A dry pressing method for a self-supporting dry electrode without current collector

Abstract

This application provides a dry pressing method for a self-supporting dry electrode without a current collector, comprising: S1, obtaining an initial attribute dataset and flow simulation basic parameters by analyzing the differences in particle size, density, and flowability of powder in a multi-layer composite structure, simulating and calculating the thickness distribution of each powder in the mold cavity, and determining a uniform distribution prediction model; S2, if the uniform distribution prediction model shows that the thickness deviation exceeds a preset thickness deviation threshold, adjusting the filling sequence parameters to obtain an optimized sequence scheme; S3, simulating the asymmetric compression distribution under the pressing direction, identifying the interface separation risk area, optimizing the powder arrangement to enhance the bonding strength, and obtaining an arrangement configuration that enhances the bonding strength; S4, extracting mechanical strength indicators from the arrangement configuration that enhances the bonding strength, determining the final dry pressing parameter set through iterative verification, generating a pressing path plan for the multi-layer composite electrode, and obtaining a position offset control scheme.

Description

Technical Field

[0001] This invention relates to the field of dry electrode technology, and in particular to a dry pressing method for a self-supporting dry electrode without current collector. Background Technology

[0002] Dry pressing technology for self-supporting dry electrodes without current collectors is a key direction for promoting the development of high-energy-density batteries. This type of electrode does not require traditional metal current collectors and is directly formed by dry pressing of powder, which can significantly reduce weight and improve the overall performance of the battery, and has an irreplaceable and important position in the field of new energy.

[0003] Current dry electrode forming methods mostly rely on single filling or simple stacking of powder. Although this method is convenient to operate, it is easy to cause positional displacement and interface separation of each functional layer during the pressing process. This is because the flow and interaction of powder in the mold cavity cannot be effectively controlled, and the bonding strength between layers is difficult to maintain stably, especially in multi-layer composite structures.

[0004] In multilayer composite current collectorless electrodes, different powders such as the active material layer, conductive layer, and binder layer need to be filled into the mold cavity in a specific spatial order. If the filling order is improper, the differences in particle size, density, and flowability of the different powders will directly affect the spreading thickness and uniformity of each layer in the mold cavity, thereby amplifying the weak bonding areas at the interlayer interfaces. For example, when the denser active material powder is filled first, the subsequently filled conductive or binder powder cannot penetrate evenly into the gaps between the lower particles, resulting in voids or stress concentrations at the interface after pressing. This weak interfacial bonding will further deteriorate with subsequent changes in the pressing direction, ultimately reducing the overall mechanical strength of the electrode and affecting cycle stability.

[0005] The filling sequence must also match the subsequent pressing direction; otherwise, unidirectional filling from bottom to top or top to bottom can easily cause asymmetrical compression of the powder at the interface, resulting in significant differences in interlayer bonding strength. Therefore, how to rationally determine the filling sequence of multiple layers of powder and design a matching arrangement, while effectively controlling the spreading thickness and uniformity of each layer to achieve strong interlayer bonding, has become a key problem that urgently needs to be solved in the dry pressing process. Summary of the Invention

[0006] To address the technical problems mentioned in the background section, the present invention provides a dry pressing method for a self-supporting dry electrode without a current collector, the method comprising:

[0007] S1. By analyzing the differences in particle size, density, and flowability of powder in the multi-layer composite structure, an initial attribute dataset and basic parameters for flow simulation are obtained. The thickness distribution of each powder in the mold cavity is simulated and calculated to determine the uniform distribution prediction model. S2. If the uniform distribution prediction model shows that the thickness deviation exceeds the preset thickness deviation threshold, the filling sequence parameters are adjusted to obtain an optimized sequence scheme. S3. The asymmetric compression distribution under the pressing direction is simulated to identify the interface separation risk area. The powder arrangement is optimized to enhance the bonding strength, resulting in an arrangement configuration that enhances the bonding strength. S4. Mechanical strength indicators are extracted from the arrangement configuration that enhances the bonding strength. The final dry molding parameter set is determined through iterative verification. The pressing path planning of the multi-layer composite electrode is generated to obtain the position offset control scheme.

[0008] Optionally, step S1 includes:

[0009] Step S11: By collecting measurement data on particle size, particle density, and flowability differences, an initial attribute dataset is constructed. The discrete element method is used to perform flow simulation, calculate the motion trajectory of each powder in the mold cavity, and determine the applicable range of the basic parameters.

[0010] Step S12: Based on the results of the flow simulation, extract the distribution data of the spreading thickness, analyze the regularity of the thickness distribution, and obtain the morphological characteristics of the powder accumulation in the mold cavity.

[0011] Step S13: If the thickness distribution data is within the preset thickness threshold range, it is determined to be a preliminary uniform distribution, and a distribution state that meets the conditions is obtained.

[0012] Step S14: If the thickness threshold range is exceeded, the basic parameters are iteratively adjusted and the flow simulation is repeated.

[0013] Step S15: By extracting features from the thickness distribution data that meets the conditions, a uniform distribution prediction model is constructed. New particle size and particle density data are input to calculate the predicted spreading thickness and determine the uniformity of powder distribution in the mold cavity.

[0014] Step S16: Based on the predicted spreading thickness, generate the corresponding optimization strategy for powder distribution in the mold cavity to obtain the final uniform distribution scheme.

[0015] Optionally, step S12 further includes:

[0016] When extracting the spread thickness distribution data, the bottom of the mold cavity is divided into multiple small areas, and the powder thickness of each area is measured.

[0017] Optionally, step S2 includes:

[0018] Step S21: Obtain the 3D model and process parameters of the printed part, and generate an initial filling path sequence scheme;

[0019] Step S22: The initial filling path sequence scheme is forward reasoned through the uniform distribution prediction model to obtain the corresponding spread thickness prediction value.

[0020] Step S23: Calculate the paving thickness deviation of each area from the predicted paving thickness value, and determine whether the paving thickness deviation is lower than the preset thickness deviation threshold.

[0021] Step S24: If all are below the preset thickness deviation threshold, the current filling path sequence scheme is directly output as the final scheme.

[0022] Step S25: If the spreading thickness deviation exceeds the preset thickness deviation threshold, the initial filling path sequence parameters are optimized and iterated to obtain an optimized filling sequence scheme.

[0023] Step S26: The optimized filling sequence scheme is forward-reasoned again using the uniform distribution prediction model to obtain a new predicted value of the spreading thickness and a new spreading thickness deviation.

[0024] Optionally, step S21 includes:

[0025] The initial filling path sequence is generated using a regional parallel scanning strategy, and the initial filling path sequence adopts a raster-style filling from left to right and from bottom to top.

[0026] Optionally, step S25 includes:

[0027] A genetic algorithm was used to optimize the initial filling path sequence parameters. The scanning direction angle, sub-region sequence number, and jumping strategy were used as gene encodings. The initial population size was 50, the crossover rate was 0.8, and the mutation rate was 0.05. After 20 generations of iteration, the optimized filling sequence scheme was obtained.

[0028] Optionally, step S3 includes:

[0029] Step S31: Apply load to the powder crushing model under the given pressing direction and optimized sequence scheme, and obtain stress distribution data on the entire interface;

[0030] Step S32: Based on the comparison between the interface stress distribution data and the allowable tensile stress threshold of the material, if the tensile stress in a certain area exceeds the allowable tensile stress threshold, then mark the area as an interface separation risk area and obtain a set of risk areas.

[0031] Step S33: Discretize the risk area set into multiple sub-regions, obtain the coordinate range and risk level of each sub-region, and construct a fitness function based on the risk area set and the risk level of the sub-regions;

[0032] Step S34: Using the fitness function as the evaluation criterion, perform population initialization, selection, crossover and mutation operations to regenerate the powder filling model, update the material property distribution in the finite element model, and obtain the crushing model with improved bonding strength.

[0033] Step S35: Apply the same compressive load in the same direction to the crushed model with improved bonding strength through finite element analysis, obtain new interface stress distribution data, and determine whether the remaining separation risk area is an empty set.

[0034] Optionally, step S33 includes:

[0035] The risk area is discretized into multiple sub-regions using a grid partitioning method. The coordinate range of each sub-region is determined by the grid number. The risk level is divided into three levels: high, medium, and low, based on the degree of tensile stress exceeding the standard.

[0036] Optionally, step S33 further includes:

[0037] The fitness function aims to minimize the tensile stress in all sub-regions, and a genetic algorithm is used with the powder spatial arrangement parameters as chromosome genes.

[0038] Optionally, step S4 includes:

[0039] Step S41: Extract mechanical strength data from the arrangement configuration that enhances bonding strength, separate key indicators, and obtain a preliminary strength dataset.

[0040] Step S42: Classify the preliminary strength dataset to determine the core parameter set related to the binding strength;

[0041] Step S43: For the core parameter group, implement an iterative verification process. If the verification result does not match the preset parameter threshold, adjust the parameter weights to obtain the optimized parameter combination.

[0042] Step S44: Extract key elements of dry molding from the optimized parameter combination to generate a molding parameter set suitable for multi-layer composite structures;

[0043] Step S45: Based on the molding parameter set, construct a path planning model for electrode pressing and obtain path distribution data during the pressing process.

[0044] Step S46: Analyze the potential risks of position offset using path distribution data. If the offset exceeds the preset range, generate a corresponding control scheme.

[0045] Step S47: Based on the generated control scheme, output the final position offset adjustment strategy and determine the pressing execution scheme of the multilayer composite electrode.

[0046] The technical solution provided by this invention has the following beneficial effects:

[0047] This invention discloses a dry pressing method for self-supporting dry electrodes without current collectors, aiming to solve the comprehensive problems of uneven spreading thickness, interface separation risk, and insufficient mechanical strength caused by differences in powder particle properties. This invention constructs an initial property dataset by analyzing differences in powder particle size, density, and flowability. Particle simulation is used to predict the spreading thickness distribution and establish a uniform distribution model. If the deviation exceeds the standard, the filling sequence is optimized. Subsequently, finite element analysis is used to evaluate the compression distribution under the pressing direction, identify interface separation risk areas, and a genetic algorithm is used to optimize the powder arrangement to enhance bonding strength. Finally, mechanical strength indicators are extracted and iteratively verified to generate pressing path planning and position offset control schemes. The core innovation of this invention lies in achieving full-process control from powder distribution to pressing path through multi-algorithm collaborative optimization, effectively improving the molding accuracy and structural stability of multilayer composite electrodes, and providing technical support for the efficient preparation of high-performance composite materials. Attached Figure Description

[0048] Figure 1 This is a flowchart of a dry pressing method for a self-supporting dry electrode without current collector according to the present invention.

[0049] Figure 2 This is a schematic diagram of a dry pressing method for a self-supporting dry electrode without current collector according to the present invention.

[0050] Figure 3 This is another schematic diagram of a dry pressing method for a self-supporting dry electrode without current collector according to the present invention. Detailed Implementation

[0051] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] like Figures 1-3 As shown, the present invention provides a dry pressing method for a self-supporting dry electrode without current collectors, specifically including:

[0053] S1. By analyzing the differences in particle size, density, and flowability of powder in a multi-layer composite structure, an initial attribute dataset and basic parameters for flow simulation are obtained. The thickness distribution of each powder in the mold cavity is simulated and calculated to determine a uniform distribution prediction model.

[0054] Specifically, by analyzing the differences in particle size, density, and flowability of each powder in the multi-layer composite structure, an initial attribute dataset is obtained to acquire the basic parameters for powder flow simulation. Based on the initial attribute dataset, the particle simulation method is used to calculate the thickness distribution of each powder in the mold cavity and determine the uniform distribution prediction model.

[0055] Optionally, this step also includes:

[0056] Step S11 involves collecting measurement data on particle size, particle density, and flowability differences, using the discrete element method to simulate flow, calculating the trajectory of each powder within the mold cavity, and determining the applicable range of the basic parameters.

[0057] Step S12: Based on the results of the flow simulation, extract the distribution data of the spreading thickness, analyze the regularity of the thickness distribution, and obtain the morphological characteristics of the powder accumulation in the mold cavity.

[0058] Step S13: If the thickness distribution data is within the preset thickness threshold range, it is determined to be a preliminary uniform distribution, and a distribution state that meets the conditions is obtained.

[0059] In step S14, if the thickness threshold is exceeded, the basic parameters are iteratively adjusted and the flow simulation is repeated.

[0060] Step S15: By extracting features from the thickness distribution data that meets the conditions, a uniform distribution prediction model is constructed. New particle size and particle density data are input to calculate the predicted spreading thickness and determine the uniformity of powder distribution in the mold cavity.

[0061] Step S16: Based on the predicted spreading thickness, generate the corresponding optimization strategy for powder distribution in the mold cavity to obtain the final uniform distribution scheme.

[0062] For example, when constructing the initial attribute dataset, the characteristics of powders with different particle sizes and densities can be measured experimentally. Assuming the particle size ranges from 50 to 200 micrometers and the particle density from 2.5 to 3.5 grams per cubic centimeter, flowability parameters such as the angle of repose can be recorded to preliminarily describe the powder characteristics. The purpose of this step is to provide a reliable data foundation for subsequent simulations, ensuring that the simulation results closely approximate reality.

[0063] In one possible implementation, when using the discrete element method (DEM) for flow simulation, the powder particles can be treated as discrete units, and their motion trajectories within the mold cavity can be simulated. Assuming the mold cavity is rectangular with dimensions of 10 cm x 5 cm, and the powder falls freely from the top center, the motion trajectory data is obtained by calculating the collisions and friction between particles. This helps determine the applicability of basic parameters such as the interparticle friction coefficient within the range of 0.3 to 0.5, providing a basis for subsequent analysis.

[0064] In one embodiment, when extracting the spread thickness distribution data, the bottom of the mold cavity can be divided into multiple small regions, and the powder thickness in each region can be measured. Assuming the thickness data ranges from 1.0 to 3.0 mm, the distribution pattern can be analyzed, revealing that the thickness is higher in the central region and lower at the edges. This regularity reflects the morphological characteristics of powder accumulation, providing data support for uniformity determination.

[0065] In one possible implementation, if the preset thickness threshold is 1.5 to 2.5 mm, and the actual data exceeds this range, the basic parameters need to be iteratively adjusted, such as reducing the particle friction coefficient to 0.2, and the flow is re-simulated until the thickness distribution meets the requirements. This can effectively improve the uniformity of distribution and reduce uneven accumulation within the mold cavity.

[0066] For example, when constructing a uniform distribution prediction model, features such as average thickness and standard deviation can be extracted based on suitable thickness distribution data. Assuming an average thickness of 2.0 mm and a standard deviation of 0.2 mm, machine learning methods can be used to determine the combination of model parameters. This provides a precise tool for subsequent predictions and significantly improves prediction efficiency.

[0067] In one possible implementation, a predictive model is used to predict the particle size (e.g., 100 micrometers) and density (e.g., 3.0 g / cm³) of newly input powder, calculating a spreading thickness of approximately 2.1 mm, and assessing its distribution uniformity. If the results show poor uniformity, optimization strategies can be generated, such as adjusting the mold cavity tilt angle to 5 degrees to improve powder flow, ultimately achieving a uniform distribution. This not only improves production efficiency but also reduces material waste.

[0068] Specifically, the generation of optimization strategies can support the uniformity target from multiple perspectives, such as combining mold cavity design adjustments and powder parameter optimization to jointly ensure the distribution effect. Through these methods, the powder's accumulation state within the mold cavity can be significantly improved, providing reliable technical support for industrial production.

[0069] S2. If the uniform distribution prediction model shows that the spreading thickness deviation exceeds the preset thickness deviation threshold, then adjust the filling sequence parameters to obtain the optimized sequence scheme.

[0070] Optionally, this step also includes:

[0071] Step S21: Obtain the 3D model and process parameters of the printed part, and generate an initial filling path sequence scheme.

[0072] Step S22: The initial filling path sequence scheme is forward reasoned using a uniform distribution prediction model to obtain the corresponding spread thickness prediction value.

[0073] Step S23: Calculate the paving thickness deviation of each area from the predicted paving thickness value, and determine whether the paving thickness deviation is lower than the preset thickness deviation threshold.

[0074] Step S24: If all are below the preset thickness deviation threshold, the current filling path sequence scheme is directly output as the final scheme.

[0075] Step S25: If the spreading thickness deviation exceeds the preset thickness deviation threshold, the initial filling path sequence parameters are optimized and iterated to obtain an optimized filling sequence scheme.

[0076] Step S26: The optimized filling sequence scheme is forward-reasoned again using the uniform distribution prediction model to obtain a new predicted value of the spreading thickness and a new spreading thickness deviation.

[0077] For example, when acquiring the 3D model and process parameters of a printed part, the STL model of the part can be exported from CAD software first, while recording key process parameters such as laser power of 250W, scanning spacing of 0.1mm, layer thickness of 40 micrometers, and powder material of 316L stainless steel. These parameters, together with the model, constitute the initial input, providing a complete data foundation for subsequent path planning.

[0078] In one possible implementation, the initial fill path sequence can be generated using a regional parallel scanning strategy. For example, each slice can be divided into 5×5mm square sub-regions, and the initial fill path sequence can use a raster-style fill from left to right and from bottom to top, with the jump distance controlled within 3mm to reduce heat accumulation. This initial approach can quickly form an executable path while reserving adjustment space for subsequent optimization.

[0079] Specifically, when performing forward inference on the initial filling path sequence scheme using a trained uniform distribution prediction model, the model quickly outputs the predicted spreading thickness value for the entire layer based on the path direction, spacing, and jump order. For example, the prediction results show that the thickness in the central area is about 2.1 mm, the edge area is only 1.6 mm, and the four corner areas are as low as 1.4 mm, thus obtaining a complete thickness distribution cloud map.

[0080] For example, when calculating the paving thickness deviation of each area from the predicted paving thickness value, the target thickness can be set to 2.0mm. Then, the deviation in the center area is +0.1mm, the deviation in the edge area is -0.4mm, and the deviation in the four corner areas is -0.6mm. Through gridded statistics, it was found that the absolute value of the deviation in about 15% of the areas exceeds the preset thickness deviation threshold of 0.3mm, so the current solution is determined to be unqualified.

[0081] In one possible implementation, if all the thickness deviations are below the preset threshold, the current filling path sequence is directly output as the final solution without further iteration, thereby significantly shortening the planning time and improving production efficiency.

[0082] Specifically, when regions with deviations exceeding the thickness deviation threshold exist, these out-of-tolerance regions are automatically marked and formed into a set. For example, 12 sub-regions, including the four corners and parts of the edges, are marked as out-of-tolerance regions. This set will serve as the input of the penalty term for the genetic algorithm, directly guiding the direction of subsequent optimization.

[0083] In one embodiment, a genetic algorithm is used to optimize and iterate the initial filling path sequence parameters. The scanning direction angle, sub-region sequence number, and skipping strategy can be used as gene encoding. The initial population size is 50, the crossover rate is 0.8, and the mutation rate is 0.05. After 20 generations of iteration, an optimized filling sequence scheme is obtained. For example, it can be changed to 45-degree alternating scanning combined with island-shaped skipping sequence to make the heat input more uniform.

[0084] In one possible implementation, the optimized filling sequence scheme is again forward-reasoned through a uniform distribution prediction model to obtain new thickness prediction values. For example, the overall thickness distribution is between 1.85 and 2.15 mm, with a maximum deviation of only 0.15 mm. The deviation of all areas is lower than the thickness deviation threshold of 0.3 mm, which achieves highly uniform powder spreading, providing an excellent foundation for subsequent melting and densification, and significantly reducing the risk of warping and hole defects in printed parts.

[0085] S3 simulates the asymmetric compression distribution under the pressing direction, identifies the risk area of ​​interface separation, optimizes the powder arrangement to enhance the bonding strength, and obtains the arrangement configuration that enhances the bonding strength.

[0086] Specifically, for the optimized sequence scheme, the asymmetric compression distribution under the pressing direction is simulated by finite element analysis to identify the interface separation risk area. The data of the interface separation risk area are obtained, and the powder arrangement is optimized by genetic algorithm to obtain an arrangement configuration that enhances the bonding strength.

[0087] Optionally, this step also includes:

[0088] Step S31: Apply load to the powder crushing model under a given pressing direction and optimized sequence scheme through finite element analysis to obtain stress distribution data on the entire interface.

[0089] Step S32: Based on the comparison between the interface stress distribution data and the allowable tensile stress threshold of the material, if the tensile stress in a certain area exceeds the allowable tensile stress threshold, then mark the area as an interface separation risk area and obtain a set of risk areas.

[0090] Step S33: Discretize the risk area set into multiple sub-regions, obtain the coordinate range and risk level of each sub-region, and construct a fitness function based on the risk area set and the risk level of the sub-regions.

[0091] Preferably, the fitness function aims to minimize the tensile stress in all sub-regions and employs a genetic algorithm with the powder spatial arrangement parameters as chromosome genes.

[0092] Step S34: Using the fitness function as the evaluation criterion, perform population initialization, selection, crossover, and mutation operations to regenerate the powder filling model, update the material property distribution in the finite element model, and obtain the crushing model with improved bonding strength.

[0093] Step S35: Apply the same compressive load in the same direction to the crushed model with improved bonding strength through finite element analysis, obtain new interface stress distribution data, and determine whether the remaining separation risk area is an empty set.

[0094] For example, when applying loads to a powder crushing model using finite element analysis, a three-dimensional model can be constructed first to simulate the stress on the powder under a specific pressing direction. Assuming the pressing direction is vertically downward, a uniformly distributed pressure is applied, with an initial value of 10 MPa, and the stress distribution across the entire interface is analyzed. In principle, finite element analysis divides the model into multiple small elements, calculates the stress state of each element individually, and finally summarizes the results to form an overall stress distribution diagram.

[0095] For example, if the maximum tensile stress in a certain area is found to reach 8 MPa, while the allowable tensile stress threshold of the material is 6 MPa, then the area is marked as a risk area for interface separation.

[0096] For example, after marking risk areas and obtaining a set of risk areas, a grid partitioning method can be used to discretize the risk areas into multiple sub-regions. Assuming the entire interface is divided into 100 grid cells, 10 of which are marked as risk areas, the coordinate range of each sub-region can be determined by the grid number, and the risk level is divided into high, medium, and low levels based on the degree of tensile stress exceeding the limit. This partitioning helps to accurately locate problem areas and provides data support for subsequent optimization.

[0097] For example, when constructing the fitness function, the goal can be to minimize the tensile stress in all sub-regions, assigning higher weights to regions with higher risk levels. Assuming a high-risk region has a weight of 3, medium-risk a weight of 2, and low-risk a weight of 1, the fitness function calculates the weighted stress values ​​for all sub-regions. This approach prioritizes high-risk regions, ensuring a more targeted optimization strategy.

[0098] For example, when using a genetic algorithm to optimize the spatial arrangement parameters of powder materials, particle size, distribution density, and stacking pattern can be encoded as genes. The initial population size is set to 60, the crossover rate to 0.7, and the mutation rate to 0.1. The optimal parameter combination is selected through multiple generations of evolution. In principle, the genetic algorithm simulates the biological evolution process, iteratively updating parameters until the fitness function value reaches its optimum. This method can effectively explore complex parameter spaces.

[0099] For example, after regenerating the filling model based on the optimal powder arrangement parameters, the material property distribution in the finite element model can be updated. Suppose the optimized powder particle density increases from 0.6 g / cm³ to 0.75 g / cm³, reanalysis reveals a more uniform distribution of interfacial stress. This update provides a more stable foundation for the subsequent pressing process.

[0100] For example, when the optimized crushing model is subjected to the same load again for finite element analysis, if the new interface stress distribution shows that the tensile stress in all regions is below the allowable tensile stress threshold of 6 MPa, then the remaining separation risk areas are empty sets. This indicates that the optimized model has significantly improved the bonding strength, providing a reliable guarantee for subsequent processes. Such repeated verification ensures the feasibility of the final solution.

[0101] For example, from a business perspective, controlling the risk of interface separation during powder pressing is crucial. Excessive tensile stress can cause the powder layer to separate from the matrix, affecting the overall structural stability. Through the series of analyses and optimizations described above, stress concentration in risk areas can be effectively reduced, laying a solid foundation for subsequent processing. This method has wide applicability in the field of powder molding.

[0102] S4. Extract mechanical strength indicators from the arrangement configuration that enhances bonding strength, determine the final dry forming parameter set through iterative verification, generate the pressing path planning of multilayer composite electrodes, and obtain the position offset control scheme.

[0103] Optionally, this step also includes:

[0104] Step S41: Extract mechanical strength data from the arrangement configuration that enhances bonding strength, and use data filtering methods to separate key indicators to obtain a preliminary strength dataset.

[0105] Step S42: Classify the preliminary strength dataset to determine the core parameter set related to the binding strength.

[0106] Step S43: For the core parameter group, implement an iterative verification process. If the verification result does not match the preset parameter threshold, adjust the parameter weights to obtain the optimized parameter combination.

[0107] Step S44: Extract key elements of dry molding from the optimized parameter combination to generate a molding parameter set suitable for multi-layer composite structures.

[0108] Step S45: Based on the molding parameter set, construct a path planning model for electrode pressing and obtain path distribution data during the pressing process.

[0109] Step S46: Analyze the potential risks of position offset using path distribution data. If the offset exceeds the preset range, generate a corresponding control scheme.

[0110] Step S47: Based on the generated control scheme, output the final position offset adjustment strategy and determine the pressing execution scheme of the multilayer composite electrode.

[0111] For example, in the field of dry pressing of multilayer composite electrodes, when extracting mechanical strength data from the arrangement and configuration that enhances bonding strength, the main indicators collected are interfacial bonding strength, radial compressive strength and interlaminar shear strength.

[0112] Specifically, a set of data points with a maximum tensile stress of 2.8 MPa and an average compressive strength of 45 MPa can be directly output through finite element post-processing to form a preliminary strength dataset.

[0113] In one possible implementation, when using a support vector machine to classify the dataset, samples with a binding strength greater than or equal to 4 MPa are first labeled as high-strength samples, and those less than 2 MPa are labeled as low-strength samples. The model is then trained using a radial basis function kernel. Ultimately, the model identifies that the proportion of active material, the density of binder distribution, and the curvature of the compression path are the three parameters with the highest correlation to the binding strength, achieving an accuracy of over 92%.

[0114] For example, when implementing an iterative verification process for the core parameter set, it was found that the contribution of the proportion of active material to the binding strength under the initial weight was overestimated, resulting in a prediction bias of 15%.

[0115] It should be noted that by introducing Lagrange multipliers to adjust the weights, the weight of the active material ratio decreased from 0.45 to 0.32, the weight of the binder distribution density increased to 0.38, and the validation set error rapidly decreased to within 4%, thus obtaining an optimized parameter combination that better meets the preset parameter thresholds.

[0116] In one embodiment, key elements of dry molding are extracted from the optimized parameter combination, including controlling the mass ratio of active material to binder at 94:6, single roll pressing speed at 0.8 m / min, and pressing temperature at 65°C. These elements directly constitute a molding parameter set suitable for multi-layer composite structures, which can improve the electrode thickness uniformity to over 98%.

[0117] Specifically, when constructing the electrode pressing path planning model based on the forming parameter set, cubic spline interpolation is used to generate continuous rolling tracks. The path distribution data shows that the offset in the central area is only 0.12 mm, and the maximum offset in the edge area is 0.47 mm.

[0118] Preferably, when the edge offset is detected to exceed the preset range of 0.4mm, the system automatically generates a control scheme, which controls the offset to within 0.25mm by adding an auxiliary pressure roller at the edge and reducing the local line pressure by 15%.

[0119] For example, after applying this position offset adjustment strategy in actual production, the interlayer bonding strength of the multilayer composite electrode increased from the initial 3.2MPa to 5.1MPa, the interface separation rate decreased to below 0.3%, and the yield increased by 12%, effectively ensuring the structural integrity and cycle stability of the high-capacity battery electrode.

[0120] The above description is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several improvements and additions without departing from the principle of the present invention, and these improvements and additions should also be considered within the scope of protection of the present invention.

Claims

1. A dry pressing method for a self-supporting dry electrode without current collector, characterized in that, The method includes: S1. By analyzing the differences in particle size, density, and flowability of powder in the multi-layer composite structure, an initial attribute dataset and basic parameters for flow simulation are obtained. The thickness distribution of each powder in the mold cavity is simulated and calculated to determine the uniform distribution prediction model. S2. If the uniform distribution prediction model shows that the thickness deviation exceeds the preset thickness deviation threshold, the filling sequence parameters are adjusted to obtain an optimized sequence scheme. S3. The asymmetric compression distribution under the pressing direction is simulated to identify the interface separation risk area. The powder arrangement is optimized to enhance the bonding strength, resulting in an arrangement configuration that enhances the bonding strength. S4. Mechanical strength indicators are extracted from the arrangement configuration that enhances the bonding strength. The final dry molding parameter set is determined through iterative verification. The pressing path planning of the multi-layer composite electrode is generated to obtain the position offset control scheme.

2. The method according to claim 1, characterized in that, Step S1 includes: Step S11: By collecting measurement data on particle size, particle density, and flowability differences, an initial attribute dataset is constructed. The discrete element method is used to perform flow simulation, calculate the motion trajectory of each powder in the mold cavity, and determine the applicable range of the basic parameters. Step S12: Based on the results of the flow simulation, extract the distribution data of the spreading thickness, analyze the regularity of the thickness distribution, and obtain the morphological characteristics of the powder accumulation in the mold cavity. Step S13: If the thickness distribution data is within the preset thickness threshold range, it is determined to be a preliminary uniform distribution, and a distribution state that meets the conditions is obtained. Step S14: If the thickness threshold range is exceeded, the basic parameters are iteratively adjusted and the flow simulation is repeated. Step S15: By extracting features from the thickness distribution data that meets the conditions, a uniform distribution prediction model is constructed. New particle size and particle density data are input to calculate the predicted spreading thickness and determine the uniformity of powder distribution in the mold cavity. Step S16: Based on the predicted spreading thickness, generate the corresponding optimization strategy for powder distribution in the mold cavity to obtain the final uniform distribution scheme.

3. The method according to claim 2, characterized in that, Step S12 further includes: When extracting the spread thickness distribution data, the bottom of the mold cavity is divided into multiple small areas, and the powder thickness of each area is measured.

4. The method according to claim 1, characterized in that, Step S2 includes: Step S21: Obtain the 3D model and process parameters of the printed part, and generate an initial filling path sequence scheme; Step S22: The initial filling path sequence scheme is forward reasoned through the uniform distribution prediction model to obtain the corresponding spread thickness prediction value. Step S23: Calculate the paving thickness deviation of each area from the predicted paving thickness value, and determine whether the paving thickness deviation is lower than the preset thickness deviation threshold. Step S24: If all are below the preset thickness deviation threshold, the current filling path sequence scheme is directly output as the final scheme. Step S25: If the spreading thickness deviation exceeds the preset thickness deviation threshold, the initial filling path sequence parameters are optimized and iterated to obtain an optimized filling sequence scheme. Step S26: The optimized filling sequence scheme is forward-reasoned again using the uniform distribution prediction model to obtain a new predicted value of the spreading thickness and a new spreading thickness deviation.

5. The method according to claim 4, characterized in that, Step S21 includes: The initial filling path sequence is generated using a regional parallel scanning strategy, and the initial filling path sequence adopts a raster-style filling from left to right and from bottom to top.

6. The method according to claim 5, characterized in that, Step S25 includes: A genetic algorithm was used to optimize the initial filling path sequence parameters. The scanning direction angle, sub-region sequence number, and jumping strategy were used as gene encodings. The initial population size was 50, the crossover rate was 0.8, and the mutation rate was 0.

05. After 20 generations of iteration, the optimized filling sequence scheme was obtained.

7. The method according to claim 1, characterized in that, Step S3 includes: Step S31: Apply load to the powder crushing model under the given pressing direction and optimized sequence scheme, and obtain stress distribution data on the entire interface; Step S32: Based on the comparison between the interface stress distribution data and the allowable tensile stress threshold of the material, if the tensile stress in a certain area exceeds the allowable tensile stress threshold, then mark the area as an interface separation risk area and obtain a set of risk areas. Step S33: Discretize the risk area set into multiple sub-regions, obtain the coordinate range and risk level of each sub-region, and construct a fitness function based on the risk area set and the risk level of the sub-regions; Step S34: Using the fitness function as the evaluation criterion, perform population initialization, selection, crossover and mutation operations to regenerate the powder filling model, update the material property distribution in the finite element model, and obtain the crushing model with improved bonding strength. Step S35: Apply the same compressive load in the same direction to the crushed model with improved bonding strength through finite element analysis, obtain new interface stress distribution data, and determine whether the remaining separation risk area is an empty set.

8. The method according to claim 7, characterized in that, Step S33 includes: The risk area is discretized into multiple sub-regions using a grid partitioning method. The coordinate range of each sub-region is determined by the grid number. The risk level is divided into three levels: high, medium, and low, based on the degree of tensile stress exceeding the standard.

9. The method according to claim 8, characterized in that, Step S33 further includes: The fitness function aims to minimize the tensile stress in all sub-regions, and a genetic algorithm is used with the powder spatial arrangement parameters as chromosome genes.

10. The method according to claim 1, characterized in that, Step S4 includes: Step S41: Extract mechanical strength data from the arrangement configuration that enhances bonding strength, separate key indicators, and obtain a preliminary strength dataset. Step S42: Classify the preliminary strength dataset to determine the core parameter set related to the binding strength; Step S43: For the core parameter group, implement an iterative verification process. If the verification result does not match the preset parameter threshold, adjust the parameter weights to obtain the optimized parameter combination. Step S44: Extract key elements of dry molding from the optimized parameter combination to generate a molding parameter set suitable for multi-layer composite structures; Step S45: Based on the molding parameter set, construct a path planning model for electrode pressing and obtain path distribution data during the pressing process. Step S46: Analyze the potential risks of position offset using path distribution data. If the offset exceeds the preset range, generate a corresponding control scheme. Step S47: Based on the generated control scheme, output the final position offset adjustment strategy and determine the pressing execution scheme of the multilayer composite electrode.

Images