Method and device for optimizing material preparation route and electronic equipment

Abstract

The invention provides a material preparation route optimization method and device and electronic equipment, and relates to the field of material synthesis, and the method comprises the steps: inputting an initial text command into a large language model LLM; generating an input file and a structure file required for reflecting molecular dynamics simulation according to the initial text command; executing reaction molecular dynamics simulation based on the input file and the structure file to generate an output file; according to the output file, a simulation result score is output, and the simulation result score is used for reflecting the preparation effect of preparing the target material based on the reaction condition parameters; and outputting optimized reaction condition parameters by utilizing a large language model (LLM) based on the simulation result score. According to the present invention, the repeatability and the reliability of the material preparation are improved, the environmental protection and safety modern industrial requirements are met, and the efficient and environmental protection new path is provided for the synthesis of MXenes and other advanced materials.

Description

Technical Field

[0001] The present disclosure relates to the field of material synthesis, and in particular to a method, device and electronic equipment for optimizing a material preparation route. Background Art

[0002] Two-dimensional transition metal carbides and nitrides such as MXenes have shown great application potential in energy storage, sensors, catalysts, and electromagnetic shielding due to their excellent electrical conductivity, mechanical properties, and chemical stability. Although these two-dimensional materials have many excellent properties, their synthesis process still faces many technical challenges and environmental risks.

[0003] Related preparation methods rely on the use of toxic reactants (such as hydrofluoric acid), which not only poses a serious threat to the health of operators, but also causes environmental pollution during its use and disposal. Its preparation process not only has high safety risks, but also requires complex waste liquid treatment facilities, which increases the cost of material synthesis and the environmental burden. Therefore, a safer, environmentally friendly and self-optimizing material preparation route is urgently needed. Summary of the invention

[0004] The embodiments of the present disclosure provide a method, device and electronic device for optimizing a material preparation route, aiming to solve the problems existing in the above-mentioned background technology.

[0005] In order to solve the above technical problems, the present disclosure is implemented as follows:

[0006] In a first aspect, an embodiment of the present disclosure provides a method for optimizing a material preparation route, the method comprising:

[0007] Inputting an initial text command into a large language model LLM, wherein the initial text command specifies reaction condition parameters for preparing a target material;

[0008] Using the large language model LLM, generating input files and structure files required for reaction molecular dynamics simulation according to the initial text command;

[0009] Perform reaction molecular dynamics simulation based on the input file and the structure file to generate an output file;

[0010] Outputting a simulation result score according to the output file using the large language model LLM, wherein the simulation result score is used to reflect the preparation effect of the target material based on the reaction condition parameters;

[0011] Based on the simulation result score, the large language model LLM is used to output optimized reaction condition parameters.

[0012] Optionally, the method further comprises:

[0013] Generate a parameter extraction text command according to the output file, wherein the parameter extraction text command is used to instruct the large language model LLM to extract evaluation parameters from the output file, wherein the evaluation parameters include reactant consumption, product purity, reaction time, and number of contaminated product categories;

[0014] Generate a scoring text command according to the evaluation parameters, wherein the scoring text command is used to instruct the large language model LLM to calculate the simulation result score based on the evaluation parameters;

[0015] Generate an optimization text command according to the input file, the structure file and the simulation result score, wherein the optimization text command is used to instruct the large language model LLM to optimize the reaction condition parameters;

[0016] A new initial text command is generated according to the optimized reaction condition parameters, and the new initial text command is used to instruct the large language model LLM to generate input files and structure files required for a new round of reaction molecular dynamics simulation.

[0017] Optionally, after outputting the optimized reaction condition parameters using the large language model LLM based on the simulation result score, the method further includes:

[0018] According to the optimized reaction condition parameters, using the large language model LLM to generate input files and structure files required for a new round of reaction molecular dynamics simulation;

[0019] Based on the input files and structure files required for the new round of reaction molecular dynamics simulation, executing reaction molecular dynamics simulation to generate new output files;

[0020] Outputting a new simulation result score according to the new output file using the large language model LLM, wherein the new simulation result score is higher than the simulation result score;

[0021] Based on the new simulation result score, using the large language model LLM to output optimized reaction condition parameters;

[0022] Execution steps: according to the optimized reaction condition parameters, using the large language model LLM to generate input files and structure files required for a new round of reaction molecular dynamics simulation; until the simulation result score output by the large language model LLM is higher than a preset score threshold.

[0023] Optionally, using the large language model LLM, outputting a simulation result score according to the output file includes:

[0024] Extracting evaluation parameters from the output file using the large language model LLM, wherein the evaluation parameters include reactant consumption, product purity, reaction time, and number of contaminated product categories;

[0025] According to the evaluation parameters, the reaction molecular dynamics simulation process is evaluated using the large language model LLM, and the simulation results are output for scoring.

[0026] Optionally, according to the evaluation parameters, the reaction molecular dynamics simulation process is evaluated using the large language model LLM, and the simulation results are output for scoring, including:

[0027] Using the large language model LLM, the reactant consumption, product purity, reaction time and number of contaminated product categories in the reaction molecular dynamics simulation process are scored respectively to obtain scores for each evaluation parameter;

[0028] The large language model LLM is used to weight the scores of the respective evaluation parameters to synthesize the simulation result score, and the simulation result score is output.

[0029] Optionally, the reaction condition parameters include reaction temperature, reaction pressure, reactant type and reactant concentration;

[0030] Using the large language model LLM, the input files and structure files required by the LAMMPS software are generated according to the initial text command, including:

[0031] parse the reaction temperature, reaction pressure, reactant type and reactant concentration from the initial text command using the large language model LLM;

[0032] Generate a structure file using the large language model LLM according to the type of reactant, wherein the structure file is used to reflect the initial molecular structure and atomic coordinates of the target material generated based on the reaction condition parameters;

[0033] According to the reaction temperature, the reaction pressure and the reactant concentration, the large language model LLM is used to generate an input file, wherein the input file is used to define simulation parameters required for performing the molecular dynamics simulation of the reaction, and the simulation parameters include time step, simulation duration, temperature value and pressure value.

[0034] Optionally, performing a reaction molecular dynamics simulation based on the input file and the structure file to generate an output file includes:

[0035] Load the structure file into LAMMPS software, initialize the simulation environment, and configure the physical and chemical environment parameters required for the reaction molecular dynamics simulation;

[0036] Performing a reaction molecular dynamics simulation, using the LAMMPS software to perform numerical calculations based on the physical and chemical environment parameters and the simulation parameters, and generating an output file.

[0037] Optionally, the method further comprises:

[0038] Using the large language model LLM, data processing is performed on the data in the output file to determine potential problems in the reaction molecular dynamics simulation process, wherein the data processing method includes but is not limited to at least one of the following: data clustering, trend analysis, and anomaly detection;

[0039] The large language model LLM is used to adjust the simulation parameters in the input file according to potential problems in the reaction molecular dynamics simulation process to generate an adjusted input file.

[0040] In a second aspect, an embodiment of the present disclosure provides a device for optimizing a material preparation route, the device comprising:

[0041] An input module, used for inputting an initial text command into a large language model LLM, wherein the initial text command specifies reaction condition parameters for preparing a target material;

[0042] A first generation module is used to generate input files and structure files required for reaction molecular dynamics simulation according to the initial text command by using the large language model LLM;

[0043] A second generating module, used for performing a reaction molecular dynamics simulation based on the input file and the structure file to generate an output file;

[0044] A first output module is used to output a simulation result score according to the output file using the large language model LLM, wherein the simulation result score is used to reflect the preparation effect of the target material based on the reaction condition parameters;

[0045] The second output module is used to output the optimized reaction condition parameters based on the simulation result score using the large language model LLM.

[0046] Optionally, the device further comprises:

[0047] A first text generation module is used to generate a parameter extraction text command according to the output file, wherein the parameter extraction text command is used to instruct the large language model LLM to extract evaluation parameters from the output file, wherein the evaluation parameters include reactant consumption, product purity, reaction time, and number of contaminated product categories;

[0048] A second text generation module is used to generate a scoring text command according to the evaluation parameters, and the scoring text command is used to instruct the large language model LLM to calculate the simulation result score based on the evaluation parameters;

[0049] A third text generation module is used to generate an optimization text command according to the input file, the structure file and the simulation result score, wherein the optimization text command is used to instruct the large language model LLM to optimize the reaction condition parameters;

[0050] The fourth text generation module is used to generate new initial text commands according to the optimized reaction condition parameters, and the new initial text commands are used to instruct the large language model LLM to generate input files and structure files required for a new round of reaction molecular dynamics simulation.

[0051] Optionally, the device further comprises:

[0052] A fourth generation module is used to generate input files and structure files required for a new round of reaction molecular dynamics simulation using the large language model LLM according to the optimized reaction condition parameters;

[0053] A fifth generation module, used to perform reaction molecular dynamics simulation based on the input files and structure files required for the new round of reaction molecular dynamics simulation, and generate new output files;

[0054] A third output module is used to output a new simulation result score according to the new output file using the large language model LLM, wherein the new simulation result score is higher than the simulation result score;

[0055] A fourth output module, configured to output optimized reaction condition parameters using the large language model LLM based on the new simulation result score;

[0056] The module is repeatedly executed to execute the steps of: generating input files and structure files required for a new round of reaction molecular dynamics simulation using the large language model LLM according to the optimized reaction condition parameters; until the simulation result score output by the large language model LLM is higher than a preset score threshold.

[0057] Optionally, the first output module includes:

[0058] An extraction submodule, for extracting evaluation parameters from the output file using the large language model LLM, wherein the evaluation parameters include reactant consumption, product purity, reaction time, and number of contaminated product categories;

[0059] The evaluation submodule is used to evaluate the reaction molecular dynamics simulation process according to the evaluation parameters using the large language model LLM, and output the simulation results for scoring.

[0060] Optionally, the first output module includes:

[0061] The sub-item scoring submodule is used to score the reactant consumption, product purity, reaction time and number of contaminated product categories in the reaction molecular dynamics simulation process respectively by using the large language model LLM to obtain scores for each evaluation parameter;

[0062] The weighting submodule is used to use the large language model LLM to weight the scores of each evaluation parameter to synthesize the simulation result score, and output the simulation result score.

[0063] Optionally, the reaction condition parameters include reaction temperature, reaction pressure, reactant type and reactant concentration; the first generation module includes:

[0064] A parsing submodule, for parsing the reaction temperature, reaction pressure, reactant type and reactant concentration from the initial text command using the large language model LLM;

[0065] A structure file generation submodule, used to generate a structure file using the large language model LLM according to the type of reactants, wherein the structure file is used to reflect the initial molecular structure and atomic coordinates of the target material generated based on the reaction condition parameters;

[0066] An input file generation submodule is used to generate an input file using the large language model LLM according to the reaction temperature, the reaction pressure and the reactant concentration, wherein the input file is used to define simulation parameters required for performing the reaction molecular dynamics simulation, and the simulation parameters include time step, simulation duration, temperature value and pressure value.

[0067] Optionally, the second generating module includes:

[0068] An initial simulation submodule, used to load the structure file into the LAMMPS software, initialize the simulation environment, and configure the physical and chemical environment parameters required for the reaction molecular dynamics simulation;

[0069] The simulation execution submodule is used to execute the reaction molecular dynamics simulation, use the LAMMPS software to perform numerical calculations based on the physical and chemical environmental parameters and the simulation parameters, and generate an output file.

[0070] Optionally, the device further comprises:

[0071] A potential problem determination module is used to process the data in the output file using the large language model LLM to determine potential problems in the reaction molecular dynamics simulation process, wherein the data processing method includes but is not limited to at least one of the following: data clustering, trend analysis, and anomaly detection;

[0072] The adjustment module is used to use the large language model LLM to adjust the simulation parameters in the input file according to potential problems in the reaction molecular dynamics simulation process to generate an adjusted input file.

[0073] In a third aspect, an embodiment of the present disclosure provides an electronic device, comprising: a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of a method for optimizing a material preparation route.

[0074] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects:

[0075] The present disclosure automatically generates the required reaction molecular dynamics simulation input files and structure files by integrating the application of the large language model LLM. According to the output files generated by the simulation, the large language model LLM can evaluate the effect of preparing the target material and output the simulation result score. The scoring system of the large language model LLM directly reflects the efficiency and quality of material preparation under different reaction conditions and parameters, making the parameter optimization in the preparation process more scientific and accurate, while ensuring the performance of the material, reducing resource waste and environmental pollution. It not only improves the repeatability and reliability of material preparation, but also meets the modern industrial requirements of environmental protection and safety, and provides a new and more efficient and environmentally friendly path for the synthesis of MXenes and other advanced materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for describing the embodiments of the present disclosure are briefly introduced below.

[0077] Figure 1 It is a schematic diagram of the steps of a method for optimizing a material preparation route provided by an embodiment of the present disclosure;

[0078] Figure 2 It is a schematic diagram of a LLM-Prompt self-optimization process provided by an embodiment of the present disclosure;

[0079] Figure 3 It is a schematic diagram of an iterative architecture of a material preparation route optimization method provided by an embodiment of the present disclosure;

[0080] Figure 4It is a structural block diagram of a device for optimizing a material preparation route provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0081] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0082] The core of the present disclosure is to use the large language model LLM to automatically generate and optimize the simulation process of target material preparation, thereby reducing dependence on hazardous chemicals, reducing environmental pollution, and improving the accuracy and repeatability of the preparation process. Through the LLM-Prompt self-optimization technology, the present disclosure realizes a complete automated process from setting the initial reaction conditions, automatically generating simulation files, to scoring simulation results and optimizing parameters. In addition, the method can also gradually improve material performance through continuous iterative optimization to ensure high quality and high performance of the final product.

[0083] For ease of understanding, the embodiments of the present disclosure take the typical two-dimensional material MXenes as an example, that is, taking the target material as MXenes, and explain in detail the optimization method of the MXenes preparation route, but it does not mean that the optimization method of the material preparation route of the present disclosure is limited to the application of MXenes preparation. Figure 1 is a schematic diagram of the steps of a method for optimizing a material preparation route provided by an embodiment of the present disclosure, such as Figure 1 As shown, the method includes:

[0084] Step S101 : inputting an initial text command into a large language model LLM, wherein the initial text command specifies reaction condition parameters for preparing a target material.

[0085] Step S102, using the large language model LLM, generating input files and structure files required for reaction molecular dynamics simulation according to the initial text command.

[0086] Step S103, performing reaction molecular dynamics simulation based on the input file and the structure file to generate an output file.

[0087] Step S104, using the large language model LLM, according to the output file, outputting a simulation result score, wherein the simulation result score is used to reflect the preparation effect of the target material based on the reaction condition parameters.

[0088] Step S105: Based on the simulation result score, the large language model LLM is used to output optimized reaction condition parameters.

[0089] In step S101, the initial text command is the basis for instructing the large language model (LLM) to generate the input file required for the reaction molecular dynamics simulation. The initial text command contains the reaction condition parameters required for the synthesis of the target material. The reaction condition parameters specify the necessary parameters required for the simulation reaction, such as reaction temperature, reaction pressure, reactant type, etc. Exemplarily, the initial text command can be as follows:

[0090] <Please help me generate two files, namely the LAMMPS in file and the structure file, for the reaction molecular dynamics simulation of the MXenes preparation process. The specific parameter settings are as follows: reaction temperature: 300K; reaction pressure: 1KPa; reactant type: Ti3AlC2, HF; reactant concentration: Ti3AlC2 1mol / L, HF 3mol / L. >. The large language model LLM can use open source pre-trained large language model LLM, such as Yuan2.0, LLaMa, Vicuna, etc.

[0091] Involving step S102, reaction molecular dynamics simulation (RMD) is a commonly used simulation method. By simulating the reaction process, the synthesis path of the material can be predicted and optimized. After receiving the initial text command input in step S101, the corresponding input file (in file) and structure file will be generated according to the command content for performing reaction molecular dynamics (RMD) simulation in the open source LAMMPS software. In the embodiment of the present disclosure, the reaction molecular dynamics simulation is implemented with the help of LAMMPS software. According to the example of the initial text command in step S101, the generated input file can be as follows:

[0092] units metal

[0093] atom_style atomic

[0094] boundary ppp

[0095] read_data structure_file.data

[0096] pair_style eam

[0097] pair_coeff**potentials / TiAlC2.eam.alloy Ti Al C

[0098] fix 1all nvt temp 300 300 100

[0099] fix 2all press / berendseniso 1 1 1000

[0100] run 10000; in this input file, units metal sets the physical units of the simulation to metal units, atom_style atomic sets the atomic style to simple atom style; boundary ppp sets the boundary condition of the simulation box to periodic boundary; read_data structure_file.data reads the structure file and defines the initial atomic position; pair_style eam: specifies the interatomic interaction potential as EAM potential; pair_coeff**potentials / TiAlC2.eam.alloy Ti Al C: sets the parameter file of EAM potential and the corresponding elements; fix 1all nvt temp300 300 100: uses the NVT integration method to keep the system temperature at 300K; fix 2all press / berendsen iso 11 1000: uses the Berendsen method to keep the system isothermal and isobaric (1KPa); run 10000: runs 10,000 time steps.

[0101] Correspondingly, the generated structure file can be as follows:

[0102] LAMMPS data file for Ti3AlC2 HF reaction

[0103] 4atoms

[0104] 2atom types

[0105] 0.0 10.0xlo xhi

[0106] 0.0 10.0ylo yhi

[0107] 0.0 10.0zlo zhi

[0108] Atoms 1 1 1.0 1.0 1.0 2 1 2.0 2.0 2.0 3 2 3.0 3.0 3.0

[0112] 4 2 4.0 4.0 4.0; In this structure file, LAMMPS data file for Ti3AlC2 HFreaction is a file header comment, which explains the purpose of the file; 4atoms defines that there are 4 atoms in the simulation; 2atom types defines that there are 2 types of atoms in the simulation; 0.0 10.0xlo xhi, 0.0 10.0ylo yhi, 0.0 10.0zlo zhi defines the size of the simulation box; Atoms: The following is the atomic information part; 1 1 1.0 1.01.0 represents the first atom, type 1, coordinates (1.0, 1.0, 1.0); 2 1 2.0 2.0 2.0 represents the second atom, type 1, coordinates (2.0, 2.0, 2.0); 3 23.0 3.0 3.0 represents the third atom, type 2, coordinates (3.0, 3.0, 3.0); 4 2 4.0 4.0 4.0 represents the 4th atom, type 2, and coordinates (4.0, 4.0, 4.0).

[0113] Involving step S103, according to the input file and structure file generated in step S102, the simulation parameters and the initial state of the system are obtained. Initialize the atoms in the simulation system according to the atomic coordinates and types defined in the structure file. Set the temperature, pressure and other conditions of the simulation according to the parameters in the input file. Use the defined force field (such as EAM potential) to calculate the interaction force between atoms at each time step. Update the velocity and position of the atoms to complete the integral calculation of a time step. During the simulation process, handle the problem of atomic crossing the boundary. If the atom crosses the boundary of the simulation box, it is reintroduced into the simulation box according to the boundary conditions (such as periodic boundaries). At each specified time step, record the state parameters of the simulation system, such as atomic position and velocity, total energy of the system, kinetic energy, potential energy, temperature, pressure and other thermodynamic parameters. Correspondingly, the generated output file can be as follows:

[0114] Step Temp Press PotEng KinEng TotEng

[0115] 0 300.0 1.0 -2000.0 1000.0 -1000.0

[0116] 1000 310.0 1.1 -1995.0 1020.0 -975.0

[0117] 2000 320.0 1.2 -1980.0 1040.0 -940.0

[0118] ....

[0119] In step S104, in the process of outputting the simulation result score according to the output file, the scoring basis is multi-dimensional. The large language model LLM will score multiple indicators and weight the score of each indicator to obtain the final simulation result score. The scoring items can be as follows:

[0120] Reactant consumption: 20 points (15 is higher); Product purity: 90 points (85% is higher); Reaction time: 30 points (1000 is longer); Number of contaminated product categories: 80 points (2 is fewer). The final simulation result score will be used to reflect the preparation effect of the target material based on the reaction condition parameters, serve as the basis for subsequent optimization iterations, and guide the adjustment and optimization of the reaction conditions.

[0121] In step S105, the large language model LLM is used in combination with the simulation result score to comprehensively consider the influence of each parameter on the reaction effect, propose the optimization direction, and generate the optimized reaction condition parameters. The optimized reaction conditions can be as follows: reaction temperature: 320K; reaction pressure: 1.2KPa; reactant type: Ti3AlC2, HF; reactant concentration: Ti3AlC2 1.2mol / L, HF 2.8mol / L, which improves material performance and preparation efficiency, reduces experimental costs and manual intervention, and promotes the optimization of MXenes material synthesis routes.

[0122] The present disclosure automatically generates the required reaction molecular dynamics simulation input files and structure files by integrating the application of the large language model LLM. According to the output files generated by the simulation, the large language model LLM can evaluate the effect of preparing the target material and output the simulation result score. The scoring system of the large language model LLM directly reflects the efficiency and quality of material preparation under different reaction conditions and parameters, making the parameter optimization in the preparation process more scientific and accurate, while ensuring the performance of the material, reducing resource waste and environmental pollution. It not only improves the repeatability and reliability of material preparation, but also meets the modern industrial requirements of environmental protection and safety, and provides a new and more efficient and environmentally friendly path for the synthesis of MXenes and other advanced materials.

[0123] Exemplarily, the method further includes: generating a parameter extraction text command according to the output file, the parameter extraction text command being used to instruct the large language model LLM to extract evaluation parameters from the output file, the evaluation parameters including reactant consumption, product purity, reaction time and number of contaminated product categories; generating a scoring text command according to the evaluation parameters, the scoring text command being used to instruct the large language model LLM to calculate the simulation result score based on the evaluation parameters; generating an optimization text command according to the input file, the structure file and the simulation result score, the optimization text command being used to instruct the large language model LLM to optimize the reaction condition parameters; generating a new initial text command according to the optimized reaction condition parameters, the new initial text command being used to instruct the large language model LLM to generate the input file and structure file required for a new round of reaction molecular dynamics simulation.

[0124] Figure 2 FIG. 1 is a schematic diagram of a LLM-Prompt self-optimization process provided by an embodiment of the present disclosure, such as Figure 2 As shown, a parameter extraction text command prompt_1 for extracting evaluation parameters is set. Prompt_1 is used to guide the large language model LLM to extract specific parameters from the output file. The parameter extraction text command Prompt_1 can be as follows:

[0125] <#out file

[0126] The out file of the molecular dynamics simulation comes from the reaction molecular dynamics simulation of the MXenes preparation process. Please extract the following parameters from the out file of the molecular dynamics simulation:

[0127] Reactant consumption, product purity, reaction time, number of polluted product categories. >. The parameter extraction text command is used to guide the large language model LLM to extract four evaluation parameters, namely, reactant consumption, product purity, reaction time, and number of polluted product categories, from the out file. The evaluation parameters will be used to evaluate the reaction effect.

[0128] A scoring text command Prompt_2 is generated according to the evaluation parameters. Prompt_2 is used to guide the large language model LLM to score according to the extracted evaluation parameters. The scoring text command Prompt_2 can be as follows:

[0129] <#Reactant consumption:

[0130] Product purity:

[0131] #Reaction time:

[0132] #Number of pollution product categories:

[0133] The above four parameters are the simulation results of the reaction molecular dynamics simulation of the MXenes preparation process. Please score the above parameters, with a score range of 0-100. The less the reactant consumption, the better, the higher the product purity, the better, the shorter the reaction time, the better, and the fewer the number of polluted product categories, the better. >. That is, the large language model LLM is guided to score according to the four evaluation parameters. The scoring criteria for each parameter are: the less the reactant consumption, the better, the higher the product purity, the better, and the shorter the reaction time, the better.

[0134] The fewer the number of contamination product categories, the better.

[0135] According to the input file, structure file and simulation result score, an optimization text command Prompt_3 is generated. Prompt_3 is used to guide the large language model LLM to optimize the reaction conditions according to the evaluation parameters and scores. The optimization text command Prompt_3 can be as follows:

[0136] <#in file

[0137] Structure File

[0138] Results scoring of four parameters

[0139] The above are the in files, structure files and simulation result scores of the reaction molecular dynamics simulation of the MXenes preparation process. Please optimize the reaction temperature, reaction pressure, reactant type and reactant concentration according to the above reference information, and give the optimized reaction temperature, reaction pressure, reactant type and reactant concentration. >. The optimization text command guides the large language model LLM to optimize the reaction temperature, pressure, reactant type and concentration according to the current in files, structure files and simulation result scores, and give the optimized parameters.

[0140] A new initial text command Prompt_4 is generated for generating a new round of simulation files according to the optimized reaction condition parameters. Prompt_4 guides the large language model LLM to generate new input files and structure files. The initial text command Prompt_4 can be as follows:

[0141] Please help me generate two files, namely the LAMMPS in file and the structure file, for the reaction molecular dynamics simulation of the MXenes preparation process. The specific parameter settings are as follows:

[0142] Reaction temperature: #Optimized reaction temperature; Reaction pressure: #Optimized reaction pressure; Reactant type: #Optimized reactant type; Reactant concentration: #Optimized reactant concentration. >. The initial text command Prompt_4 guides the large language model LLM to generate new in files and structure files. The new files will be used for a new round of reaction molecular dynamics simulations. The specific parameters are based on the results of the previous round of optimization.

[0143] Exemplarily, after the large language model LLM is used to output the optimized reaction condition parameters based on the simulation result score, the method further includes: generating input files and structure files required for a new round of reaction molecular dynamics simulation based on the optimized reaction condition parameters using the large language model LLM; performing reaction molecular dynamics simulation based on the input files and structure files required for the new round of reaction molecular dynamics simulation to generate new output files; outputting a new simulation result score based on the new output file using the large language model LLM, the new simulation result score being higher than the simulation result score; outputting the optimized reaction condition parameters based on the new simulation result score using the large language model LLM; executing the steps of: generating input files and structure files required for a new round of reaction molecular dynamics simulation based on the optimized reaction condition parameters using the large language model LLM; until the simulation result score output by the large language model LLM is higher than a preset score threshold.

[0144] Figure 3 is a schematic diagram of an iterative architecture of an optimization method for a material preparation route provided by an embodiment of the present disclosure, such as Figure 3As shown, the optimization method of the material preparation route provided by the present disclosure is a process of continuous iteration and optimization. In each iteration, the input file and structure file generated by the large language model LLM are subjected to reaction molecular dynamics simulation in the open source LAMMPS software to generate an output file, and then the simulation results are scored and optimized by the large language model LLM. Specifically, according to the scoring results of the simulation results, the reaction conditions (temperature, pressure, reactant type, reactant concentration) are optimized. The optimized parameters generate new input files and structure files, and a new round of reaction molecular dynamics simulation is performed. Each iteration repeats the steps of the previous round, but uses the optimized reaction conditions to perform a new reaction molecular dynamics simulation. Reaction molecular dynamics simulation is performed by LAMMPS to generate a new output file. The large language model LLM extracts relevant parameters from the new output file and scores. According to the new scoring results, the reaction conditions are optimized again, and new input files and structure files are generated. Each iteration includes executing reaction molecular dynamics simulation, extracting and scoring results, optimizing reaction conditions, generating new input files and structure files, and iterating continuously until the score of the simulation result reaches a preset score threshold. When the simulation result score of a certain iteration reaches or exceeds the preset score threshold, the iteration is stopped. At this time, the reaction conditions obtained are the expected, optimized optimal conditions.

[0145] It can be seen from the above steps that the method disclosed in the present invention is a process of continuous iterative optimization. In each iteration, the preparation conditions of the target material are gradually optimized through the steps of simulation execution, result extraction, scoring, condition optimization, etc. Through multiple iterations, the optimal reaction conditions are finally found, making the material preparation process more efficient, safe and environmentally friendly.

[0146] Exemplarily, the large language model LLM is used to output a simulation result score according to the output file, including: using the large language model LLM to extract evaluation parameters from the output file, the evaluation parameters including reactant consumption, product purity, reaction time and the number of contaminated product categories; based on the evaluation parameters, the large language model LLM is used to evaluate the reaction molecular dynamics simulation process, and the simulation result is output for scoring.

[0147] After the reaction molecular dynamics simulation is completed, the generated output file contains the evaluation parameters of the entire simulation process, including reactant consumption, product purity, reaction time, and the number of contaminated product categories. The large language model LLM is used to parse the output file and extract the evaluation parameters from the output file. Using the extracted evaluation parameters, the large language model LLM evaluates and scores the simulation process. The scoring criteria are pre-set, taking into account the importance and ideal value of each parameter. Based on the scoring results, the large language model LLM optimizes the reaction conditions to improve the simulation result score. Using the optimized reaction conditions, new input files and structure files are generated to start a new round of reaction molecular dynamics simulation.

[0148] Exemplarily, according to the evaluation parameters, the large language model LLM is used to evaluate the reaction molecular dynamics simulation process, and the simulation results are output for scoring, including: using the large language model LLM to score the reactant consumption, product purity, reaction time and number of contaminated product categories in the reaction molecular dynamics simulation process, respectively, to obtain a score for each evaluation parameter; using the large language model LLM to weight the scores of each evaluation parameter to synthesize the simulation result score, and output the simulation result score.

[0149] According to the extracted evaluation parameters, the large language model LLM is used to score each parameter. The scoring criteria can be: the less the reactant consumption, the better (full score 30 points); the higher the product purity, the better (full score 40 points); the shorter the reaction time, the better (full score 20 points); the fewer the number of contaminated product categories, the better (full score 10 points). The scores of each evaluation parameter obtained according to the scoring criteria are: reactant consumption score 25, product purity score 38, reaction time score 15, and number of contaminated product categories score 8. The large language model LLM is used to synthesize the scores of each evaluation parameter according to the weights to form a total score. The weights of reactant consumption, product purity, reaction time and number of contaminated product categories are 30%, 40%, 20% and 10% respectively. Total score = reactant consumption score × 0.3 + product purity score × 0.4 + reaction time score × 0.2 + contamination product category score × 0.1 = 25 × 0.3 + 38 × 0.4 + 15 × 0.2 + 8 × 0.1 = 7.5 + 15.2 + 3 + 0.8 = 26.5. Therefore, the final simulation result score is 26.5 points.

[0150] Exemplarily, the reaction condition parameters include reaction temperature, reaction pressure, reactant type and reactant concentration; using the large language model LLM, generating input files and structure files required by LAMMPS software according to the initial text command, including: using the large language model LLM to parse the reaction temperature, reaction pressure, reactant type and reactant concentration from the initial text command; using the large language model LLM to generate a structure file according to the reactant type, the structure file is used to reflect the initial molecular structure and atomic coordinates of the target material generated based on the reaction condition parameters; using the large language model LLM to generate an input file according to the reaction temperature, the reaction pressure and the reactant concentration, the input file is used to define simulation parameters required to perform the reaction molecular dynamics simulation, the simulation parameters including time step, simulation duration, temperature value and pressure value.

[0151] In the disclosed embodiment, the reaction condition parameters include reaction temperature, reaction pressure, reactant type and reactant concentration. As mentioned above, the structure file reflects the initial molecular structure and atomic coordinates of the target material generated based on the reaction condition parameters. The large language model LLM generates this file based on the reactant type. The input file defines the simulation parameters required to perform the reaction molecular dynamics simulation, including the time step, simulation duration, temperature value and pressure value.

[0152] The large language model LLM generates the content of the structure file based on the analyzed reactant types. Specifically, the large language model LLM determines the atomic arrangement and coordinates of each substance based on the reactant types (such as Ti3AlC2, HF). According to the crystal structure and chemical composition of the material, the corresponding structure file is generated. For example, the structure file of Ti3AlC2 is:

[0153] #Ti3AlC2 structure file 3 1 0 0 0

[0155] Al 0.0 0.0 0.0

[0156] Ti 1.0 0.0 0.0

[0157] C 0.5 0.5 0.5.

[0158] The large language model LLM uses the parsed reaction temperature, reaction pressure and reactant concentration to generate the content of the input file. Specifically, the large language model LLM reads the reaction temperature (300K), reaction pressure (1KPa), reactant concentration (Ti3AlC21mol / L, HF 3mol / L), and defines the simulation parameters required for LAMMPS simulation based on these parameters, including time step, simulation duration, temperature control and pressure control, etc. For example, the generated input file is:

[0159] #LAMMPS input file

[0160] units metal

[0161] dimension 3

[0162] boundary ppp

[0163] atom_style atomic

[0164] # Initialization settings

[0165] read_data Ti3AlC2.structure

[0166] pair_style eam

[0167] pair_coeff**Ti3AlC2.eam

[0168] #Define simulation parameters

[0169] timestep 0.001

[0170] thermo 100

[0171] run 10000

[0172] #Temperature and pressure settings

[0173] fix 1all npt temp 300 300 0.1iso 1.0 1.0 1.0

[0174] ....

[0175] Exemplarily, performing a reaction molecular dynamics simulation based on the input file and the structure file to generate an output file includes: loading the structure file into the LAMMPS software, initializing the simulation environment, and configuring the physical and chemical environment parameters required for the reaction molecular dynamics simulation; performing a reaction molecular dynamics simulation, using the LAMMPS software to perform numerical calculations based on the physical and chemical environment parameters and the simulation parameters, and generating an output file.

[0176] The structure file generated by the large language model LLM is loaded into the LAMMPS software. Since the structure file contains the initial molecular structure and atomic coordinates of the target material, it is equivalent to importing the initial state of the material into the simulation environment for subsequent molecular dynamics simulation. After loading the structure file, the simulation environment needs to be initialized. This includes defining the physical and chemical environment parameters required for molecular dynamics simulation, such as force field type, temperature control, pressure control, etc. After initializing the simulation environment, it is necessary to further configure specific physical and chemical environment parameters, which come from the input file generated by the large language model LLM, including reaction temperature, reaction pressure, simulation duration, time step, etc. After the configuration is completed, the reaction molecular dynamics simulation is performed. Based on the set physical and chemical parameters, the LAMMPS software simulates the interaction between molecules through numerical calculations and tracks the motion trajectory of molecules, thereby simulating the dynamic reaction process of the material. During the simulation execution process, the LAMMPS software records the simulation results and generates an output file. The output file contains detailed data of the molecular dynamics simulation, including atomic coordinate changes, energy changes, reactant consumption, product formation, etc.

[0177] Exemplarily, the method further includes: using the large language model LLM to perform data processing on the data in the output file to determine potential problems in the reaction molecular dynamics simulation process, wherein the data processing method includes but is not limited to at least one of the following: data clustering, trend analysis, and anomaly detection; using the large language model LLM to adjust the simulation parameters in the input file according to the potential problems in the reaction molecular dynamics simulation process to generate an adjusted input file.

[0178] The data in the output file is processed using the large language model LLM. Since the output file contains a large amount of simulation data, valuable information can be extracted from it through methods such as data clustering, trend analysis and anomaly detection to identify potential problems in the simulation process. Specifically, data clustering is to cluster the output data to identify different types of atomic or molecular behaviors during the reaction process. For example, the atomic coordinate data can be clustered to distinguish different reaction areas. Trend analysis is to analyze the changing trend of simulation data over time, such as energy change, temperature change, etc., to identify the main stages and key points of the reaction. Anomaly detection is to detect outliers or abnormal patterns in the output data to identify abnormal behaviors that may affect the reaction effect, such as abnormal reactant consumption, abnormal product purity, etc. Through data processing, the large language model LLM can identify potential problems in the simulation process. For example, according to the clustering results of data clustering, it is found that the atomic behavior in a specific area is abnormal, which may be due to unsuitable local reaction conditions; for example, according to the trend analysis results of trend analysis, it is found that the energy change trend is unreasonable, and the reaction process is stagnant or incomplete.

[0179] Based on the potential problems identified, the large language model (LLM) is used to adjust the simulation parameters in the input file to optimize the next round of simulation.

[0180] For example, according to the energy change trend, the reaction temperature and pressure are adjusted to avoid reaction stagnation; according to the abnormal detection results of reactant consumption, the reactant concentration is adjusted to improve the reaction efficiency; according to the cluster analysis results, the simulation duration and time step are adjusted to ensure that the reaction process is fully carried out.

[0181] Through the above steps, the large language model LLM can be used to process the data in the output file, identify potential problems, and optimize the simulation parameters in the input file based on these problems, thereby improving the effect of reaction molecular dynamics simulation.

[0182] Figure 4 is a structural block diagram of a material preparation route optimization device provided by an embodiment of the present disclosure, such as Figure 4 As shown, the device comprises:

[0183] An input module 201 is used to input an initial text command into a large language model LLM, wherein the initial text command specifies reaction condition parameters for preparing a target material;

[0184] A first generating module 202 is used to generate input files and structure files required for reaction molecular dynamics simulation according to the initial text command by using the large language model LLM;

[0185] A second generating module 203 is used to perform a reaction molecular dynamics simulation based on the input file and the structure file to generate an output file;

[0186] A first output module 204 is used to output a simulation result score according to the output file using the large language model LLM, wherein the simulation result score is used to reflect the preparation effect of the target material based on the reaction condition parameters;

[0187] The second output module 205 is used to output the optimized reaction condition parameters based on the simulation result score using the large language model LLM.

[0188] Exemplarily, the device further comprises:

[0189] A first text generation module is used to generate a parameter extraction text command according to the output file, wherein the parameter extraction text command is used to instruct the large language model LLM to extract evaluation parameters from the output file, wherein the evaluation parameters include reactant consumption, product purity, reaction time, and number of contaminated product categories;

[0190] A second text generation module is used to generate a scoring text command according to the evaluation parameters, and the scoring text command is used to instruct the large language model LLM to calculate the simulation result score based on the evaluation parameters;

[0191] A third text generation module is used to generate an optimization text command according to the input file, the structure file and the simulation result score, wherein the optimization text command is used to instruct the large language model LLM to optimize the reaction condition parameters;

[0192] The fourth text generation module is used to generate new initial text commands according to the optimized reaction condition parameters, and the new initial text commands are used to instruct the large language model LLM to generate input files and structure files required for a new round of reaction molecular dynamics simulation.

[0193] Exemplarily, the device further comprises:

[0194] A fourth generation module is used to generate input files and structure files required for a new round of reaction molecular dynamics simulation using the large language model LLM according to the optimized reaction condition parameters;

[0195] A fifth generation module, used to perform reaction molecular dynamics simulation based on the input files and structure files required for the new round of reaction molecular dynamics simulation, and generate new output files;

[0196] A third output module is used to output a new simulation result score according to the new output file using the large language model LLM, wherein the new simulation result score is higher than the simulation result score;

[0197] A fourth output module, configured to output optimized reaction condition parameters using the large language model LLM based on the new simulation result score;

[0198] The module is repeatedly executed to execute the steps of: generating input files and structure files required for a new round of reaction molecular dynamics simulation using the large language model LLM according to the optimized reaction condition parameters; until the simulation result score output by the large language model LLM is higher than a preset score threshold.

[0199] Exemplarily, the first output module includes:

[0200] An extraction submodule, for extracting evaluation parameters from the output file using the large language model LLM, wherein the evaluation parameters include reactant consumption, product purity, reaction time, and number of contaminated product categories;

[0201] The evaluation submodule is used to evaluate the reaction molecular dynamics simulation process according to the evaluation parameters using the large language model LLM, and output the simulation results for scoring.

[0202] Exemplarily, the first output module includes:

[0203] The sub-item scoring submodule is used to score the reactant consumption, product purity, reaction time and number of contaminated product categories in the reaction molecular dynamics simulation process respectively by using the large language model LLM to obtain scores for each evaluation parameter;

[0204] The weighting submodule is used to use the large language model LLM to weight the scores of each evaluation parameter to synthesize the simulation result score, and output the simulation result score.

[0205] Exemplarily, the reaction condition parameters include reaction temperature, reaction pressure, reactant type and reactant concentration; the first generation module includes:

[0206] A parsing submodule, for parsing the reaction temperature, reaction pressure, reactant type and reactant concentration from the initial text command using the large language model LLM;

[0207] A structure file generation submodule, used to generate a structure file using the large language model LLM according to the type of reactants, wherein the structure file is used to reflect the initial molecular structure and atomic coordinates of the target material generated based on the reaction condition parameters;

[0208] An input file generation submodule is used to generate an input file using the large language model LLM according to the reaction temperature, the reaction pressure and the reactant concentration, wherein the input file is used to define simulation parameters required for performing the reaction molecular dynamics simulation, and the simulation parameters include time step, simulation duration, temperature value and pressure value.

[0209] Exemplarily, the second generating module includes:

[0210] An initial simulation submodule, used to load the structure file into the LAMMPS software, initialize the simulation environment, and configure the physical and chemical environment parameters required for the reaction molecular dynamics simulation;

[0211] The simulation execution submodule is used to execute the reaction molecular dynamics simulation, use the LAMMPS software to perform numerical calculations based on the physical and chemical environmental parameters and the simulation parameters, and generate an output file.

[0212] Exemplarily, the device further comprises:

[0213] A potential problem determination module is used to process the data in the output file using the large language model LLM to determine potential problems in the reaction molecular dynamics simulation process, wherein the data processing method includes but is not limited to at least one of the following: data clustering, trend analysis, and anomaly detection;

[0214] The adjustment module is used to use the large language model LLM to adjust the simulation parameters in the input file according to potential problems in the reaction molecular dynamics simulation process to generate an adjusted input file.

[0215] The embodiment of the present disclosure also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and capable of running on the processor. When the computer program is executed by the processor, the various processes of the above-mentioned embodiment of the optimization method for the material preparation route are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0216] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, devices, and electronic devices. Therefore, the embodiments of the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present disclosure may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0217] The embodiments of the present disclosure are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing terminal device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 A process or multiple processes and / or boxes Figure 1These computer program instructions can also be loaded into a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable terminal device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0218] Although the preferred embodiments of the present disclosure have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present disclosure.

[0219] Finally, it should be noted that, in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "include..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements. The above is a detailed introduction to the optimization method, device and electronic device for a material preparation route provided by the present disclosure. The principles and implementation methods of the present disclosure are explained in this article using specific examples. The description of the above embodiments is only used to help understand the method and its core idea of ​​the present disclosure; at the same time, for those of ordinary skill in the art, according to the idea of ​​the present disclosure, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present disclosure.

Claims

1. A method for optimizing a material preparation route, characterized in that: The method comprises: Inputting an initial text command into a large language model LLM, wherein the initial text command specifies reaction condition parameters for preparing a target material; Using the large language model LLM, generating input files and structure files required for reaction molecular dynamics simulation according to the initial text command; Perform reaction molecular dynamics simulation based on the input file and the structure file to generate an output file; Outputting a simulation result score according to the output file using the large language model LLM, wherein the simulation result score is used to reflect the preparation effect of the target material based on the reaction condition parameters; Based on the simulation result score, the large language model LLM is used to output optimized reaction condition parameters.

2. The method according to claim 1, characterized in that The method further comprises: Generate a parameter extraction text command according to the output file, wherein the parameter extraction text command is used to instruct the large language model LLM to extract evaluation parameters from the output file, wherein the evaluation parameters include reactant consumption, product purity, reaction time, and number of contaminated product categories; Generate a scoring text command according to the evaluation parameters, wherein the scoring text command is used to instruct the large language model LLM to calculate the simulation result score based on the evaluation parameters; Generate an optimization text command according to the input file, the structure file and the simulation result score, wherein the optimization text command is used to instruct the large language model LLM to optimize the reaction condition parameters; A new initial text command is generated according to the optimized reaction condition parameters, and the new initial text command is used to instruct the large language model LLM to generate input files and structure files required for a new round of reaction molecular dynamics simulation.

3. The method according to claim 1, characterized in that After outputting the optimized reaction condition parameters using the large language model LLM based on the simulation result score, the method further includes: According to the optimized reaction condition parameters, the large language model LLM is used to generate input files and structure files required for a new round of reaction molecular dynamics simulation; Based on the input files and structure files required for the new round of reaction molecular dynamics simulation, executing reaction molecular dynamics simulation to generate new output files; Outputting a new simulation result score according to the new output file using the large language model LLM, wherein the new simulation result score is higher than the simulation result score; Based on the new simulation result score, using the large language model LLM to output optimized reaction condition parameters; Execution steps: according to the optimized reaction condition parameters, using the large language model LLM to generate input files and structure files required for a new round of reaction molecular dynamics simulation; until the simulation result score output by the large language model LLM is higher than a preset score threshold.

4. The method according to claim 1, characterized in that: Using the large language model LLM, according to the output file, outputting the simulation result score, including: Extracting evaluation parameters from the output file using the large language model LLM, wherein the evaluation parameters include reactant consumption, product purity, reaction time, and number of contaminated product categories; According to the evaluation parameters, the reaction molecular dynamics simulation process is evaluated using the large language model LLM, and the simulation results are output for scoring.

5. The method according to claim 4, characterized in that According to the evaluation parameters, the reaction molecular dynamics simulation process is evaluated using the large language model LLM, and the simulation results are output for scoring, including: Using the large language model LLM, the reactant consumption, product purity, reaction time and number of contaminated product categories in the reaction molecular dynamics simulation process are scored respectively to obtain scores for each evaluation parameter; The large language model LLM is used to weight the scores of the respective evaluation parameters to synthesize the simulation result score, and the simulation result score is output.

6. The method according to claim 1, characterized in that The reaction condition parameters include reaction temperature, reaction pressure, reactant type and reactant concentration; Using the large language model LLM, the input files and structure files required by the LAMMPS software are generated according to the initial text command, including: parse the reaction temperature, reaction pressure, reactant type and reactant concentration from the initial text command using the large language model LLM; Generate a structure file using the large language model LLM according to the type of reactant, wherein the structure file is used to reflect the initial molecular structure and atomic coordinates of the target material generated based on the reaction condition parameters; According to the reaction temperature, the reaction pressure and the reactant concentration, the large language model LLM is used to generate an input file, wherein the input file is used to define simulation parameters required for performing the molecular dynamics simulation of the reaction, and the simulation parameters include time step, simulation duration, temperature value and pressure value.

7. The method according to claim 6, characterized in that Performing a reaction molecular dynamics simulation based on the input file and the structure file to generate an output file includes: Load the structure file into LAMMPS software, initialize the simulation environment, and configure the physical and chemical environment parameters required for the reaction molecular dynamics simulation; Performing a reaction molecular dynamics simulation, using the LAMMPS software to perform numerical calculations based on the physical and chemical environment parameters and the simulation parameters, and generating an output file.

8. The method according to claim 6, characterized in that The method further comprises: Using the large language model LLM, data processing is performed on the data in the output file to determine potential problems in the reaction molecular dynamics simulation process, wherein the data processing method includes but is not limited to at least one of the following: data clustering, trend analysis, and anomaly detection; The large language model LLM is used to adjust the simulation parameters in the input file according to potential problems in the reaction molecular dynamics simulation process to generate an adjusted input file.

9. A device for optimizing a material preparation route, characterized in that: The device comprises: An input module, used for inputting an initial text command into a large language model LLM, wherein the initial text command specifies reaction condition parameters for preparing a target material; A first generation module is used to generate input files and structure files required for reaction molecular dynamics simulation according to the initial text command by using the large language model LLM; A second generating module, used for performing a reaction molecular dynamics simulation based on the input file and the structure file to generate an output file; A first output module is used to output a simulation result score according to the output file using the large language model LLM, wherein the simulation result score is used to reflect the preparation effect of the target material based on the reaction condition parameters; The second output module is used to output the optimized reaction condition parameters based on the simulation result score using the large language model LLM.

10. An electronic device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program implements the steps of the method according to any one of claims 1 to 8 when executed by the processor.

Images