to the embodiments of the present invention will screen the first simulation trajectory file output by the molecular dynamics simulation module 13 based on the current topology file simulation (the first initial topology file or the previous first optimized topology file), and when the screening is unqualified, the force field parameter optimization module 15 will optimize the force field parameters of the current topology file and send the optimized topology file, that is, the next first optimized topology file, back to the molecular dynamics simulation module 13; after receiving the first optimized topology file sent back by the force field parameter optimization module 15, the molecular dynamics simulation module 13 will perform a new molecular dynamics simulation with this file as the current topology file simulation.

[0050] It should be noted that the molecular dynamics simulation module 13 of the system according to the embodiments of the present invention has multiple trajectory file output methods when outputting the first simulation trajectory file;

[0051] One of the output methods is: preset a total number of simulation iteration steps N1, and at the end of each iteration, sample the three-dimensional structure of the electrolyte at the current moment and use the sampled three-dimensional structure snapshot as a corresponding first single-frame trajectory file F i , and when the number of iterations reaches the total number of simulation iteration steps N, the N first single-frame trajectory files F i are combined to form the corresponding first simulation trajectory file for output; where, 1≤N1, 1≤i≤N1;

[0052] Another output method is: preset a maximum number of simulation iteration steps M1, and at the end of each iteration, sample the three-dimensional structure of the electrolyte at the current moment and use the sampled three-dimensional structure snapshot as a corresponding first single-frame trajectory file F j , and use the first single-frame trajectory file F j obtained in the current iteration as the corresponding first simulation trajectory file for output, and stop the simulation when the number of iterations is greater than the maximum number of simulation iteration steps M; where, 1≤M1, 1≤j≤M1;

[0053] Another output method is: preset a total number of simulation iteration steps N2 and / or a starting sampling iteration step N0 and / or a starting sampling condition (such as a preset one or more types of electrochemical property threshold conditions), and after the number of iterations reaches the starting sampling iteration step N0 and / or the electrochemical property of the simulation structure meets the preset starting sampling condition, sample the three-dimensional structure of the electrolyte at the current moment at the end of each iteration and use the sampled three-dimensional structure snapshot as a corresponding first single-frame trajectory file F k , and when the number of iterations reaches the total number of simulation iteration steps N2, all the first single-frame trajectory files F k obtained are combined to form the corresponding first simulation trajectory file for output; where, 1≤N0≤N2, 1≤k;

[0054] Another output method is as follows: preset a maximum number of simulation iteration steps M2 and / or a starting sampling iteration step M0 and / or a starting sampling condition (such as a preset threshold condition for one or more types of electrochemical properties), and after the number of iterations reaches the starting sampling iteration step M0 and / or the electrochemical properties of the simulated structure meet the preset starting sampling condition, sample the three-dimensional structure of the electrolyte at the end of each iteration and use the sampled three-dimensional structure snapshot as a corresponding first single-frame trajectory file F h , and use the first single-frame trajectory file F obtained in that instance h as the corresponding first simulation trajectory file for output, and stop the simulation when the number of iterations is greater than the maximum number of simulation iteration steps M2; where 1 ≤ M0 ≤ M2, 1 ≤ h;

[0055] Here, the computing power resources and simulation accuracies occupied by the above several methods supported by the molecular dynamics simulation module 13 of the system in the embodiments of the present invention are different, and users can select the output method based on actual application requirements and resource conditions.

[0056] (4) Property prediction module 14:

[0057] The property prediction module 14 includes a plurality of first property prediction units 141, and each first property prediction unit 141 corresponds to a property type. The property types mentioned here at least include density, dielectric constant, diffusion coefficient, conductivity, and transference number. Additionally, other types of electrochemical properties can be added according to specific implementation requirements.

[0058] Here, when calculating the corresponding electrochemical properties, each first property prediction unit 141 of the system in the embodiments of the present invention defaults to using an electrochemical property calculation method based on the DFT theory for corresponding property calculations, and can also use an electrochemical property calculation method based on other theories for calculations based on actual application requirements.

[0059] The property prediction module 14 is used to obtain the corresponding first property type and first property expected value from the data sharing area 12 when receiving the first simulation trajectory file; and the first property prediction unit 141 corresponding to the first property type performs property prediction on the first simulation trajectory file to obtain the corresponding first property prediction result; and calculates the prediction error based on the first property prediction result and the first property expected value to generate the corresponding first error; and identifies whether the first error exceeds the preset first error range; if it exceeds, the first property prediction result is sent to the force field parameter optimization module 15; if it does not exceed, the first simulation trajectory file is sent to the electrochemical window prediction module 16.

[0060] Here, the first error range is a pre-set error value range. The first error range may only include a pre-set error threshold, or may be a value range composed of a pair of pre-set minimum and maximum error thresholds. If the first error range only includes a pre-set error threshold, then when identifying whether the first error exceeds the pre-set first error range, if the first error exceeds the pre-set error threshold, it is confirmed that the first error exceeds the first error range; conversely, if the first error does not exceed the pre-set error threshold, it is confirmed that the first error does not exceed the first error range. If the first error range is a value range composed of a pair of pre-set minimum and maximum error thresholds, then when identifying whether the first error exceeds the pre-set first error range, if the first error is less than the minimum error threshold of the first error range or greater than the maximum error threshold of the first error range, it is confirmed that the first error exceeds the first error range; conversely, if the first error is greater than or equal to the minimum error threshold of the first error range and less than or equal to the maximum error threshold of the first error range, it is confirmed that the first error does not exceed the first error range.

[0061] In another specific implementation manner of the embodiment of the present invention, the property prediction module 14 is specifically configured to, when the first property prediction unit 141 corresponding to the first property type performs property prediction on the first simulation trajectory file to obtain a corresponding first property prediction result:

[0062] Step B1, if the first property type is density, call the first property prediction unit 141 corresponding to density to calculate the density of the electrolyte according to the three-dimensional structure given by each first single-frame trajectory file of the first simulation trajectory file to obtain a corresponding first density; and perform weighted average calculation on all the obtained first densities according to a pre-set weighted average calculation method and output the calculation result as the corresponding first property prediction result.

[0063] Here, the pre-set weighted average calculation method of the system in the embodiment of the present invention supports a variety of weighted average calculation methods, and one of them is: X is the total number of the first single-frame trajectory files, a is the density index, 1 ≤ a ≤ X, ρ a is the first density, w a is the density weight parameter, w a-1 ≤ w a ;

[0064] Step B2, if the first property type is the dielectric constant, call the first property prediction unit 141 corresponding to the dielectric constant to calculate the dielectric constant of the electrolyte according to the three-dimensional structures given by the respective first single-frame trajectory files of the first simulation trajectory file, obtaining the corresponding first dielectric constant; and perform a weighted average calculation on all the obtained first dielectric constants according to a preset weighted average calculation method, and output the calculation result as the corresponding first property prediction result;

[0065] The weighted average calculation method used here is similar to the weighted average calculation method used in Step B1;

[0066] Step B3, if the first property type is the diffusion coefficient, call the first property prediction unit 141 corresponding to the diffusion coefficient to calculate the diffusion coefficient of the electrolyte according to the three-dimensional structures given by the respective first single-frame trajectory files of the first simulation trajectory file, obtaining the corresponding first diffusion coefficient; and perform a weighted average calculation on all the obtained first diffusion coefficients according to a preset weighted average calculation method, and output the calculation result as the corresponding first property prediction result;

[0067] The weighted average calculation method used here is similar to the weighted average calculation method used in Step B1;

[0068] Step B4, if the first property type is the conductivity, call the first property prediction unit 141 corresponding to the conductivity to calculate the conductivity of the electrolyte according to the three-dimensional structures given by the respective first single-frame trajectory files of the first simulation trajectory file, obtaining the corresponding first conductivity; and perform a weighted average calculation on all the obtained first conductivities according to a preset weighted average calculation method, and output the calculation result as the corresponding first property prediction result;

[0069] The weighted average calculation method used here is similar to the weighted average calculation method used in Step B1;

[0070] Step B5, if the first property type is the transference number, call the first property prediction unit 141 corresponding to the transference number to calculate the transference number of the electrolyte according to the three-dimensional structures given by the respective first single-frame trajectory files of the first simulation trajectory file, obtaining the corresponding first transference number; and perform a weighted average calculation on all the obtained first transference numbers according to a preset weighted average calculation method, and output the calculation result as the corresponding first property prediction result;

[0071] The weighted average calculation method used here is similar to the weighted average calculation method used in Step B1.

[0072] In another specific implementation manner of the embodiment of the present invention, the property prediction module 14 is specifically configured to, when calculating a prediction error according to the first property prediction result and the first property expected value to generate a corresponding first error, use the absolute difference between the first property prediction result and the first property expected value as the corresponding first error.

[0073] (5) Force field parameter optimization module 15:

[0074] The force field parameter optimization module 15 locally pre-sets a plurality of property-force field parameter optimization objective functions, and each property-force field parameter optimization objective function corresponds to a property type.

[0075] Here, the property-force field parameter optimization objective function of the system in the embodiment of the present invention is used to estimate a property with the force field parameters of the force field topology file as independent variables and the corresponding property type as the dependent variable, and iteratively optimize the force field parameters of the force field topology file in the direction of minimizing the error between the property estimated value and the property expected value, and output the force field topology file with the optimized force field parameters as the corresponding optimized topology file.

[0076] The force field parameter optimization module 15 is configured to save the first initial topology file as the corresponding first force field topology file when receiving it.

[0077] Here, when the force field parameter optimization module 15 of the system in the embodiment of the present invention obtains the first initial topology file from the simulation preprocessing module 11, it will save it as the target force field topology file for the next force field parameter optimization, that is, the first force field topology file, but will not perform optimization processing on it at this time.

[0078] The force field parameter optimization module 15 is further configured to, when receiving the first property prediction result, use the first property prediction result as the corresponding first property initial value, obtain the corresponding first property type and the first property expected value from the data sharing area 12, use the property-force field parameter optimization objective function corresponding to the first property type as the corresponding first optimization objective function, input the first force field topology file, the first property initial value and the first property expected value into the first optimization objective function for force field parameter optimization, and send the optimized topology file output by the function as the corresponding first optimized topology file to the molecular dynamics simulation module 13, and replace the first force field topology file saved locally with the first optimized topology file.

[0079] Here, the force field parameter optimization module 15 of the system according to the embodiment of the present invention will perform a force field parameter optimization on the locally saved first force field topology file only when receiving the first property prediction result sent by the property prediction module 14 each time, and send the optimization result, i.e., the first optimized topology file, to the molecular dynamics simulation module 13. The molecular dynamics simulation module 13 will re - simulate based on the first optimized topology file, and use the locally obtained first force field topology file as the target force field topology file for the next force field parameter optimization, that is, replace the locally stored first force field topology file with the first force field topology file obtained this time.

[0080] (6) Electrochemical window prediction module 16:

[0081] The electrochemical window prediction module 16 includes a cluster structure extraction unit 161 and a plurality of first method prediction units 162. Each first method prediction unit 162 corresponds to a prediction method type; the cluster structure extraction unit 161 is respectively connected to the property prediction module 14 and each first method prediction unit 162.

[0082] The cluster structure extraction unit 161 is used to extract the single - frame trajectory cluster structure of the first simulation trajectory file to obtain a plurality of first cluster structure files; and send each first cluster structure file to the first method prediction unit 162 corresponding to the first prediction method type in the data sharing area 12.

[0083] In another specific implementation manner of the embodiment of the present invention, when the cluster structure extraction unit 161 extracts the single - frame trajectory cluster structure of the first simulation trajectory file to obtain a plurality of first cluster structure files: each first single - frame trajectory file of the first simulation trajectory file is used as a three - dimensional electrolyte sampling structure file at a sampling time point; and the cluster structure of each three - dimensional electrolyte sampling structure file is identified to obtain the corresponding first cluster structure file; wherein, the first cluster structure file includes a plurality of three - dimensional first cluster structures.

[0084] In another specific implementation manner of the embodiment of the present invention, when the cluster structure extraction unit 161 sends each first cluster structure file to the first method prediction unit 162 corresponding to the first prediction method type in the data sharing area 12, it obtains the corresponding first prediction method type from the data sharing area 12; and traverses each first type switch of the first prediction method type; and during the traversal, the currently traversed first type switch is used as the corresponding current switch; and when the switch state of the current switch is the on state, each first cluster structure file is sent to the first method prediction unit 162 corresponding to the prediction method type corresponding to the current switch.

[0085] Here, the electrochemical window prediction module 16 of the system according to the embodiments of the present invention can use one or more methods to predict the electrochemical window according to the switching states of the respective first type switches in the first prediction method type and output their respective prediction results.

[0086] The first method prediction unit 162 is used to predict the electrochemical window according to each first cluster structure file and output the corresponding first prediction result.

[0087] Here, the electrochemical window prediction module 16 of the system according to the embodiments of the present invention supports at least four prediction method types: HOMO / LUMO prediction method, IP / EA prediction method, thermodynamic cycle prediction method combined with reorganization energy and solvation energy correction, and deep learning model prediction method; that is, it includes at least 4 first method prediction units 162.

[0088] The prediction principles of these 4 first method prediction units 162 are briefly introduced as follows:

[0089] 1) The first method prediction unit 162 corresponding to the HOMO / LUMO prediction method predicts the electrochemical window of the input first cluster structure file based on the highest occupied molecular orbital (HUMO) and the lowest unoccupied molecular orbital (LUMO). The principle of this method can be understood by referring to the technical article "Lithium insertion compounds [M]. MRS Online Proceeding" and related technical articles derived therefrom, and will not be elaborated here in detail;

[0090] 2) The first method prediction unit 162 corresponding to the IP / EA prediction method uses the Electron Affinity (EA) / Ionization Potential (IP) to replace the Highest Occupied Molecular Orbital (HUMO) / Lowest Unoccupied Molecular Orbital (LUMO) to predict the electrochemical window of the input first cluster structure file. This method includes two calculation methods: the calculation method based on Vertical Electron Affinity (VEA) / Vertical Ionization Potential (VIP), and the calculation method based on Adiabatic Electron Affinity (AEA) / Adiabatic Ionization Potential (AIP). The principle of this method can be understood by referring to the technical article "Electrochemical Windows of Room-Temperature Ionic Liquids from Molecular Dynamics and Density Functional Theory Calculations" and related technical articles derived therefrom, and will not be elaborated in detail here;

[0091] 3) The first method prediction unit 162 corresponding to the thermodynamic cycle prediction method combined with the reorganization energy and solvation energy correction predicts the electrochemical window of the first cluster structure file input in the thermodynamic cycle iteration manner, and uses the reorganization energy parameter λ IP / EA and the solvation energy parameter ΔG sol to correct the prediction result. The principle of this method can be understood by referring to the technical articles "A Thermodynamic Cycle-Based Electrochemical Windows Database of 308 Electrolyte Solvents for Rechargeable Batteries", "The solvation structure, transport properties and reduction behavior of carbonate-based electrolytes of lithium-ion batteries[J]" and related technical articles, and will not be elaborated in detail here;

[0092] 4) The first method prediction unit 162 corresponding to the deep learning model prediction method pre-constructs a Machine Learning Potential (MLP) function with a machine learning model (convolutional neural network or multi-layer perceptron neural network), and uses the machine learning potential function to perform machine learning molecular dynamics (MLMD) simulation on the input first cluster structure file to obtain the corresponding MLMD trajectory file. The ab initio molecular dynamics (AIMD) combined with free energy perturbation (FEP) and thermodynamic integration (TI) method is used to predict the electrochemical window of the MLMD trajectory file. The principle of this method can be understood by referring to the technical article "Switching of Redox Levels Leads to High Reductive Stability in Water-in-Salt Electrolytes" and related technical articles derived therefrom, which will not be elaborated here.

[0093] It should be noted that it should be understood that the division of each module of the above system is only a logical function division. In actual implementation, it can be fully or partially integrated into a physical entity, or physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also all be implemented in hardware; they can also be partially implemented in the form of software called by processing elements and partially implemented in hardware. For example, the simulation preprocessing module can be a separately established processing element, or can be integrated in a certain chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a certain processing element of the above system to perform the functions of the above determined module. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together or independently implemented. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each method step of the foregoing method or each module processing step of the foregoing system can be completed by the integrated logic circuit in the processor element or the instruction in software form.

[0094] For example, these modules of the above system can be one or more integrated circuits configured to implement the foregoing method. For example: one or more Application Specific Integrated Circuits (ASICs), or one or more Digital Signal Processors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs), etc. Again, when a certain module of the above system is implemented in the form of a processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processors that can call program code. Again, these modules can be integrated together and implemented in the form of a System-on-a-chip (SOC).

[0095] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the foregoing method embodiments are generated in whole or in part. The above computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The above computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the above computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (such as coaxial cable, optical fiber, Digital Subscriber Line (DSL)) or wireless (such as infrared, wireless, Bluetooth, microwave, etc.) means. The above computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more integrated available media. The above available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)), etc.

[0096] An embodiment of the present invention provides a processing system for predicting the electrochemical window of a lithium battery electrolyte, including: a simulation preprocessing module, a data sharing area, a molecular dynamics simulation module, a property prediction module, a force field parameter optimization module, and an electrochemical window prediction module; wherein, the simulation preprocessing module is used to prepare a force field topology file and a three-dimensional structure file for the electrolyte formulation; the molecular dynamics simulation module performs molecular dynamics simulation according to the force field topology file and the three-dimensional structure file, and transmits the simulation trajectory file to the property prediction module at the end of the simulation; the property prediction module calculates the specified property according to the simulation trajectory file, and judges whether the error between the property calculation result and the preset expected value meets the preset error range. If it meets, the simulation trajectory file is transmitted to the electrochemical window prediction module for processing. If it does not meet, the property calculation result is transmitted to the force field parameter optimization module for processing; the force field parameter optimization module selects a property-force field parameter optimization objective function matching the electrolyte formulation to optimize the force field parameters of the force field topology file, and transmits the optimized force field topology file to the molecular dynamics simulation module for re-simulation; the electrochemical window prediction module first regards each frame of the simulation trajectory file as the three-dimensional structure of the electrolyte at a time point, identifies the cluster structure for each three-dimensional structure of the electrolyte to obtain the corresponding cluster structure file, and then uses one or more selected electrochemical window prediction processing units to perform electrochemical window prediction based on each cluster structure file. By the present invention, the problem that the conventional test method cannot perform high-throughput analysis is solved, the analysis efficiency of the electrochemical window is improved, and the artificial experiment cost is reduced.

[0097] Those skilled in the art should also be able to further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of the present invention.

[0098] The steps of the methods or algorithms described in combination with the embodiments disclosed herein can be implemented by hardware, software modules executed by a processor, or a combination of the two. The software modules can be placed in a random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium well-known in the technical field.

[0099] The specific embodiments described above have further elaborated on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A processing system for predicting the electrochemical window of a lithium battery electrolyte, characterized in that The system includes: an analog preprocessing module, a data sharing area, a molecular dynamics simulation module, a property prediction module, a force field parameter optimization module, and an electrochemical window prediction module; the data sharing area is respectively connected to the analog preprocessing module, the molecular dynamics simulation module, the property prediction module, the force field parameter optimization module, and the electrochemical window prediction module; The analog preprocessing module is respectively connected to the molecular dynamics simulation module and the force field parameter optimization module; the analog preprocessing module is used to store the input first electrolyte formulation, first property type, first property expected value, and first prediction method type into the data sharing area; and prepare a force field topology file and a three-dimensional structure file according to the first electrolyte formulation to obtain a corresponding first initial topology file and a first electrolyte three-dimensional structure file; and send the first initial topology file and the first electrolyte three-dimensional structure file to the molecular dynamics simulation module, and send the first initial topology file to the force field parameter optimization module; The molecular dynamics simulation module is respectively connected to the property prediction module and the force field parameter optimization module; the molecular dynamics simulation module is used to, when receiving the first initial topology file and the first electrolyte three-dimensional structure file sent by the analog preprocessing module, set the electrolyte lithium salt simulation system according to the first electrolyte three-dimensional structure file and the first electrolyte formulation in the data sharing area to obtain a corresponding current simulation system, and perform molecular dynamics simulation according to the first initial topology file in the current simulation system to obtain a corresponding first simulation trajectory file and send it to the property prediction module; the molecular dynamics simulation module is also used to, when receiving the first optimized topology file sent by the force field parameter optimization module, perform molecular dynamics simulation according to the first optimized topology file in the current simulation system to obtain the corresponding first simulation trajectory file and send it to the property prediction module; The property prediction module is respectively connected to the force field parameter optimization module and the electrochemical window prediction module; the property prediction module includes a plurality of first property prediction units, and each first property prediction unit corresponds to a property type; the property prediction module is used to, when receiving the first simulation trajectory file, obtain the corresponding first property type and the first property expected value from the data sharing area; and perform property prediction on the first simulation trajectory file by the first property prediction unit corresponding to the first property type to obtain a corresponding first property prediction result; and calculate a prediction error according to the first property prediction result and the first property expected value to generate a corresponding first error; and identify whether the first error exceeds a preset first error range; if it exceeds, send the first property prediction result to the force field parameter optimization module; if it does not exceed, send the first simulation trajectory file to the electrochemical window prediction module; The force field parameter optimization module locally pre-sets multiple property-force field parameter optimization objective functions, and each of the property-force field parameter optimization objective functions corresponds to a property type; the property-force field parameter optimization objective function is used to estimate the property with the force field parameters of the force field topology file as the independent variable and the corresponding property type as the dependent variable, and iteratively optimize the force field parameters of the force field topology file in the direction of minimizing the error between the property estimated value and the property expected value, and output the force field topology file with optimized force field parameters as the corresponding optimized topology file; the force field parameter optimization module is used to save the first initial topology file as the corresponding first force field topology file when receiving it; the force field parameter optimization module is also used to, when receiving the first property prediction result, take the first property prediction result as the corresponding first property initial value, obtain the corresponding first property type and the first property expected value from the data sharing area, take the property-force field parameter optimization objective function corresponding to the first property type as the corresponding first optimization objective function, input the first force field topology file, the first property initial value and the first property expected value into the first optimization objective function to perform force field parameter optimization, and send the optimized topology file output by the function as the corresponding first optimized topology file to the molecular dynamics simulation module, and replace the first force field topology file saved locally with the first optimized topology file; The electrochemical window prediction module includes a cluster structure extraction unit and multiple first method prediction units, and each of the first method prediction units corresponds to a prediction method type; the cluster structure extraction unit is respectively connected to each of the first method prediction units; the cluster structure extraction unit is used to extract the single-frame trajectory cluster structure of the first simulation trajectory file to obtain multiple first cluster structure files; and send each of the first cluster structure files to the first method prediction unit corresponding to the first prediction method type of the data sharing area; the first method prediction unit is used to predict the electrochemical window according to each of the first cluster structure files and output the corresponding first prediction result.

2. The processing system for predicting the electrochemical window of a lithium battery electrolyte according to claim 1, wherein, The first electrolyte formulation includes multiple first electrolyte components; the first electrolyte component includes a first component name, a first component type, a first component SMILES sequence, and a first component proportion; the first component type includes an electrolyte lithium salt type, a solvent type, and an additive type; the first component proportion is the formulation proportion of the current component in the current electrolyte; The first property type includes at least density, dielectric constant, diffusion coefficient, conductivity, and transference number; The first prediction method type includes a plurality of first-type switches; the switch states of the first-type switches include an on state and an off state; each of the first-type switches corresponds to one of the prediction method types; the prediction method types at least include HOMO / LUMO prediction methods, IP / EA prediction methods, thermodynamic cycle prediction methods combined with reorganization energy and solvation energy corrections, and deep learning model prediction methods; The first simulation trajectory file includes a plurality of first single-frame trajectory files, and each of the first single-frame trajectory files corresponds to a sampling time point.

3. The processing system for predicting the electrochemical window of a lithium battery electrolyte according to claim 2, wherein The simulation preprocessing module is specifically configured to, when preparing a force field topology file and a three-dimensional structure file according to the first electrolyte formula to obtain a corresponding first initial topology file and a first electrolyte three-dimensional structure file, generate a corresponding first component three-dimensional structure file for each of the first component SMILES sequences in the first electrolyte formula based on a molecular three-dimensional structure generation software; And use a configuration optimization tool based on DFT theory to optimize each of the first component three-dimensional structure files to obtain corresponding second component three-dimensional structure files; and perform atomic electrostatic potential fitting based on each of the second component three-dimensional structure files to obtain corresponding first component atomic charge files; And use an electrolyte three-dimensional structure generation software to generate an electrolyte three-dimensional structure according to all the second component three-dimensional structure files and the first component ratios to obtain the corresponding first electrolyte three-dimensional structure file; And construct a GAFF2 force field topology file according to the first electrolyte three-dimensional structure file and the first component atomic charge file, and use the constructed GAFF2 force field topology file as the corresponding first initial topology file.

4. The processing system for predicting the electrochemical window of a lithium battery electrolyte according to claim 2, wherein The molecular dynamics simulation module is specifically configured to, when setting an electrolyte lithium salt simulation system according to the first electrolyte three-dimensional structure file and the first electrolyte formula in the data sharing area to obtain a corresponding current simulation system, extract the first component name of the first electrolyte component whose first component type is the electrolyte lithium salt type in the first electrolyte formula as the corresponding current electrolyte lithium salt name; And use the electrolyte lithium salt simulation system corresponding to the current electrolyte lithium salt name as the corresponding current simulation system; and construct a system structure for the current simulation system based on the first electrolyte three-dimensional structure file.

5. The processing system for predicting the electrochemical window of a lithium battery electrolyte according to claim 2, wherein The cluster structure extraction unit is specifically configured to, when extracting single-frame trajectory cluster structures from the first simulation trajectory file to obtain a plurality of first cluster structure files, use each of the first single-frame trajectory files of the first simulation trajectory file as a three-dimensional electrolyte sampling structure file at a sampling time point; Cluster structure recognition is performed on each of the three-dimensional electrolyte sampling structure files to obtain the corresponding first cluster structure files; The first cluster structure files include a plurality of three-dimensional first cluster structures.

6. The processing system for predicting the electrochemical window of a lithium battery electrolyte according to claim 2, wherein When sending each of the first cluster structure files to the first method prediction unit corresponding to the first prediction method type in the data sharing area, the cluster structure extraction unit is specifically configured to obtain the corresponding first prediction method type from the data sharing area; traverse each of the first type switches of the first prediction method type; and when traversing, use the currently traversed first type switch as the corresponding current switch; and when the switch state of the current switch is the on state, send each of the first cluster structure files to the first method prediction unit corresponding to the prediction method type corresponding to the current switch.

Images