An ai algorithm for automatically generating a solid-state battery material formula and application

Abstract

The application provides an AI algorithm for automatically generating a solid-state battery material formula and an application, including a solid-state battery material formula generation method, device, equipment, medium and product, and relates to the technical field of battery materials. The method comprises the following steps: generating a plurality of initial parameter groups based on a candidate particle size set and a candidate particle number set; obtaining a particle packing constraint condition; inputting the plurality of initial parameter groups and the particle packing constraint condition into a constraint layer to obtain a plurality of constrained parameter groups that meet the particle packing constraint condition; inputting each of the constrained parameter groups into a battery performance parameter prediction model to obtain battery performance parameters corresponding to each of the constrained parameter groups; determining a target parameter group from the plurality of constrained parameter groups based on the battery performance parameters corresponding to each of the constrained parameter groups, and generating a solid-state battery positive electrode formula based on the target parameter group. The embodiment of the application can optimize the positive electrode material system and improve the overall performance of the battery.

Description

Technical Field

[0001] This application relates to the field of battery materials technology, and in particular to an AI algorithm and application for automatic generation of solid-state battery material formulations, including solid-state battery material formulation generation methods, apparatus, equipment, media and products. Background Technology

[0002] The performance of solid-state batteries is highly dependent on the scientific design of the material system. Its core structure consists of a solid electrolyte, a cathode material, a cathode material, and interface modification materials. The formulation ratio and interactions of different material components directly determine key indicators such as ionic conductivity, cycle stability, energy density, and safety characteristics of the battery. Therefore, the selection and optimization of the material system is a core aspect of solid-state battery research and development.

[0003] In particular, the compaction density of the cathode material in a solid-state battery determines the overall capacity of the battery. However, in practical applications of cathode materials, existing methods often use particulate powder with a relatively uniform particle size distribution as the cathode active material. This makes it difficult for the cathode layer to achieve a higher density after compaction, and the compaction density is difficult to exceed the existing threshold. Consequently, the actual battery capacity is significantly lower than the theoretical design value, resulting in poor battery performance. Summary of the Invention

[0004] This application provides an AI algorithm and application for automatically generating solid-state battery material formulations, including a method, apparatus, equipment, medium, and product for generating solid-state battery material formulations, which can improve the overall performance of the battery.

[0005] A first aspect of this application provides a method for generating a solid-state battery material formulation, comprising:

[0006] Based on the candidate particle size set and the candidate particle number set, multiple initial parameter sets are generated. Each initial parameter set contains the mean particle size and the mean particle number of each type of cathode powder. For each type of cathode powder, the particle size follows a normal distribution around the mean particle size, and the particle number is within the range of the mean particle number. The mean particle size of each type of cathode powder is selected from the candidate particle size set, and the mean particle number of each type of cathode powder is selected from the candidate particle number set.

[0007] Obtain particle packing constraints, which reflect the constraint relationships between the average particle sizes of various positive electrode powder particles and between the average particle numbers of various positive electrode powder particles.

[0008] Multiple initial parameter sets and particle packing constraints are input into the constraint layer to obtain multiple constrained parameter sets that meet the particle packing constraints.

[0009] Each set of constrained parameters is input into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each set of constrained parameters.

[0010] Based on the battery performance parameters corresponding to each constrained parameter group, a target parameter group is determined in each constrained parameter group, and a solid-state battery cathode formula is generated based on the target parameter group.

[0011] Furthermore, this application also proposes that the battery performance parameter prediction model includes a first sub-model, a second sub-model, a third sub-model, and a battery performance parameter prediction sub-model.

[0012] Each constrained parameter set is input into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each constrained parameter set, including:

[0013] Obtain the set of candidate particulate materials;

[0014] For each constrained parameter group, select the particle material of each of the various cathode powder particles from the candidate particle material set, and form an expanded parameter group based on the constrained parameter group and the particle material of each of the various cathode powder particles.

[0015] For each expanded parameter group, the average particle size and the average number of each positive electrode powder particle in the expanded parameter group are input into the first sub-model to obtain the first sub-predicted value.

[0016] The average particle size and particle material of each of the various positive electrode powder particles in the expanded parameter group are input into the second sub-model to obtain the second sub-predicted value.

[0017] The mean number of each type of cathode powder particle in the expanded parameter group and the particle material of each type of cathode powder particle are input into the third sub-model to obtain the third sub-predicted value.

[0018] The first, second, and third sub-predicted values ​​are input into the battery performance parameter prediction sub-model to obtain the battery performance parameters.

[0019] Furthermore, this application also proposes that the battery performance parameter prediction sub-model includes a compaction density prediction sub-model, an ionic conductivity prediction sub-model, a cycle stability prediction sub-model, and an overall battery performance parameter prediction sub-model.

[0020] The first, second, and third sub-predicted values ​​are input into the battery performance parameter prediction sub-model to obtain the battery performance parameters, including:

[0021] Input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the compaction density prediction sub-model to obtain the predicted compaction density;

[0022] Input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the ionic conductivity prediction sub-model to obtain the predicted ionic conductivity;

[0023] Input the first sub-prediction, the second sub-prediction, and the third sub-prediction into the cyclic stability prediction sub-model to obtain the predicted cyclic stability.

[0024] The compaction density, ionic conductivity, and cycle stability are input into the overall battery performance parameter prediction sub-model to obtain the predicted battery performance parameters.

[0025] Furthermore, this application also proposes that the overall battery performance parameter prediction sub-model includes a positive electrode performance parameter prediction sub-model, an electrolyte compatibility prediction sub-model, and a negative electrode compatibility prediction sub-model;

[0026] The compaction density, ionic conductivity, and cycle stability are input into the overall battery performance parameter prediction sub-model to obtain the predicted battery performance parameters, including:

[0027] The compaction density, ionic conductivity, and cycle stability are input into the cathode performance parameter prediction model to obtain the predicted cathode performance parameters.

[0028] The electrolyte compatibility prediction model is obtained by inputting the particle material of various positive electrode powders and the electrolyte material of solid-state batteries into the electrolyte compatibility prediction model.

[0029] By inputting the particle materials of various positive electrode powders and the negative electrode materials of solid-state batteries into the negative electrode fit prediction model, the predicted negative electrode fit is obtained.

[0030] Battery performance parameters are calculated based on positive electrode performance parameters, electrolyte compatibility, and negative electrode compatibility.

[0031] Furthermore, this application also proposes a method for generating a solid-state battery cathode formulation based on a set of target parameters, including:

[0032] From the target parameter set, obtain the particle material of each of the various positive electrode powder particles;

[0033] From the material isomorphism diagram, obtain a group of isomorphic materials with the same microstructure as each of the obtained particulate materials;

[0034] For each particulate material, replace the particulate material with other materials that are different from the particulate material in the isomorphic material group to obtain multiple equivalent parameter groups of the target parameter group;

[0035] By inputting multiple equivalent parameter groups into the battery performance parameter prediction model, the battery performance parameters corresponding to each equivalent parameter group are obtained.

[0036] Based on the battery performance parameters corresponding to each equivalent parameter group, a formula parameter group is determined in each equivalent parameter group, and a solid-state battery cathode formula is generated based on the formula parameter group.

[0037] Furthermore, this application also proposes that the battery performance parameter prediction model be pre-trained in the following manner:

[0038] Multiple solid-state battery cathode samples were obtained, and each solid-state battery cathode sample was made from the average sample particles of various cathode powder particles.

[0039] Measure the mean sample particle size and mean sample particle number for each type of positive electrode powder particles;

[0040] Measure the sample battery performance parameters of each solid-state battery cathode sample;

[0041] The mean sample particle size and mean sample particle number of each type of cathode powder particles are input into the battery performance parameter prediction model to obtain the predicted battery performance parameters of the solid-state battery cathode sample.

[0042] Based on the predicted battery performance parameters and the sample battery performance parameters, a loss function is generated, and a battery performance parameter prediction model is trained based on the loss function.

[0043] Furthermore, this application proposes that the initial parameter set be generated in the following manner:

[0044] Calculate the first population mean and first population standard deviation of the candidate particle size set;

[0045] From the candidate particle size set, the mean particle size of each of the various positive electrode powder particles is selected, such that the average absolute value of the deviation of the mean particle size of each of the various positive electrode powder particles from the standard deviation is less than the first threshold. The deviation of the mean particle size of the positive electrode powder particles from the standard deviation is equal to: the difference between the mean particle size of the positive electrode powder particles and the first population mean, divided by the first population standard deviation.

[0046] Calculate the second population mean and second population standard deviation of the candidate particle number set;

[0047] From the candidate particle number set, select the mean particle number of each of the various positive electrode powder particles, such that the average absolute value of the deviation of the mean particle number of each of the various positive electrode powder particles from the standard deviation is less than the second threshold. The deviation of the mean particle number of positive electrode powder particles from the standard deviation is equal to: the difference between the mean particle number of positive electrode powder particles and the second population mean, divided by the second population standard deviation.

[0048] An initial parameter set is generated based on the average particle size and average number of particles of various positive electrode powders.

[0049] A second aspect of this application provides a solid-state battery, including a positive electrode, a negative electrode, and an electrolyte, wherein the electrolyte is located between the positive electrode and the negative electrode, and the positive electrode is generated based on the solid-state battery material formulation method provided in any of the above aspects.

[0050] A third aspect of this application provides a solid-state battery material formulation generation apparatus, comprising:

[0051] The parameter group generation module is used to generate multiple initial parameter groups based on the candidate particle size set and the candidate particle number set. Each initial parameter group contains the mean particle size and the mean particle number of each type of positive electrode powder. For each type of positive electrode powder, the particle size follows a normal distribution around the mean particle size, and the particle number is within the fluctuation range of the mean particle number. The mean particle size of each type of positive electrode powder is selected from the candidate particle size set, and the mean particle number of each type of positive electrode powder is selected from the candidate particle number set.

[0052] The condition acquisition module is used to acquire particle packing constraint conditions. The particle packing constraint conditions reflect the constraint relationships between the average particle sizes of various positive electrode powder particles and between the average particle numbers of various positive electrode powder particles.

[0053] The parameter group filtering module is used to input multiple initial parameter groups and particle packing constraints into the constraint layer to obtain multiple constrained parameter groups that meet the particle packing constraints.

[0054] The performance prediction module is used to input each constrained parameter group into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each constrained parameter group.

[0055] The formulation generation module is used to determine the target parameter group in each constrained parameter group based on the battery performance parameters corresponding to each constrained parameter group, and generate a solid-state battery cathode formulation based on the target parameter group.

[0056] A fourth aspect of this application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the solid-state battery material formulation generation method as described above.

[0057] A fifth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the solid-state battery material formulation generation method as described above.

[0058] A sixth aspect of this application provides a computer program product comprising a computer program that is read and executed by a processor of a computer device, causing the computer device to perform the solid-state battery material formulation generation method as described above.

[0059] The solid-state battery material formulation generation method provided in this application firstly generates multiple initial parameter sets based on a set of candidate particle sizes and a set of candidate particle numbers. These initial parameter sets cover the average particle size and average particle number of various cathode powder particles. This diverse combination increases the diversity of cathode powder particles, overcomes the limitations of traditional single particle size, and helps improve the compaction density of the cathode layer. Secondly, particle packing constraints are obtained, and a constrained parameter set that meets the conditions is obtained using the constraint layer. This ensures the rationality and stability of particle packing and avoids microstructural defects caused by improper particle combination. Then, the constrained parameter set is input into a battery performance parameter prediction model, which can quickly evaluate the battery performance parameters corresponding to different parameter sets. Finally, the target parameter set is determined based on the battery performance parameters, and a solid-state battery cathode formulation is generated. This allows for the selection of the formulation that optimizes battery performance. Through this systematic approach, considering factors such as particle size, number, and packing constraints, the cathode material system is optimized, thereby effectively improving key performance indicators such as ionic conductivity, cycle stability, and energy density, and enhancing the overall battery performance.

[0060] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0061] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0062] Figure 1 This is a schematic flowchart of a solid-state battery material formulation generation method provided in one embodiment of this application;

[0063] Figure 2 This is a schematic flowchart of S400 provided in one embodiment of this application;

[0064] Figure 3This is a schematic flowchart of S460 provided in one embodiment of this application;

[0065] Figure 4 This is a schematic flowchart of S464 provided in one embodiment of this application;

[0066] Figure 5 This is a schematic flowchart of S500 provided in one embodiment of this application;

[0067] Figure 6 This is a schematic flowchart of S100 provided in one embodiment of this application;

[0068] Figure 7 This is a schematic diagram of the structure of a solid-state battery provided in one embodiment of this application;

[0069] Figure 8 This is a schematic diagram of the solid-state battery material formulation generation apparatus provided in one embodiment of this application;

[0070] Figure 9 This is a schematic diagram of the structure of a solid-state battery material formulation generation device provided in one embodiment of this application. Detailed Implementation

[0071] To further illustrate the technical means and effects adopted by this application to achieve the intended inventive purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation methods, structures, features, and effects of a solid-state battery material formulation generation method, apparatus, equipment, medium, and product proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0072] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0073] It should be noted that the acquisition, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of laws and regulations.

[0074] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0075] In traditional solid-state battery cathode material manufacturing processes, the particle size and number configuration of cathode powder lack a multi-dimensional synergistic optimization mechanism. The unplanned particle size distribution of traditional cathode materials leads to the formation of numerous interstitial pores during particle stacking, resulting in poor compaction density and thus affecting battery performance.

[0076] For example, in solid-state battery systems using lithium nickel cobalt manganese oxide as the positive electrode active material, when the particle size distribution of the positive electrode powder is unreasonable and the overall situation is similar to a single particle size dominance (for example, a large number of particles are concentrated at around 15 micrometers), the measured compaction density only reaches 68% of the theoretical value, resulting in battery performance that is significantly lower than the theoretical design.

[0077] To address the aforementioned issues, this application first analyzes the root cause of particle packing gaps due to unreasonable particle size distribution in traditional processes. By studying the synergistic mechanism of particle size and number distribution, it was found that single-parameter optimization cannot resolve the contradiction between pore filling and mechanical stability. Therefore, this application proposes a multimodal particle system, generating multiple initial parameter sets based on candidate particle size and candidate particle number sets. Each initial parameter set includes the mean particle size and mean particle number for various cathode powder particles. For each type of cathode powder particle, the particle size follows a normal distribution around the mean particle size, and the particle number fluctuates within the mean particle number range. Simultaneously, particle packing constraints are introduced to ensure geometric compatibility of particles with different sizes. Furthermore, a battery performance parameter prediction model is combined to establish a mapping relationship from microscopic particle parameters to macroscopic battery performance, achieving systematic optimization of the material formulation.

[0078] In this regard, such as Figure 1 As shown, this application provides a schematic flowchart of a method for generating a solid-state battery material formulation. This method can be applied to a solid-state battery material formulation generating apparatus or equipment, and may include the following steps S100 to S500:

[0079] S100: Based on the candidate particle size set and the candidate particle number set, generate multiple initial parameter sets. Each initial parameter set contains the average particle size and the average particle number of each type of positive electrode powder. For each type of positive electrode powder, the particle size follows a normal distribution around the average particle size, and the particle number is within the fluctuation range of the average particle number. The average particle size of each type of positive electrode powder is selected from the candidate particle size set, and the average particle number of each type of positive electrode powder is selected from the candidate particle number set.

[0080] S200, obtain particle packing constraint conditions. Particle packing constraint conditions reflect the constraint relationship between the average particle size of various positive electrode powder particles and the average number of particles of various positive electrode powder particles.

[0081] S300: Input multiple initial parameter sets and particle packing constraints into the constraint layer to obtain multiple constrained parameter sets that meet the particle packing constraints.

[0082] S400: Input each constrained parameter group into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each constrained parameter group.

[0083] S500 determines the target parameter group in each constrained parameter group based on the battery performance parameters corresponding to each constrained parameter group, and generates a solid-state battery cathode formula based on the target parameter group.

[0084] In this embodiment, the candidate particle size set refers to the range of possible values ​​for the particle size of the cathode powder in a pre-defined manner. Specifically, it can be achieved by experimental measurement or literature review to determine the particle size range corresponding to different materials, covering the particle size characteristics of different material systems and ensuring the diversity of parameter combinations.

[0085] The candidate particle number set refers to the possible range of values ​​for the number of positive electrode powder particles set in advance. Specifically, it can be determined by statistically analyzing the particle distribution pattern under different ratios or process condition limitations, providing basic data support for parameter generation.

[0086] Normal distribution refers to the symmetrical probability distribution of particle size around the mean. Specifically, it can be achieved by controlling the uniformity of the powder preparation process, simulating the natural fluctuations in actual production, and improving the feasibility of parameter combinations.

[0087] The quantity fluctuation range refers to a numerical interval set around the average particle number. This range limits the degree of fluctuation in the actual particle number of each cathode powder particle relative to the average particle number. By setting upper and lower limits, it ensures that when generating the initial parameter set, the particle number of each cathode powder particle maintains a certain degree of randomness (to simulate fluctuations in actual production) without deviating too much from the average particle number. This ensures the rationality and reliability of subsequent operations such as particle packing simulation and battery performance prediction based on these parameter sets. Assuming the candidate particle number set is {100, 200, 300}, we select an average particle number of 200 for one type of cathode powder particle. The quantity fluctuation range is set to ±20% of the average particle number, meaning the particle number of this cathode powder particle falls within the fluctuation range of 160 to 240. When generating the initial parameter set, the particle number of this cathode powder particle can be any integer within this fluctuation range, such as 180, 200, 220, etc.

[0088] Particle packing constraints refer to the physical compatibility rules between different particle sizes and particle numbers. Specifically, they can be defined using particle gradation theory or packing density models to avoid structural defects caused by excessively large or small particle gaps due to parameter combinations.

[0089] The constraint layer refers to the algorithm module that filters parameter groups that meet the particle packing conditions. Specifically, a rule engine or optimization algorithm can be used to eliminate parameter combinations that do not meet the constraints, ensuring that the generated parameters meet the actual process requirements.

[0090] Battery performance parameter prediction models are mathematical models that predict battery performance indicators based on parameter combinations. Specifically, they can establish a mapping relationship between input parameters and performance by training historical data through machine learning, and quickly evaluate the potential effects of parameter combinations.

[0091] The core innovation of this application lies in the systematic optimization of the particle gradation design of cathode materials by combining and constraining the generation of multimodal particle parameters and combining them with performance prediction models, thereby overcoming the problem of insufficient contact area caused by single particle size and improving compaction density and battery performance.

[0092] For example, an example is provided in which the particle size of the positive electrode powder particles in the initial parameter set is normally distributed around the mean particle size, and the number of particles is within the range of the mean particle number.

[0093] In the design of solid-state battery cathode material formulations, we selected three types of cathode powder particles, denoted as particle A, particle B, and particle C. Through preliminary research and the selection of candidate particle size and number sets, the average particle size and average particle number for each type were determined. Particle A has an average particle size of 10 μm and an average particle number of 100. Particle B has an average particle size of 15 μm and an average particle number of 80. Particle C has an average particle size of 5 μm and an average particle number of 120.

[0094] Simultaneously, standard deviations were set for each type of particle: 1 μm for particle A, 1.5 μm for particle B, and 0.8 μm for particle C. The quantity of each type of particle was set to fluctuate within ±10% of the mean particle count.

[0095] For particle A, according to the normal distribution, most particle sizes will be concentrated in the range of 10 ± 1 μm. For example, the actual generated particle A size might be 9.2 μm, 10.5 μm, 9.8 μm, etc., these values ​​are distributed around the mean of 10 μm, and the closer the value is to the mean, the higher the probability of its occurrence. The size of particle B is mainly distributed in the range of 15 ± 1.5 μm, such as 14 μm, 15.3 μm, 13.8 μm, etc., which are possible particle sizes and follow the normal distribution law. The size of particle C is concentrated in the range of 5 ± 0.8 μm, such as 4.5 μm, 5.2 μm, 4.8 μm, etc., which will occur with a relatively high probability.

[0096] The number of particles A will fluctuate between 90 and 110, the number of particles B will fluctuate between 72 and 88, and the number of particles C will fluctuate between 108 and 132.

[0097] As an example, first, a set of candidate particle sizes is established, including five particle sizes: 5μm, 10μm, 15μm, 20μm, and 25μm. A set of candidate particle numbers is also established, including five numbers: 1000, 2000, 3000, 4000, and 5000.

[0098] An initial parameter set is generated based on the above set. For example, an initial parameter set may contain three types of positive electrode powder particles: the first type has an average particle size of 10 μm and an average particle number of 3000; the second type has an average particle size of 20 μm and an average particle number of 2000; and the third type has an average particle size of 5 μm and an average particle number of 4000.

[0099] Next, obtain the particle packing constraints. For example, specify that the ratio of the largest to the smallest particle size should not exceed 5, and the total number of particles should be between 5000 and 10000. Input the initial parameter set and particle packing constraints into the constraint layer, and filter out the constrained parameter sets that meet the particle packing constraints.

[0100] A battery performance parameter prediction model is then established, with the average particle size and average particle number as input variables, and battery performance parameters such as compaction density and ionic conductivity as output variables. The constrained parameter set is then input into the battery performance parameter prediction model to obtain the predicted battery performance parameters.

[0101] Finally, based on the predicted battery performance parameters, the set of constrained parameters that yields the best battery performance is selected as the target parameter set. For example, the set of constrained parameters with the highest compaction density and the highest ionic conductivity is selected as the target parameter set. The solid-state battery cathode formulation is then generated based on the target parameter set, and the proportions of cathode powders with various particle sizes are determined.

[0102] In this embodiment, firstly, multiple initial parameter sets are generated based on candidate particle size and candidate particle number sets, covering the average particle size and average particle number of various cathode powder particles. This diverse combination increases the diversity of cathode powder particles, overcoming the limitations of traditional single particle size and helping to improve the compaction density of the cathode layer. Secondly, particle packing constraints are obtained, and a constrained parameter set that meets the conditions is obtained using the constraint layer. This ensures the rationality and stability of particle packing and avoids microstructural defects caused by improper particle combination. Then, the constrained parameter set is input into the battery performance parameter prediction model, which can quickly evaluate the battery performance parameters corresponding to different parameter sets. Finally, the target parameter set is determined based on the battery performance parameters, and a solid-state battery cathode formulation is generated, allowing for the selection of the formulation that optimizes battery performance. Through this systematic approach, considering factors such as particle size, number, and packing constraints, the cathode material system is optimized, thereby effectively improving key performance indicators such as ionic conductivity, cycle stability, and energy density of the battery, and enhancing the overall battery performance.

[0103] In some of the schemes described above in this application, the initial parameter set generation and constraint screening process did not consider the influence of the type of cathode powder material on battery performance, resulting in the battery performance parameter prediction model being unable to accurately reflect the synergistic effect of different material combinations on battery performance parameters such as compaction density and ionic conductivity.

[0104] In this regard, such as Figure 2 As shown, this application further proposes a battery performance parameter prediction model that includes a first sub-model, a second sub-model, a third sub-model, and a battery performance parameter prediction sub-model.

[0105] S400 may specifically include the following S410 to S460:

[0106] S410, Obtain the candidate particulate material set;

[0107] S420, for each constrained parameter group, selects the particle material of each of the various positive electrode powder particles from the candidate particle material set, and forms an expanded parameter group based on the constrained parameter group and the particle material of each of the various positive electrode powder particles;

[0108] S430, for each expanded parameter group, input the average particle size of each of the various positive electrode powder particles in the expanded parameter group and the average number of each of the various positive electrode powder particles into the first sub-model to obtain the first sub-predicted value;

[0109] S440: Input the mean particle size and particle material of each of the various positive electrode powder particles in the expanded parameter group into the second sub-model to obtain the second sub-predicted value.

[0110] S450: Input the average number of particles of various positive electrode powder particles in the expanded parameter group and the particle material of various positive electrode powder particles into the third sub-model to obtain the third sub-predicted value.

[0111] S460 inputs the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the battery performance parameter prediction sub-model to obtain the battery performance parameters.

[0112] In this embodiment, the first sub-model is a component of the battery performance parameter prediction model. It receives the average particle size and average number of various cathode powder particles from the expanded parameter set as input, and after internal calculation and processing, outputs the first sub-predicted value, which is used to analyze the specific influence relationship between particle size and particle number on battery performance.

[0113] The second sub-model is also one of the sub-models of the battery performance parameter prediction model. Its input is the mean particle size and particle material of various cathode powder particles in the expanded parameter set, and the output is the second sub-predicted value, which is used to analyze how the two factors of particle size and particle material jointly affect battery performance.

[0114] The third sub-model is also one of the sub-models of the battery performance parameter prediction model. Its input is the mean particle number and particle material of various positive electrode powder particles in the expanded parameter set, and the output is the third sub-predicted value, which is used to analyze the effect mechanism of particle number and particle material on battery performance.

[0115] The candidate particulate material set includes positive electrode active materials such as lithium nickel manganese cobalt oxide, lithium iron phosphate, or lithium cobalt oxide. The expanded parameter set combines material properties with particle distribution parameters to form input data with both material and structural characteristics. The first sub-model handles the combined effect of particle size and number, the second sub-model analyzes the matching relationship between material type and particle size, and the third sub-model evaluates the suitability of material type and particle number. The battery performance parameter prediction sub-model establishes a nonlinear mapping relationship between multidimensional features and performance parameters by integrating the outputs of the three sub-models.

[0116] As an example, the first step is to obtain a set of candidate particulate materials. This set of candidate particulate materials includes common cathode materials such as lithium nickel manganese cobalt oxide, lithium iron phosphate, or lithium cobalt oxide.

[0117] Then, for each constrained parameter group, the particle material for each of the various cathode powder particles is selected from the candidate particle material set. For example, lithium nickel manganese cobalt oxide can be selected as the material for the first cathode powder particle, and lithium iron phosphate as the material for the second cathode powder particle. An expanded parameter group is formed based on the constrained parameter group and the selected particle materials.

[0118] Next, for each expanded parameter group, the average particle size and average number of various cathode powder particles in the expanded parameter group are input into the first sub-model to obtain the first sub-predicted value. The first sub-model can be a neural network model used to predict the impact of particle size and particle number on battery performance.

[0119] The mean particle size and material of each of the various cathode powder particles in the expanded parameter set are input into the second sub-model to obtain the second sub-prediction value. The second sub-model can be a machine learning model used to predict the impact of particle size and material on battery performance.

[0120] The mean particle number and particle material of each of the various cathode powder particles in the expanded parameter set are input into the third sub-model to obtain the third sub-predicted value. The third sub-model can be a regression model used to predict the impact of particle number and material on battery performance.

[0121] Finally, the first, second, and third sub-predicted values ​​are input into the battery performance parameter prediction sub-model to obtain the battery performance parameters. The battery performance parameter prediction sub-model can be an ensemble model that comprehensively considers the prediction results of the first three sub-models to provide the final battery performance parameter prediction.

[0122] This embodiment comprehensively considers the influence of cathode powder particle size, number, and material on battery performance, improving the accuracy of battery performance parameter prediction. By employing stepwise and ensemble prediction methods, it fully utilizes the interactions between different features, avoiding the problem of a single model potentially overlooking certain important features. Furthermore, this method incorporates material information by expanding the parameter set, making the prediction results more comprehensive and reliable, providing strong support for the optimization of solid-state battery cathode formulations.

[0123] In some of the solutions described above in this application, the battery performance parameters are predicted using only a single model for overall evaluation, which cannot be broken down into specific performance indicators of multiple performance dimensions such as compaction density, ionic conductivity, and cycle stability. This makes it impossible to adjust the combination of material parameters for different performance dimensions.

[0124] In this regard, such as Figure 3 As shown, this application further proposes a battery performance parameter prediction sub-model, including a compaction density prediction sub-model, an ionic conductivity prediction sub-model, a cycle stability prediction sub-model, and an overall battery performance parameter prediction sub-model;

[0125] S460 specifically includes the following S461 to S464:

[0126] S461, input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the compaction density prediction sub-model to obtain the predicted compaction density;

[0127] S462, input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the ionic conductivity prediction sub-model to obtain the predicted ionic conductivity;

[0128] S463, input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the cyclic stability prediction sub-model to obtain the predicted cyclic stability;

[0129] S464 inputs compaction density, ionic conductivity, and cycle stability into the overall battery performance parameter prediction sub-model to obtain the predicted battery performance parameters.

[0130] In this embodiment, the compaction density prediction sub-model is configured to output compaction density based on the mean particle size and number distribution data of the cathode powder particles by calculating the relationship between particle packing efficiency and porosity; the ionic conductivity prediction sub-model is configured to output ionic conductivity by combining particle material properties and particle size distribution data and simulating the connectivity of ion transport paths; the cycle stability prediction sub-model is configured to output cycle stability by combining particle material properties and number distribution data and analyzing the stress distribution at the particle interface; and the overall battery performance parameter prediction sub-model is configured to weightedly fuse compaction density, ionic conductivity, and cycle stability to generate a comprehensive performance score.

[0131] Specifically, the first sub-predicted value includes particle size and number distribution characteristics, the second sub-predicted value includes particle material and size correlation characteristics, and the third sub-predicted value includes particle material and number correlation characteristics. When these features are input into the compaction density prediction sub-model, the model calculates the packing density under different particle size combinations to generate compaction density values. For example, when the particle size distribution meets a multi-peak distribution, the compaction density increases to more than 95% of the theoretical maximum value. When features are input into the ionic conductivity prediction sub-model, the model simulates the migration path length of ions in the interparticle spaces based on the matching relationship between particle material type and size to generate conductivity values. For example, when the particulate material is lithium cobalt oxide and the size gradient is 0.5-5 micrometers, the conductivity increases to the order of 10^-3 S / cm. When the feature is input into the cycle stability prediction sub-model, the model calculates the stress concentration coefficient at the particle interface during charging and discharging based on the particle number distribution and material mechanical strength, and generates a stability value. For example, when the standard deviation of the particle number is less than 10% and the material is an oxide, the cycle stability exceeds 500 cycles. Finally, the overall battery performance parameter prediction sub-model integrates the above three indicators according to the weight coefficients to generate a comprehensive score, which is used to screen the optimal parameter combination, thereby solving the problem of formulation optimization limitations caused by a single evaluation dimension.

[0132] As an example, the first, second, and third sub-predictions are input into the compaction density prediction sub-model to obtain the predicted compaction density. The compaction density prediction sub-model can employ a multilayer perceptron structure containing three hidden layers, with 64, 32, and 16 neurons in each layer, respectively.

[0133] The first, second, and third sub-predictions are input into the ionic conductivity prediction sub-model to obtain the predicted ionic conductivity. The ionic conductivity prediction sub-model can employ a convolutional neural network structure, containing two convolutional layers and two fully connected layers.

[0134] The first, second, and third sub-predictions are input into the cyclic stability prediction sub-model to obtain the predicted cyclic stability. The cyclic stability prediction sub-model can employ a long short-term memory (LSTM) network structure, comprising one LSM network layer and one fully connected layer.

[0135] Finally, the compaction density, ionic conductivity, and cycle stability are input into the overall battery performance parameter prediction sub-model to obtain the predicted battery performance parameters. The overall battery performance parameter prediction sub-model can adopt a random forest structure containing 100 decision trees.

[0136] In this embodiment, multiple sub-predicted values ​​are input into different specialized prediction models to obtain intermediate parameters such as compaction density, ionic conductivity, and cycle stability. Overall battery performance is then predicted based on these intermediate parameters, improving the accuracy and interpretability of the prediction. This overcomes the limitations of directly predicting overall performance parameters and better captures the impact of different factors on battery performance through a step-by-step prediction approach.

[0137] In some of the schemes described above in this application, the overall battery performance parameter prediction sub-model is based only on compaction density, ionic conductivity and cycle stability, without considering the compatibility between the cathode material and the electrolyte and anode material. This results in the prediction results failing to reflect the compatibility of the interfacial reactions between materials, affecting the accuracy of battery performance evaluation.

[0138] In this regard, such as Figure 4 As shown, this application further proposes a sub-model for predicting overall battery performance parameters, including a positive electrode performance parameter prediction sub-model, an electrolyte compatibility prediction sub-model, and a negative electrode compatibility prediction sub-model.

[0139] S464 specifically includes the following S4641 to S4644:

[0140] S4641 inputs compaction density, ionic conductivity, and cycle stability into the cathode performance parameter prediction model to obtain the predicted cathode performance parameters;

[0141] S4642 inputs the particle material of various positive electrode powder particles and the electrolyte material of solid-state battery into the electrolyte compatibility prediction model to obtain the predicted electrolyte compatibility.

[0142] S4643 inputs the particle material of various positive electrode powders and the negative electrode material of solid-state batteries into the negative electrode fit prediction model to obtain the predicted negative electrode fit.

[0143] S4644 calculates battery performance parameters based on positive electrode performance parameters, electrolyte compatibility, and negative electrode compatibility.

[0144] In this embodiment, the cathode performance parameter prediction model converts compaction density, ionic conductivity and cycle stability into cathode performance parameters through linear weighting or nonlinear mapping, for example, by using a three-layer fully connected neural network to achieve feature fusion.

[0145] Electrolyte compatibility prediction uses a material chemistry compatibility database to match the interfacial reaction activity parameters of the cathode particle material and the electrolyte material, for example, using the lattice matching degree between lithium cobalt oxide and sulfide electrolyte as input features.

[0146] The anode compatibility prediction model calculates the difference in expansion coefficients and charge transfer impedance between the cathode and anode materials. For example, the support vector machine model is used to predict the interface stability between graphite anode and lithium nickel manganese oxide cathode.

[0147] Specifically, the positive electrode performance parameter prediction first receives the values ​​of compaction density, ionic conductivity, and cycle stability from the model, and normalizes them using an activation function to generate a positive electrode performance parameter score in the 0-1 range. Electrolyte compatibility prediction uses thermodynamic data from a materials database to calculate the volume change difference and interfacial impedance increment between the positive electrode particles and the electrolyte during charge and discharge, outputting an electrolyte compatibility score. Negative electrode compatibility prediction extracts the crystal structure parameters of the positive and negative electrode materials from the model, extracts interfacial contact features using a convolutional neural network, and generates a negative electrode compatibility score. Finally, the overall battery performance parameters are obtained by multiplying the positive electrode performance score, electrolyte compatibility score, and negative electrode compatibility score.

[0148] As an example, the cathode performance parameter prediction model inputs compaction density, ionic conductivity, and cycle stability to obtain the predicted cathode performance parameters. The cathode performance parameter prediction model adopts a neural network structure. The input layer contains three neurons for compaction density, ionic conductivity, and cycle stability, the hidden layer contains 10 neurons, and the output layer is a single neuron that outputs the cathode performance parameters.

[0149] The electrolyte fit prediction model inputs the particle materials of various cathode powders and the electrolyte materials of solid-state batteries to obtain the predicted electrolyte fit. The electrolyte fit prediction model uses a support vector machine algorithm, with input features being a combination of cathode powder particle materials and electrolyte materials, and outputting an electrolyte fit score.

[0150] The anode material of various positive electrode powders and the anode material of the solid-state battery are input into the anode fit prediction model to obtain the predicted anode fit. The anode fit prediction model adopts the random forest algorithm, and the input features are the combined encoding of the positive electrode powder particles and the anode material, and the output is the anode fit score.

[0151] Finally, based on the positive electrode performance parameters, electrolyte compatibility, and negative electrode compatibility, the battery performance parameters are calculated. Specifically, the positive electrode performance parameters, electrolyte compatibility, and negative electrode compatibility are used as inputs, and the final battery performance parameters are obtained by weighted summation, with weighting coefficients of 0.5, 0.3, and 0.2, respectively.

[0152] This embodiment achieves comprehensive prediction of solid-state battery performance. By establishing separate prediction models for positive electrode performance, electrolyte compatibility, and negative electrode compatibility, and combining them, the interactions between various battery components can be fully considered, improving the accuracy and reliability of battery performance prediction. Furthermore, this method has strong versatility and can be applied to the performance prediction of solid-state batteries with different material systems, providing effective guidance for the optimized design of solid-state battery materials.

[0153] In some of the solutions described above in this application, when generating solid-state battery cathode formulations based on target parameter sets, only initially selected particulate materials are used for formulation construction, resulting in a limited range of material selection and an inability to optimize formulation performance through substitution of similar materials.

[0154] In this regard, such as Figure 5 As shown, this application further proposes that S500 specifically includes the following S510 to S550:

[0155] S510: Obtain the particle material of various positive electrode powder particles from the target parameter set;

[0156] S520, From the material isomorphism diagram, obtain a group of isomorphic materials with the same microstructure as each of the obtained particulate materials;

[0157] S530, for each particulate material, replace the particulate material with other materials in the isomorphic material group that are different from the particulate material to obtain multiple equivalent parameter groups of the target parameter group;

[0158] S540 inputs multiple equivalent parameter groups into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each equivalent parameter group.

[0159] S550 determines the formula parameter group in each equivalent parameter group based on the battery performance parameters corresponding to each equivalent parameter group, and generates a solid-state battery cathode formula based on the formula parameter group.

[0160] In this embodiment, the material isomorphism map stores data on material groups with the same crystal structure or chemical bonding mode. Each isomorphic material group contains three or more materials with similar ion migration paths. During the generation of the isomorphic parameter group, the replacement operation for each particulate material is performed independently. The battery performance parameter prediction model compares the predicted battery performance parameters of the isomorphic parameter group with the original predicted values ​​of the target parameter group, and selects the isomorphic parameter group with the best battery performance parameters as the formulation parameter group.

[0161] Specifically, the material isomorphism map is constructed using a crystal database, where lithium cobalt oxide and lithium nickel oxide are grouped into isomorphic material groups due to their identical layered structures. When a particle material in the target parameter group is lithium cobalt oxide, lithium nickel oxide is automatically retrieved as a substitute material to generate an equivalent parameter group. The battery performance parameter prediction model performs parallel calculations of compaction density, ionic conductivity, and cycle stability for each equivalent parameter group. By comparing the predicted battery performance parameters of each equivalent parameter group with the original battery performance parameters of the target parameter group, the equivalent parameter group with the best battery performance parameters is finally selected as the formulation parameter group.

[0162] As an example, we first obtain the particle materials corresponding to various cathode powder particles from the target parameter set. For example, the particle materials for the three cathode powder particles obtained are lithium cobalt oxide, lithium nickel oxide, and spinel-type lithium manganese oxide.

[0163] Then, from the material isomorphism diagram, isomorphic material groups with the same microstructure as each type of particle material were obtained. Specifically, for lithium cobalt oxide, isomorphic material groups including lithium nickel oxide and lithium manganese oxide were obtained; for lithium nickel oxide, isomorphic material groups including lithium cobalt oxide and lithium manganese oxide were obtained; for spinel-type lithium manganese oxide, isomorphic material groups including lithium dicobalt oxide and lithium dinickel oxide were obtained.

[0164] For each type of particulate material, the particulate material is replaced with other materials from the isomorphic material group that are different from the particulate material itself, resulting in multiple equivalent parameter groups for the target parameter group. Specifically, lithium cobalt oxide can be replaced with lithium nickel oxide and lithium manganese oxide to obtain two equivalent parameter groups respectively; lithium nickel oxide can be replaced with lithium cobalt oxide and lithium manganese oxide to obtain two equivalent parameter groups respectively; and spinel-type lithium manganese oxide can be replaced with lithium dicobalt oxide and lithium dinickel oxide to obtain two equivalent parameter groups respectively, thus obtaining a total of 6 equivalent parameter groups.

[0165] Then, multiple equivalent parameter sets are input into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each equivalent parameter set. For example, for each equivalent parameter set, the corresponding battery performance parameters such as compaction density, ionic conductivity, and cycle stability are predicted.

[0166] Finally, based on the battery performance parameters corresponding to each equivalent parameter group, a formulation parameter group is determined within each equivalent parameter group, and a solid-state battery cathode formulation is generated based on the formulation parameter group. Specifically, the equivalent parameter group with the optimal battery performance parameters is selected as the formulation parameter group, and a solid-state battery cathode formulation is generated according to the particulate material, particle size, and particle number in the formulation parameter group.

[0167] This embodiment optimizes the cathode formulation for solid-state batteries. This allows for the exploration of more potential material combinations while maintaining microstructural similarity, leading to a higher-performance cathode formulation. Furthermore, this approach, combining isomorphic material substitution and performance prediction, efficiently screens for the optimal material combination, avoiding extensive experimental verification and improving the development efficiency of solid-state battery cathode materials.

[0168] In some of the solutions described above in this application, an initial parameter set is generated and filtered before being input into a battery performance parameter prediction model to determine the target parameter set. However, the prediction accuracy of the battery performance parameter prediction model directly affects the reliability of the generated formula. If the model has not been effectively trained, the prediction results may be biased, resulting in the final formula performance not meeting expectations.

[0169] To address this, this application further proposes a pre-training method for the battery performance parameter prediction model:

[0170] Multiple solid-state battery cathode samples were obtained, and each solid-state battery cathode sample was made from the average sample particles of various cathode powder particles.

[0171] Measure the mean sample particle size and mean sample particle number for each type of positive electrode powder particles;

[0172] Measure the sample battery performance parameters of each solid-state battery cathode sample;

[0173] The mean sample particle size and mean sample particle number of each type of cathode powder particles are input into the battery performance parameter prediction model to obtain the predicted battery performance parameters of the solid-state battery cathode sample.

[0174] Based on the predicted battery performance parameters and the sample battery performance parameters, a loss function is generated, and a battery performance parameter prediction model is trained based on the loss function.

[0175] In this embodiment, the solid-state battery cathode samples used in the training process are obtained through actual preparation. The average sample particle size and average sample particle number of the cathode powder particles are accurately collected through physical measurement methods, and the sample battery performance parameters are obtained through standardized testing procedures. During model training, the input parameters and measured performance parameters form a mapping relationship. The loss function calculates the error by comparing the difference between the predicted and measured values, and adjusts the model parameters in reverse to optimize prediction accuracy. The training data covers different combinations of particle sizes and numbers, as well as corresponding performance parameters, ensuring that the model can learn the nonlinear relationships between multiple factors. For example, the particle size distribution and number ratio can be adjusted during sample preparation to generate cathode samples with different compaction densities and ionic conductivity, thereby expanding the diversity of the training data.

[0176] Specifically, the training process establishes the correlation between model input and output using actual sample data. During model training, the mean sample particle size and mean sample particle number of each type of cathode powder are used as input features, and the corresponding sample battery performance parameters are used as the target output. Iterative optimization is employed to make the model's predicted values ​​approximate the measured values. The loss function uses mean squared error or cross-entropy, and the model weight parameters are updated based on the error gradient. After training, the model can accurately predict the corresponding battery performance parameters based on the input particle parameter combinations, thus selecting the optimal parameter set in the subsequent formulation generation process. For example, when different particle size combinations are input, the model can infer the corresponding trend of ionic conductivity changes based on similar samples in the training data, thereby evaluating the potential performance advantages and disadvantages of different formulations. In this way, the model training process solves the prediction accuracy problem, ensuring the practical feasibility of the generated solid-state battery cathode formulation.

[0177] As an example, multiple solid-state battery cathode samples are obtained, each sample being made from the average sample particles of various cathode powder particles. For example, 100 solid-state battery cathode samples are obtained, each containing 3 different cathode powder particles.

[0178] Next, the mean sample particle size and mean sample particle number of each type of positive electrode powder were measured. Specifically, a laser particle size analyzer was used to measure the particle size, and an electron microscope was used to observe and count the number of particles.

[0179] Next, the sample battery performance parameters of each solid-state battery cathode sample were measured. Furthermore, parameters such as battery compaction density, ionic conductivity, and cycle stability were tested using an electrochemical workstation.

[0180] The mean particle size and mean number of cathode powder particles for each sample are then input into the battery performance parameter prediction model to obtain the predicted battery performance parameters for the solid-state battery cathode sample. The battery performance parameter prediction model can be a machine learning model based on a neural network.

[0181] Finally, based on the predicted battery performance parameters and the sample battery performance parameters, a loss function is generated, and a battery performance parameter prediction model is trained based on the loss function. Thus, the model parameters can be continuously optimized using the backpropagation algorithm to improve prediction accuracy.

[0182] This embodiment effectively improves the accuracy and reliability of battery performance parameter prediction models. By training the model using actual measured sample data, the complex relationship between the properties of solid-state battery cathode materials and battery performance can be better captured. This data-driven approach reduces the bias of human experience-based judgments, providing more scientific and objective guidance for the optimized design of solid-state battery materials. Simultaneously, by continuously accumulating and updating sample data, the model's predictive ability can be continuously improved, adapting to the development of new materials and processes. The establishment and optimization of this prediction model provides strong technical support for the rapid development and performance improvement of solid-state batteries.

[0183] In some of the above-mentioned schemes in this application, the generation process of the initial parameter set has the problem of unreasonable distribution of candidate particle size and number. The average particle size and number of the randomly selected particles may deviate from the physical constraints of the actual material system, resulting in a large number of initial parameter sets failing to pass the subsequent constraint layer screening, thus reducing the efficiency of formula generation.

[0184] In this regard, such as Figure 6 As shown, this application further proposes that S100 includes the following S110 to S150:

[0185] S110, calculate the first population mean and first population standard deviation of the candidate particle size set;

[0186] S120, from the candidate particle size set, select the mean particle size of each of the various positive electrode powder particles, such that the average absolute value of the deviation of the mean particle size of each of the various positive electrode powder particles from the standard deviation is less than the first threshold, wherein the deviation of the mean particle size of the positive electrode powder particles from the standard deviation is equal to: the difference between the mean particle size of the positive electrode powder particles and the first population mean, divided by the first population standard deviation;

[0187] S130, calculate the second population mean and second population standard deviation of the candidate particle number set;

[0188] S140, from the candidate particle number set, select the mean particle number of each of the various positive electrode powder particles, such that the average absolute value of the deviation of the mean particle number of each of the various positive electrode powder particles from the standard deviation is less than the second threshold, wherein the deviation of the mean particle number of positive electrode powder particles from the standard deviation is equal to: the difference between the mean particle number of positive electrode powder particles and the second population mean, divided by the second population standard deviation.

[0189] S150 generates an initial parameter set based on the average particle size and average number of particles of various positive electrode powders.

[0190] In this embodiment, the statistical parameters of the candidate particle size set are calculated using a normal distribution model. The first population mean is the arithmetic mean of the candidate particle size set, and the first population standard deviation reflects the dispersion of the candidate particles. For the candidate particle number set, the second population mean is the average of the candidate number values, and the second population standard deviation characterizes the fluctuation range of the number distribution. When selecting the mean particle size, the average absolute value of the deviation of each particle from the standard deviation is required to not exceed a first threshold. For example, when the candidate particle size set is 1-10 micrometers, the first threshold can be set to 1.5 times the standard deviation. When selecting the mean particle number, the average absolute value of the deviation is limited to a second threshold. For example, when the candidate number set is 1000-5000, the second threshold can be set to 1.2 times the standard deviation.

[0191] Specifically, the candidate particle size set was obtained by measuring particle sizes in historical formulations, including typical values ​​of 5 micrometers, 8 micrometers, and 12 micrometers. The calculated first population mean was 8.33 micrometers, and the first population standard deviation was 2.87 micrometers. When selecting the mean particle size of two cathode powder particles, if 5 micrometers and 12 micrometers were selected respectively, the deviation from the standard deviation was approximately (5-8.33) / 2.87≈-1.16 and (12-8.33) / 2.87≈1.28, with an average absolute value of 1.22. When the first threshold was set to 1.5, this combination met the screening criteria. The candidate particle number set included 2000, 3000, and 4000 particles, with a second population mean of 3000 and a standard deviation of 816.5. If 2500 and 3500 particles are selected as the mean particle number, the deviation from the standard deviation is approximately (2500-3000) / 816.5≈-0.61 and (3500-3000) / 816.5≈0.61, respectively, with an average absolute value of 0.61, which is lower than the second threshold of 1.2. By controlling the degree of deviation, the initial parameter set is ensured to maintain both a reasonable statistical distribution and sufficient diversity, thereby improving the effectiveness of the constraint layer screening.

[0192] As an example, we first calculate the first population mean and the first population standard deviation for the candidate particle size set. For example, if the candidate particle size set is {1μm, 2μm, 5μm, 10μm, 20μm}, the calculated first population mean is 7.6μm and the first population standard deviation is 7.8μm.

[0193] From the candidate particle size set, the mean particle size of various cathode powder particles is selected, such that the average absolute value of the deviation of the mean particle size of each cathode powder particle from the standard deviation is less than a first threshold. Specifically, the deviation of the mean particle size of the cathode powder particles from the standard deviation is equal to the difference between the mean particle size of the cathode powder particles and the first overall mean, divided by the first overall standard deviation. For example, selecting 2μm, 5μm, and 10μm as the mean particle sizes of three cathode powder particles yields standard deviations of -0.72, -0.33, and 0.31, respectively, with an average absolute value of 0.45, which is less than the set first threshold of 0.5.

[0194] Next, calculate the second population mean and second population standard deviation for the candidate particle number set. For example, if the candidate particle number set is {1000, 2000, 5000, 10000, 20000}, the calculated second population mean is 7600 and the second population standard deviation is 7800.

[0195] From the candidate particle count set, the mean particle counts of various cathode powder particles are selected, ensuring that the average absolute value of the deviation of the mean particle count of each cathode powder particle from the standard deviation is less than a second threshold. Specifically, the deviation of the mean particle count of cathode powder particles from the standard deviation is equal to the difference between the mean particle count of cathode powder particles and the second population mean, divided by the second population standard deviation. For example, selecting 2000, 5000, and 10000 as the mean particle counts of three cathode powder particles yields deviations of -0.72, -0.33, and 0.31, respectively, with an average absolute value of 0.45, which is less than the set second threshold of 0.5.

[0196] Finally, an initial parameter set is generated based on the average particle size and average particle number of various cathode powder particles. For example, the initial parameter set is generated as {(2μm, 2000), (5μm, 5000), (10μm, 10000)}.

[0197] This embodiment generates a reasonable set of initial parameters. Therefore, it allows for the selection of uniformly distributed combinations of particle size and number within the range of candidate particle size and candidate particle number sets, avoiding the generation of extreme parameter combinations. Furthermore, by controlling the average deviation from the standard deviation, it ensures that the generated initial parameter set is representative, providing a good starting point for subsequent optimization processes.

[0198] Based on the solid-state battery material formulation preparation method provided above in this application, accordingly, as... Figure 7As shown, this application further provides a schematic diagram of a solid-state battery structure, including a positive electrode 710, a negative electrode 720, and an electrolyte 730. The electrolyte 730 is located between the positive electrode 710 and the negative electrode 720. The positive electrode 710 is generated based on any of the above-mentioned solid-state battery material formulation methods.

[0199] The positive electrode 710 is generated through a formulation containing various positive electrode powder particles. This formulation generates an initial parameter set by using a set of candidate particle sizes and a set of candidate particle numbers. A constrained parameter set is then selected based on particle packing constraints. Finally, the target parameter set is determined after evaluation using a battery performance parameter prediction model. The electrolyte 730 fills the gap between the positive electrode 710 and the negative electrode 720 in solid form. The negative electrode 720 is bonded to the electrolyte 730 through physical contact.

[0200] Specifically, the positive electrode 710 material layer is composed of a mixture of positive electrode powder particles of various sizes, with both particle size and number following a normal distribution. During the preparation of the positive electrode 710, parameter combinations that satisfy space-filling efficiency are selected through particle packing constraints, and then the compaction density, ionic conductivity, and cycle stability are evaluated using a predictive model. The electrolyte 730 material is hot-pressed at high temperature to form a continuous and dense interface with the positive electrode 710. The negative electrode 720 uses lithium metal foil, which is bonded to the electrolyte 730 without gaps through a roll forming process. When the battery is charged and discharged, the hierarchical porous structure formed by particles of different sizes in the positive electrode 710 promotes the uniform diffusion of lithium ions in the electrolyte, and the tight contact interface generated by the multi-scale particle packing reduces charge transfer resistance.

[0201] This embodiment effectively solves the problem of insufficient contact area between traditional single-size cathode materials and the electrolyte. By optimizing the combination ratio of multi-size cathode powders, particles of different sizes form a complementary filling structure during compaction, significantly increasing the compaction density of the cathode layer while maintaining porosity. This optimizes the lithium-ion transport channels and improves battery performance.

[0202] Based on the solid-state battery material formulation generation method provided in this application, correspondingly, this application also provides specific embodiments of a solid-state battery material formulation generation apparatus.

[0203] like Figure 8 As shown, the solid-state battery material formulation generation device 800 provided in this application embodiment includes a parameter group generation module 810, a condition acquisition module 820, a parameter group screening module 830, a performance prediction module 840, and a formulation generation module 850.

[0204] The parameter group generation module 810 is used to generate multiple initial parameter groups based on the candidate particle size set and the candidate particle number set. Each initial parameter group includes the average particle size and the average particle number of each type of positive electrode powder particle. For each type of positive electrode powder particle, the particle size follows a normal distribution around the average particle size, and the particle number is within the fluctuation range of the average particle number. The average particle size of each type of positive electrode powder particle is selected from the candidate particle size set, and the average particle number of each type of positive electrode powder particle is selected from the candidate particle number set.

[0205] The condition acquisition module 820 is used to acquire particle packing constraint conditions. The particle packing constraint conditions reflect the constraint relationships between the average particle sizes of various positive electrode powder particles and between the average particle numbers of various positive electrode powder particles.

[0206] The parameter group filtering module 830 is used to input multiple initial parameter groups and particle packing constraints into the constraint layer to obtain multiple constrained parameter groups that meet the particle packing constraints.

[0207] The performance prediction module 840 is used to input each constrained parameter group into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each constrained parameter group.

[0208] The formula generation module 850 is used to determine the target parameter group in each constrained parameter group based on the battery performance parameters corresponding to each constrained parameter group, and generate a solid-state battery cathode formula based on the target parameter group.

[0209] In the solid-state battery material formulation generation apparatus provided in this application embodiment, firstly, multiple initial parameter sets are generated based on a set of candidate particle sizes and a set of candidate particle numbers. These initial parameter sets cover the average particle size and average particle number of various cathode powder particles. This diverse combination increases the diversity of cathode powder particles, overcomes the limitations of traditional single particle size, and helps improve the compaction density of the cathode layer. Secondly, particle packing constraints are obtained, and a constrained parameter set that meets the conditions is obtained using the constraint layer. This ensures the rationality and stability of particle packing and avoids microstructural defects caused by improper particle combination. Then, the constrained parameter set is input into a battery performance parameter prediction model, which can quickly evaluate the battery performance parameters corresponding to different parameter sets. Finally, the target parameter set is determined based on the battery performance parameters, and a solid-state battery cathode formulation is generated. This allows for the selection of the formulation that optimizes battery performance. Through this systematic approach, considering factors such as particle size, number, and packing constraints, the cathode material system is optimized, thereby effectively improving key performance indicators such as ionic conductivity, cycle stability, and energy density of the battery, and enhancing the overall battery performance.

[0210] As an optional embodiment, the battery performance parameter prediction model includes a first sub-model, a second sub-model, a third sub-model, and a battery performance parameter prediction sub-model;

[0211] The performance prediction module 840 includes the following sub-modules:

[0212] The collection acquisition submodule is used to acquire a collection of candidate particulate materials;

[0213] The material selection submodule is used to select the particle material of each of the various cathode powder particles from the candidate particle material set for each constrained parameter group, and to form an expanded parameter group based on the constrained parameter group and the particle material of each of the various cathode powder particles.

[0214] The first prediction submodule is used to input the average particle size and the average number of various positive electrode powder particles in each expanded parameter group into the first sub-model to obtain the first sub-predicted value.

[0215] The second prediction submodule is used to input the average particle size of each of the various positive electrode powder particles in the expanded parameter group and the particle material of each of the various positive electrode powder particles into the second sub-model to obtain the second sub-prediction value;

[0216] The third prediction submodule is used to input the average number of particles of various positive electrode powder particles in the expanded parameter group and the particle material of various positive electrode powder particles into the third sub-model to obtain the third sub-prediction value.

[0217] The performance prediction submodule is used to input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the battery performance parameter prediction sub-model to obtain the battery performance parameters.

[0218] As an optional embodiment, the battery performance parameter prediction sub-model includes a compaction density prediction sub-model, an ionic conductivity prediction sub-model, a cycle stability prediction sub-model, and an overall battery performance parameter prediction sub-model;

[0219] The performance prediction submodule includes the following units:

[0220] The density prediction unit is used to input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the compaction density prediction sub-model to obtain the predicted compaction density.

[0221] The conductivity prediction unit is used to input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the ion conductivity prediction sub-model to obtain the predicted ion conductivity.

[0222] The stability prediction unit is used to input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the cyclic stability prediction sub-model to obtain the predicted cyclic stability.

[0223] The performance prediction unit is used to input compaction density, ionic conductivity, and cycle stability into the overall battery performance parameter prediction sub-model to obtain the predicted battery performance parameters.

[0224] As an optional embodiment, the overall battery performance parameter prediction sub-model includes a positive electrode performance parameter prediction sub-model, an electrolyte compatibility prediction sub-model, and a negative electrode compatibility prediction sub-model;

[0225] Performance prediction unit, used for:

[0226] The compaction density, ionic conductivity, and cycle stability are input into the cathode performance parameter prediction model to obtain the predicted cathode performance parameters.

[0227] The electrolyte compatibility prediction model is obtained by inputting the particle material of various positive electrode powders and the electrolyte material of solid-state batteries into the electrolyte compatibility prediction model.

[0228] By inputting the particle materials of various positive electrode powders and the negative electrode materials of solid-state batteries into the negative electrode fit prediction model, the predicted negative electrode fit is obtained.

[0229] Battery performance parameters are calculated based on positive electrode performance parameters, electrolyte compatibility, and negative electrode compatibility.

[0230] As an optional embodiment, the recipe generation module 850 includes the following sub-modules:

[0231] The material acquisition submodule is used to acquire the particle material of various positive electrode powder particles from the target parameter group.

[0232] The material acquisition submodule is also used to acquire, from the material isomorphism diagram, a group of isomorphic materials with the same microstructure as each acquired particulate material;

[0233] The material replacement submodule is used to replace each particulate material with other materials that are different from the particulate material in the isomorphic material group for each particulate material, so as to obtain multiple equivalent parameter groups of the target parameter group.

[0234] The performance analysis submodule is used to input multiple equivalent parameter groups into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each equivalent parameter group.

[0235] The parameter group selection submodule is used to determine the formula parameter group in each equivalent parameter group based on the battery performance parameters corresponding to each equivalent parameter group, and generate the solid-state battery cathode formula based on the formula parameter group.

[0236] As an optional embodiment, the solid-state battery material formulation generation apparatus 800 further includes the following modules:

[0237] The sample acquisition module is used to acquire multiple solid-state battery cathode samples. Each solid-state battery cathode sample is made up of the average sample particles of various cathode powder particles.

[0238] The sample measurement module is used to measure the mean sample particle size and the mean sample particle number of each type of positive electrode powder particles.

[0239] The sample measurement module is also used to measure the sample battery performance parameters of each solid-state battery cathode sample;

[0240] The sample prediction module is used to input the mean sample particle size and mean sample particle number of each type of positive electrode powder particles into the battery performance parameter prediction model to obtain the predicted battery performance parameters of the solid-state battery positive electrode sample.

[0241] The model training module is used to generate a loss function based on the predicted battery performance parameters and the sample battery performance parameters, and to train a battery performance parameter prediction model based on the loss function.

[0242] As an optional embodiment, the parameter group generation module 810 includes the following sub-modules:

[0243] The particle statistics submodule is used to calculate the first population mean and the first population standard deviation of the candidate particle size set;

[0244] The particle selection submodule is used to select the mean particle size of various positive electrode powder particles from the candidate particle size set, such that the average absolute value of the deviation of the mean particle size of various positive electrode powder particles from the standard deviation is less than a first threshold. The deviation of the mean particle size of the positive electrode powder particles from the standard deviation is equal to: the difference between the mean particle size of the positive electrode powder particles and the first population mean, divided by the first population standard deviation.

[0245] The particle statistics submodule is also used to calculate the second population mean and the second population standard deviation of the candidate particle number set;

[0246] The particle selection submodule is also used to select the mean particle number of each of the various positive electrode powder particles from the candidate particle number set, so that the average absolute value of the deviation of the mean particle number of each of the various positive electrode powder particles from the standard deviation is less than the second threshold. The deviation of the mean particle number of the positive electrode powder particles from the standard deviation is equal to: the difference between the mean particle number of the positive electrode powder particles and the second population mean, divided by the second population standard deviation.

[0247] The parameter set generation submodule is used to generate an initial parameter set based on the average particle size and average number of particles of various positive electrode powders.

[0248] Based on the solid-state battery material formulation generation method provided in this application, correspondingly, this application also provides specific embodiments of solid-state battery material formulation generation equipment.

[0249] Figure 9 A schematic diagram of the hardware structure of the solid-state battery material formulation generation device provided in an embodiment of this application is shown.

[0250] The solid-state battery material formulation generation device may include a processor 901 and a memory 902 storing computer program instructions.

[0251] Specifically, the processor 901 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0252] Memory 902 may include mass storage for data or instructions. For example, and not limitingly, memory 902 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 902 may include removable or non-removable (or fixed) media. Where appropriate, memory 902 may be internal to an integrated gateway disaster recovery device. In a particular embodiment, memory 902 is non-volatile solid-state memory.

[0253] The processor 901 reads and executes computer program instructions stored in the memory 902 to implement any of the solid-state battery material formulation generation methods in the above embodiments.

[0254] In one example, the device for verifying feature matching results may further include a communication interface 903 and a bus 910. For example, Figure 9 As shown, the processor 901, memory 902, and communication interface 903 are connected through bus 910 and complete communication with each other.

[0255] The communication interface 903 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0256] Bus 910 includes hardware, software, or both, that couples components of a solid-state battery material formulation generation device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 410 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0257] Furthermore, in conjunction with the solid-state battery material formulation generation method in the above embodiments, this application embodiment can provide a computer storage medium for implementation. This computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the solid-state battery material formulation generation methods in the above embodiments.

[0258] In addition, in conjunction with the solid-state battery material formulation generation method in the above embodiments, this application embodiment can provide a computer program product to implement it. When the instructions in the computer program product are executed by the processor of an electronic device, the electronic device executes the solid-state battery material formulation generation method provided by any aspect of the above embodiments of this application.

[0259] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0260] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0261] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0262] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, 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, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0263] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for generating a solid-state battery material formulation, characterized in that, include: Based on the candidate particle size set and the candidate particle number set, multiple initial parameter sets are generated. Each initial parameter set includes the average particle size and the average particle number of each type of cathode powder. For each type of cathode powder, the particle size follows a normal distribution around the average particle size, and the particle number is within the fluctuation range of the average particle number. The average particle size of each type of cathode powder is selected from the candidate particle size set, and the average particle number of each type of cathode powder is selected from the candidate particle number set. Obtain particle packing constraints, which reflect the constraint relationships between the average particle sizes of the various positive electrode powder particles and between the average number of the various positive electrode powder particles. The multiple initial parameter sets and the particle packing constraint conditions are input into the constraint layer to obtain multiple constrained parameter sets that meet the particle packing constraint conditions. Each set of constrained parameters is input into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each set of constrained parameters. Based on the battery performance parameters corresponding to each constrained parameter group, a target parameter group is determined in each constrained parameter group, and a solid-state battery cathode formula is generated based on the target parameter group. The battery performance parameter prediction model includes a first sub-model, a second sub-model, a third sub-model, and a battery performance parameter prediction sub-model. The step of inputting each constrained parameter group into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each constrained parameter group includes: Obtain the set of candidate particulate materials; For each constrained parameter group, select the particle material of each of the various cathode powder particles from the candidate particle material set, and form an expanded parameter group based on the constrained parameter group and the particle material of each of the various cathode powder particles. For each expanded parameter group, the average particle size and the average number of each of the various positive electrode powder particles in the expanded parameter group are input into the first sub-model to obtain the first sub-predicted value. The average particle size and particle material of each of the various positive electrode powder particles in the expanded parameter group are input into the second sub-model to obtain the second sub-predicted value. The average number of each of the various positive electrode powder particles in the expanded parameter group and the particle material of each of the various positive electrode powder particles are input into the third sub-model to obtain the third sub-predicted value; The first sub-predicted value, the second sub-predicted value, and the third sub-predicted value are input into the battery performance parameter prediction sub-model to obtain the battery performance parameters; the battery performance parameter prediction sub-model includes a compaction density prediction sub-model, an ionic conductivity prediction sub-model, a cycle stability prediction sub-model, and an overall battery performance parameter prediction sub-model.

2. The method for generating solid-state battery material formulations according to claim 1, characterized in that, The step of inputting the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the battery performance parameter prediction sub-model to obtain the battery performance parameters includes: The first sub-predicted value, the second sub-predicted value, and the third sub-predicted value are input into the compaction density prediction sub-model to obtain the predicted compaction density. The first sub-predicted value, the second sub-predicted value, and the third sub-predicted value are input into the ionic conductivity prediction sub-model to obtain the predicted ionic conductivity. The first sub-predicted value, the second sub-predicted value, and the third sub-predicted value are input into the cyclic stability prediction sub-model to obtain the predicted cyclic stability. The compaction density, ionic conductivity, and cycle stability are input into the overall battery performance parameter prediction sub-model to obtain the predicted battery performance parameters.

3. The method for generating solid-state battery material formulations according to claim 1, characterized in that, The initial parameter set is generated in the following manner: Calculate the first population mean and the first population standard deviation of the candidate particle size set; From the candidate particle size set, the mean particle size of each of the various positive electrode powder particles is selected such that the average absolute value of the deviation of the mean particle size of each of the various positive electrode powder particles from the standard deviation is less than a first threshold. The deviation of the mean particle size of the positive electrode powder particles from the standard deviation is equal to: the difference between the mean particle size of the positive electrode powder particles and the first overall mean, divided by the first overall standard deviation. Calculate the second population mean and the second population standard deviation of the candidate particle number set; From the candidate particle number set, the mean particle number of each of the various positive electrode powder particles is selected such that the average absolute value of the deviation of the mean particle number of each of the various positive electrode powder particles from the standard deviation is less than a second threshold. The deviation of the mean particle number of the positive electrode powder particles from the standard deviation is equal to: the difference between the mean particle number of the positive electrode powder particles and the second population mean, divided by the second population standard deviation. The initial parameter set is generated based on the average particle size and average number of each of the various positive electrode powder particles.

4. A solid-state battery, characterized in that, It includes a positive electrode, a negative electrode, and an electrolyte, wherein the electrolyte is located between the positive electrode and the negative electrode, and the positive electrode is generated based on the solid-state battery material formulation generation method according to any one of claims 1-3.

5. A solid-state battery material formulation generation apparatus, characterized in that, include: The parameter group generation module is used to generate multiple initial parameter groups based on the candidate particle size set and the candidate particle number set. Each initial parameter group includes the average particle size and the average particle number of each type of positive electrode powder particle. For each type of positive electrode powder particle, the particle size follows a normal distribution around the average particle size, and the particle number is within the fluctuation range of the average particle number. The average particle size of each type of positive electrode powder particle is selected from the candidate particle size set, and the average particle number of each type of positive electrode powder particle is selected from the candidate particle number set. The condition acquisition module is used to acquire particle packing constraint conditions, which reflect the constraint relationship between the average particle size of each of the various positive electrode powder particles and the average number of each of the various positive electrode powder particles. The parameter group filtering module is used to input the multiple initial parameter groups and the particle packing constraint conditions into the constraint layer to obtain multiple constrained parameter groups that meet the particle packing constraint conditions. The performance prediction module is used to input each constrained parameter group into the battery performance parameter prediction model to obtain the battery performance parameters corresponding to each constrained parameter group. The formulation generation module is used to determine the target parameter group in each constrained parameter group based on the battery performance parameters corresponding to each constrained parameter group, and generate a solid-state battery cathode formulation based on the target parameter group. The battery performance parameter prediction model includes a first sub-model, a second sub-model, a third sub-model, and a battery performance parameter prediction sub-model. The performance prediction module includes: The collection acquisition submodule is used to acquire a collection of candidate particulate materials; The material selection submodule is used to select the particle material of each of the various cathode powder particles from the candidate particle material set for each constrained parameter group, and form an expanded parameter group based on the constrained parameter group and the particle material of each of the various cathode powder particles. The first prediction submodule is used to input the average particle size and the average number of each of the various positive electrode powder particles in the expanded parameter group into the first sub-model to obtain the first sub-predicted value for each expanded parameter group. The second prediction submodule is used to input the average particle size of each of the various positive electrode powder particles in the expanded parameter group and the particle material of each of the various positive electrode powder particles into the second sub-model to obtain the second sub-prediction value; The third prediction submodule is used to input the average number of each of the various positive electrode powder particles in the expanded parameter group and the particle material of each of the various positive electrode powder particles into the third sub-model to obtain the third sub-prediction value; The performance prediction submodule is used to input the first sub-predicted value, the second sub-predicted value, and the third sub-predicted value into the battery performance parameter prediction sub-model to obtain the battery performance parameters; the battery performance parameter prediction sub-model includes a compaction density prediction sub-model, an ionic conductivity prediction sub-model, a cycle stability prediction sub-model, and an overall battery performance parameter prediction sub-model.

6. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the solid-state battery material formulation generation method according to any one of claims 1 to 3.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the solid-state battery material formulation generation method according to any one of claims 1 to 3.

8. A computer program product comprising a computer program that is read and executed by a processor of a computer device, causing the computer device to perform the solid-state battery material formulation generation method according to any one of claims 1 to 3.

Images