Method for uniform roughening of stainless steel meshes

Abstract

The application is suitable for the technical field of stainless steel mesh processing, and particularly relates to a uniform roughening treatment method for a stainless steel mesh, which comprises the following steps: collecting profile point cloud data of the stainless steel mesh in real time; performing feature extraction on the stainless steel mesh and encoding to obtain a material identification code; obtaining initial process parameters based on the material identification code; determining a roughness feature vector and a uniformity index according to the profile point cloud data; obtaining a global feature vector of the stainless steel mesh according to the initial process parameters and the roughness feature vector; in the case that the uniformity index is greater than a preset threshold, performing a next process link according to the initial process parameters; in the case that the uniformity index is less than or equal to the preset threshold, obtaining optimal process parameters of the next process link according to the roughness feature vector and the global feature vector. Therefore, the method provided by the application can improve the problem of poor uniformity of the roughening treatment of the stainless steel mesh caused by the open loop of the machining process.

Description

Technical Field

[0001] This application belongs to the field of stainless steel mesh processing technology, and particularly relates to a method for uniformly roughening stainless steel mesh. Background Technology

[0002] Existing roughening techniques for stainless steel mesh mainly fall into two categories: physical and chemical methods. Physical methods, represented by sandblasting / shot blasting, use compressed air to propel hard abrasives such as quartz sand at high speed onto the mesh surface, creating a rough texture through physical impact. Mechanical polishing / grinding also employs methods, using abrasive belts or grinding wheels for directional friction; while this can smooth the surface, the uniformity of roughness is difficult to precisely control. Chemical methods, primarily based on chemical etching, utilize acidic solutions such as ferric chloride to selectively dissolve the stainless steel surface, creating a micro-textured structure by controlling temperature and time.

[0003] In existing technologies, the entire processing generally employs fixed process parameters for mass production. This lacks online digital perception of the inherent differences in properties between batches of incoming materials, and also fails to provide real-time, quantitative monitoring and feedback on the surface roughness distribution and evolution trend of stainless steel mesh during processing. Because of the inability to perceive differences in incoming materials, the initial setting of process parameters relies heavily on manual experience, often deviating from the optimal processing window for the current material, resulting in poor quality consistency. During processing, naturally occurring spatial non-uniform disturbances such as the boundary effect of sandblasting, the edge concentration effect of electric field in electroplating, and the dead zone of solution flow in chemical etching cannot be identified and corrected in real time. This continuously amplifies the roughness differences between the central and edge areas of the same mesh, and between straight sections of the wire and weaving nodes, ultimately leading to severe loss of uniformity control. Therefore, existing technologies suffer from poor uniformity in the roughening treatment of stainless steel mesh due to the open-loop processing. Summary of the Invention

[0004] This application provides a method for uniformly roughening stainless steel mesh, which can improve the problem of poor uniformity in the roughening process of stainless steel mesh caused by the open-loop processing.

[0005] In a first aspect, embodiments of this application provide a method for uniformly roughening stainless steel mesh, comprising: Real-time acquisition of the outline point cloud data of stainless steel mesh; The stainless steel mesh is feature extracted and encoded to obtain an incoming material identification code; Initial process parameters are obtained based on the incoming material identification code; wherein, the initial process parameters include processing voltage, scanning speed, solution flow rate, evaporation power, sulfidation time, grinding pressure, and feed speed; The roughness feature vector and uniformity index are determined based on the contour point cloud data; The global feature vector of the stainless steel mesh is obtained based on the initial process parameters and the roughness feature vector; wherein, the global feature vector is used to reflect the uniformity of the overall roughness distribution in the processing area of ​​the stainless steel mesh. If the uniformity index is determined to be greater than a preset threshold, the next process step is performed according to the initial process parameters. If the uniformity index is determined to be less than or equal to the preset threshold, the optimal process parameters for the next process step are obtained based on the roughness feature vector and the global feature vector.

[0006] The technical solutions described in this application embodiment have at least the following technical effects: The stainless steel mesh uniform roughening method provided in this application involves: real-time acquisition of the stainless steel mesh contour point cloud data; feature extraction and encoding of the stainless steel mesh to obtain an incoming material identification code; initial process parameters based on the incoming material identification code; determination of roughness feature vector and uniformity index based on the contour point cloud data; obtaining the global feature vector of the stainless steel mesh based on the initial process parameters and the roughness feature vector; proceeding to the next process step based on the initial process parameters when the uniformity index is determined to be greater than a preset threshold; and obtaining the optimal process parameters for the next process step based on the roughness feature vector and the global feature vector when the uniformity index is determined to be less than or equal to the preset threshold. Therefore, the stainless steel mesh uniform roughening method provided in this application ensures batch stability through adaptive initialization of incoming materials and ensures dynamic correction within batches through online closed-loop control, which helps reduce unnecessary waiting and rework. While improving product uniformity and consistency, it does not sacrifice efficiency; instead, by reducing scrap, rework, and manual intervention, it helps achieve simultaneous improvement in quality and efficiency.

[0007] In one possible implementation of the first aspect, obtaining the global feature vector of the stainless steel mesh based on the initial process parameters and the roughness feature vector includes: The processing area of ​​the stainless steel mesh is divided into spatially continuous grid units; Based on the initial process parameters, the roughness feature vector, and the adjacency relationship of the mesh cells, an adjacency matrix of a graph structure is constructed. Calculate the local feature vector of each grid cell based on the adjacency matrix, which incorporates neighborhood context information; A global pooling operation is performed based on the local feature vector to generate the global feature vector.

[0008] In one possible implementation of the first aspect, determining the roughness feature vector and uniformity index based on the contour point cloud data includes: The roughness feature vector is calculated based on the contour point cloud data; wherein, the roughness feature vector includes the arithmetic mean deviation, maximum height, skewness, and kurtosis of the contour; Calculate the local deviation value of each mesh element based on the roughness feature vector; The uniformity index is generated based on the coefficient of variation of each of the local deviation values.

[0009] In one possible implementation of the first aspect, obtaining the optimal process parameters for the next process step based on the roughness feature vector and the global feature vector includes: The process parameter adjustment amount for the next process step is obtained based on the roughness feature vector and the global feature vector. The optimal process parameters for the next process step are obtained based on the adjustment amount of the process parameters.

[0010] In one possible implementation of the first aspect, obtaining the process parameter adjustment amount for the next process step based on the roughness feature vector and the global feature vector includes: If the next process step is determined to be the vapor deposition step, the first feedforward compensation amount is obtained based on the roughness feature vector and the global feature vector. Obtain the thickness of the silver layer on the stainless steel mesh; The first thickness deviation is obtained based on the silver layer thickness; The process parameter adjustment amount is obtained based on the first thickness deviation and the first feedforward compensation amount.

[0011] In one possible implementation of the first aspect, the step of obtaining the process parameter adjustment amount for the next process step based on the roughness feature vector and the global feature vector further includes: If the next process step is determined to be the blackening step, the second feedforward compensation amount is obtained based on the silver layer thickness, the roughness feature vector, and the global feature vector. The thickness of the silver sulfide layer was obtained by acquiring the reflectance spectrum and inverting it. The second thickness deviation is obtained based on the thickness of the silver sulfide layer; The process parameter adjustment amount is obtained based on the second thickness deviation and the second feedforward compensation amount.

[0012] In one possible implementation of the first aspect, the step of obtaining the process parameter adjustment amount for the next process step based on the roughness feature vector and the global feature vector further includes: If the next process step is determined to be a polishing step, the process parameter adjustment amount is obtained based on the thickness of the silver sulfide layer and the roughness feature vector.

[0013] In one possible implementation of the first aspect, obtaining the optimal process parameters for the next process step based on the roughness feature vector and the global feature vector further includes: The partial derivative vector of the roughness feature vector with respect to the initial process parameters is used as the parameter sensitivity feature. The risk confidence level is obtained based on the local feature vector and the parameter sensitivity feature; The optimal process parameters for the next process step are generated based on the global feature vector and the risk confidence level.

[0014] In one possible implementation of the first aspect, obtaining the initial process parameters based on the incoming material identification code includes: The initial process parameters are obtained by iteratively updating the parameters based on the preset target roughness and the incoming material identification code using the gradient descent method.

[0015] In one possible implementation of the first aspect, the step of extracting features from the stainless steel mesh and encoding them to obtain an incoming material identification code includes: Multi-point sampling was performed on the surface of the stainless steel mesh to obtain the element content vector and the surface roughness baseline value; The incoming material identification code is obtained based on the element content vector and the surface roughness baseline value.

[0016] Secondly, embodiments of this application provide a stainless steel mesh uniform roughening treatment apparatus, comprising: The data acquisition module is used to acquire the contour point cloud data of stainless steel mesh in real time. The incoming material identification code module is used to extract features from the stainless steel mesh and encode them to obtain an incoming material identification code; The initial process parameter module is used to obtain initial process parameters based on the incoming material identification code; wherein, the initial process parameters include processing voltage, scanning speed, solution flow rate, evaporation power, sulfidation time, grinding pressure, and feed speed; The roughness and uniformity module is used to determine the roughness feature vector and uniformity index based on the contour point cloud data. A global feature vector module is used to obtain a global feature vector of the stainless steel mesh based on the initial process parameters and the roughness feature vector; wherein, the global feature vector is used to reflect the overall roughness distribution uniformity of the processed area of ​​the stainless steel mesh; The next process step module is used to perform the next process step according to the initial process parameters when it is determined that the uniformity index is greater than a preset threshold. The optimal process parameter module is used to obtain the optimal process parameters for the next process step based on the roughness feature vector and the global feature vector when the uniformity index is determined to be less than or equal to the preset threshold.

[0017] Thirdly, embodiments of this application provide a roughening processing apparatus, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any one of the first aspects above.

[0018] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in any of the first aspects above.

[0019] Fifthly, embodiments of this application provide a computer program product that, when run on a roughening processing device, causes the roughening processing device to perform the method described in any one of the first aspects.

[0020] It is understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic flowchart of a method for uniform roughening of stainless steel mesh provided in an embodiment of this application; Figure 2 This is a schematic diagram of the implementation process of steps S200, S300, S400 and S500 in the uniform roughening treatment method for stainless steel mesh provided in an embodiment of this application. Figure 3 This is a schematic diagram of the implementation process of steps S700 and S710 in the uniform roughening treatment method for stainless steel mesh provided in an embodiment of this application. Figure 4 This is a schematic diagram of another implementation process of step S700 in the uniform roughening treatment method for stainless steel mesh provided in one embodiment of this application; Figure 5 This is a schematic diagram of the roughness feature vector in a uniform roughening treatment method for stainless steel mesh provided in an embodiment of this application; Figure 6 This is a schematic diagram of the silver layer thickness and silver sulfide layer thickness in a uniform roughening treatment method for stainless steel mesh provided in an embodiment of this application. Figure 7 This is a schematic diagram of the stainless steel mesh uniform roughening treatment device provided in the embodiments of this application; Figure 8 This is a schematic diagram of the roughening treatment equipment provided in the embodiments of this application. Detailed Implementation

[0023] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0024] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0025] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0026] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0027] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0028] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0029] In related technologies, the entire processing generally employs fixed process parameters for mass production. This lacks online digital perception of the inherent differences in properties between batches of incoming materials, and also fails to provide real-time, quantitative monitoring and feedback on the surface roughness distribution and evolution trend of stainless steel mesh during processing. Because of the inability to perceive differences in incoming materials, the initial setting of process parameters relies heavily on manual experience, often deviating from the optimal processing window for the current material, resulting in poor quality consistency. During processing, naturally occurring spatial non-uniform disturbances such as the boundary effect of sandblasting, the edge concentration effect of electric field in electroplating, and the dead zone of solution flow in chemical etching cannot be identified and corrected in real time. This causes the roughness differences between the central and edge areas of the same mesh, and between straight sections of the wire and weaving nodes, to be continuously amplified, ultimately leading to severe loss of uniformity control. Therefore, existing technologies suffer from poor uniformity in the roughening treatment of stainless steel mesh due to the open-loop processing.

[0030] To address the aforementioned issues, this application provides a method for uniform roughening of stainless steel mesh. This method involves: real-time acquisition of the stainless steel mesh's contour point cloud data; feature extraction and encoding of the stainless steel mesh to obtain an incoming material identification code; initial process parameters based on the incoming material identification code; determination of a roughness feature vector and a uniformity index based on the contour point cloud data; obtaining a global feature vector of the stainless steel mesh based on the initial process parameters and the roughness feature vector; proceeding to the next process step based on the initial process parameters when the uniformity index is determined to be greater than a preset threshold; and obtaining the optimal process parameters for the next process step based on the roughness feature vector and the global feature vector when the uniformity index is determined to be less than or equal to the preset threshold. Therefore, the stainless steel mesh uniform roughening method provided in this application ensures batch stability through adaptive initialization of incoming materials and ensures dynamic correction within batches through online closed-loop control, which helps reduce unnecessary waiting and rework. While improving product uniformity and consistency, it does not sacrifice efficiency; instead, by reducing scrap, rework, and manual intervention, it helps achieve simultaneous improvement in quality and efficiency.

[0031] The stainless steel mesh uniform roughening treatment method provided in this application embodiment can be applied to roughening treatment equipment. In this case, the roughening treatment equipment is the main body for executing the stainless steel mesh uniform roughening treatment method provided in this application embodiment. This application embodiment does not impose any restrictions on the specific type of roughening treatment equipment.

[0032] For example, roughening equipment can be an industrial computer, programmable logic controller, embedded control system, distributed control system, tablet computer, laptop computer, ultra-mobile personal computer (UMPC), netbook, desktop computer, laptop computer, handheld computing device, etc., but is not limited to these.

[0033] To better understand the uniform roughening treatment method for stainless steel mesh provided in this application embodiment, the specific implementation process of the uniform roughening treatment method for stainless steel mesh provided in this application embodiment will be described by way of example below.

[0034] Figure 1 A schematic flowchart of a method for uniformly roughening stainless steel mesh according to an embodiment of this application is shown. The method includes: S100 collects the contour point cloud data of stainless steel mesh in real time.

[0035] For example, lidar (such as the Velodyne VLP-16) can achieve centimeter-level ranging accuracy within a 150-meter range by emitting pulsed lasers and measuring the echo time. Structured light systems (such as the Creaform HandySCAN 3D) achieve 0.03mm repeatability accuracy within a 0.1-2 meter range by projecting dynamic speckle patterns. The overall outline of a stainless steel mesh can be obtained first using lidar, and then detailed scanning of the mesh nodes can be performed using structured light. Voxel filtering (voxel_size=0.005m) and statistical filtering using the Open3D library are used to remove outliers, resulting in contour point cloud data containing coordinate and reflection intensity information.

[0036] S200 extracts and encodes features from stainless steel mesh to obtain an incoming material identification code.

[0037] For example, geometric and texture features can be extracted from contour point cloud data, and a material identification code can be generated using hash encoding. Geometric features include grid spacing, aperture size, surface curvature, etc. Principal component features are extracted using the PCA algorithm to obtain geometric features; texture features are obtained by calculating parameters such as contrast and energy using GLCM (Gray-Level Co-occurrence Matrix). The feature vectors are then standardized and input into the SHA-256 hash function to generate the material identification code.

[0038] S300 obtains initial process parameters based on the incoming material identification code. These initial process parameters include processing voltage, scanning speed, solution flow rate, evaporation power, vulcanization time, grinding pressure, and feed speed.

[0039] For example, initial parameters can be obtained by querying the process parameter database based on the identification code. The process parameter database is built using MySQL and contains the mapping relationship between seven types of parameters, such as voltage, speed, and flow rate, and the identification codes. For example, the parameters corresponding to the incoming material identification code "a3f7b9e8" are: processing voltage 220V±5V, scanning speed 45mm / s, solution flow rate 12L / min, evaporation power 600W, vulcanization time 25min, grinding pressure 18N, and feed speed 8mm / s.

[0040] S400 determines the roughness feature vector and uniformity index based on the contour point cloud data.

[0041] For example, the surface profile curve can be measured by a profilometer and the roughness feature vector can be calculated using the profile arithmetic mean deviation method, and the uniformity index can be calculated using the Shannon-Wiener model.

[0042] S500, based on the initial process parameters and roughness feature vector, obtains the global feature vector of the stainless steel mesh. The global feature vector reflects the overall uniformity of roughness distribution in the processed area of ​​the stainless steel mesh.

[0043] For example, a convolutional neural network (CNN) can be used to fuse initial process parameters and roughness features. The input consists of initial process parameters and roughness feature vectors, and the output is a global feature vector. The training data for the CNN comes from historical process records (containing process parameters, roughness feature vectors, and corresponding global feature vectors). For instance, the Adam optimizer (learning rate 0.001) can be used for 50 training iterations, with the loss function being MSE.

[0044] S600: If the uniformity index is determined to be greater than the preset threshold, proceed to the next process step according to the initial process parameters.

[0045] For example, the uniformity index of the current process step can be determined based on the real-time acquired contour point cloud data. If the uniformity index is determined to be greater than a preset threshold (such as 0.7), the initial process parameters can be directly used for the next process step.

[0046] S700, when the uniformity index is determined to be less than or equal to a preset threshold, obtains the optimal process parameters for the next process step based on the roughness feature vector and the global feature vector.

[0047] For example, optimization can be achieved using a genetic algorithm when the uniformity index is less than or equal to a preset threshold (e.g., 0.7). The optimization objective is to maximize the uniformity index, and the constraint is the range of process parameters (e.g., voltage 180-250V). For instance, the population size can be set to 50, the crossover probability to 0.8, the mutation probability to 0.1, and the iteration to 100 generations.

[0048] In one possible implementation, please refer to Figure 2 S500, based on the initial process parameters and roughness feature vector, obtains the global feature vector of the stainless steel mesh, including: S510 divides the processing area of ​​stainless steel mesh into spatially continuous grid units.

[0049] For example, an adaptive voxel mesh generation algorithm can be used to generate spatially continuous mesh cells. A 3D mesh model is constructed based on contour point cloud data. The voxel size is set to 1 mm³ using the `voxel_down_sample` function of the Open3D library to generate initial mesh cells. A region growing algorithm is then used to adaptively divide the initial mesh cells to obtain spatially continuous mesh cells. For example, using the stainless steel mesh aperture size (3 mm × 5 mm) as a reference, the minimum mesh cell size is set to 2 mm × 2 mm, and the maximum size to 8 mm × 8 mm, ensuring that each mesh cell contains at least 50 point cloud data points.

[0050] S520 constructs an adjacency matrix of a graph structure based on initial process parameters, roughness feature vectors, and the adjacency relationships of mesh cells.

[0051] For example, a weighted adjacency matrix can be constructed based on the spatial adjacency of grid cells. Two grid cells are considered adjacent if they share at least one edge. The Delaunay triangulation algorithm is used to calculate the spatial distance between grid cells, and the adjacency matrix element values ​​A are constructed. ij ={If meshes i and j are adjacent, then 1; otherwise, 0}, the adjacency matrix is ​​weighted based on the initial process parameters and roughness feature vector: A ij ′ =A ij ×(1+α ΔRa), where α=0.2 is the roughness influence factor, and ΔRa is the roughness difference between adjacent mesh elements calculated based on the roughness feature vector.

[0052] S530 calculates the local feature vector of each grid cell based on the adjacency matrix, which incorporates neighborhood context information.

[0053] For example, a graph convolutional network (GCN) can be used to fuse neighborhood context information based on the adjacency matrix to obtain local feature vectors. For instance, the GCN layer employs the following propagation rule: =σ( W (l) +b (l) ), where N(i) is the set of neighboring nodes (representing the set of all adjacent grid cells of grid cell i), d i W represents the node degree (the number of neighboring nodes of grid cell i). (l) and b (l) These are learnable parameters (which can be initialized and iteratively updated during training).

[0054] S540 performs a global pooling operation based on the local feature vectors to generate a global feature vector.

[0055] For example, hierarchical attention pooling can be used to aggregate global features and generate a global feature vector. All local feature vectors can be grouped, pooled within each group, and then global attention pooling can be used to generate the global feature vector.

[0056] Through steps S510 to S540, adaptive mesh generation and graph structure modeling facilitate the extraction and fusion of spatially continuous local features. Weighted adjacency matrices and Global Convergence Neighborhood (GCN) enable deep correlation between process parameters, roughness features, and spatial location. Neighborhood information fusion allows for dynamic adjustment of local process parameters, which helps reduce the roughness standard deviation. The global feature vector directly reflects the overall state of the processing area, which helps shorten decision-making time. Adaptive mesh generation and graph structure modeling enable the system to handle stainless steel meshes of different sizes and shapes, improving its versatility.

[0057] In one possible implementation, please refer to Figure 2 S400, based on the contour point cloud data, determines the roughness feature vector and uniformity index, including: S410, the roughness feature vector is calculated based on the contour point cloud data. The roughness feature vector includes the arithmetic mean deviation of the contour, the maximum height, the skewness, and the kurtosis.

[0058] For example, a roughness feature vector can be obtained by combining Gaussian filtering with parameter calculation. First, Gaussian filtering is applied to the input contour point cloud data for noise reduction. After noise reduction, the arithmetic mean deviation is obtained by calculating the mean absolute distance from each point to the centerline: Ra = , where y i Let be the ordinate of the point. The vertical coordinate is the midline, and N is the number of points; the maximum height is obtained by taking the average of the maximum peak and valley heights within 5 consecutive sampling lengths: Rz = y p,k -y v,k ), where y p,k and y v,k The ordinates of the highest peak and lowest valley within the k-th sampling segment; skewness Rsk = - ) 3 ;Kurtitude Rku= - ) 4 ,like Figure 5 As shown.

[0059] S420 calculates the local deviation value of each mesh element based on the roughness feature vector.

[0060] For example, the contour point cloud data can be spatially matched with the grid cells. The relationship between the ordinate and abscissa of the point cloud within the grid cell can be fitted using a quadratic polynomial. The arithmetic mean deviation can be calculated only for the points within the grid cell. The grid cell can be divided into multiple sub-regions. The maximum peak and valley height of each sub-region can be taken and the average value can be calculated to obtain the local deviation value of each grid cell.

[0061] S430 generates a uniformity index based on the coefficient of variation of each local deviation value.

[0062] For example, the coefficient of variation can be calculated for the local deviation values ​​of all grid cells: CV = σ / μ × 100%, where σ is the standard deviation of the local deviation values ​​of all grid cells, and μ is the mean of the local deviation values ​​of all grid cells. The uniformity index E = 1 − (w1) CV Ra / CV max +w2 CV Rz / CV max ), where w1 and w2 are weights (e.g., w1=0.6, w2=0.4), CV max Set the maximum allowable coefficient of variation (e.g., 30%).

[0063] Through steps S410 to S430, Gaussian filtering is used to eliminate measurement noise, ensuring the stability of parameter calculations. Four parameters quantify surface roughness from different dimensions: Ra reflects overall deviation, Rz reflects local extreme heights, and Rsk and Rku describe the distribution pattern. Compared to traditional methods that only calculate Ra, this scheme provides a more comprehensive description of surface features, offering multi-dimensional support for subsequent process optimization. Spatial resolution is improved through mesh cell division. Local centerline fitting considers the curvature trend of the contour (such as the curved edges of stainless steel mesh), which helps avoid local errors caused by global centerline fitting. The degree of dispersion is quantified through the coefficient of variation, and a weighted scoring is introduced to make the evaluation results more closely reflect actual needs.

[0064] In one possible implementation, please refer to Figure 3 S700, based on the roughness eigenvector and the global eigenvector, obtains the optimal process parameters for the next process step, including: S710, based on the roughness feature vector and the global feature vector, obtains the process parameter adjustment amount for the next process step.

[0065] For example, the roughness feature vector and the global feature vector can be normalized and concatenated into a joint feature vector as model input. A pre-trained SVR model (training data is historical process data, including the joint feature vector obtained from the roughness feature vector and the global feature vector and the corresponding process parameter adjustment amount. A radial basis function (RBF) kernel is used as the kernel function of the SVR model, the SVR model is trained using the Sequential Minimum Optimization (SMO) algorithm, and the hyperparameters of the SVR model are tuned using 5-fold cross-validation) is used to predict the process parameter adjustment amount of the next process step.

[0066] S720 obtains the optimal process parameters for the next process step based on the adjustment amount of the process parameters.

[0067] For example, the current process parameters of the next process step can be updated according to the process parameter adjustment amount. Check whether the updated process parameters are within the preset allowable range of the equipment. If they are outside the range, generate multiple candidate parameter combinations within the preset allowable range, calculate the corresponding Ra and Rz, and select the combination that makes both lower as the optimal parameter.

[0068] Through steps S710 to S720 above, by fusing roughness and global features, and utilizing SVR to capture nonlinear mapping relationships, the accuracy of adjustment amounts is improved. Parameter feasibility is checked through boundary constraints to avoid equipment overload or process failure.

[0069] Optionally, please refer to Figure 3 S710, based on the roughness eigenvector and the global eigenvector, obtains the process parameter adjustment amount for the next process step, including: S711, when it is determined that the next process step is the vapor deposition step, the first feedforward compensation amount is obtained based on the roughness feature vector and the global feature vector.

[0070] For example, if the next process step is determined to be vapor deposition, the roughness feature vector and the global feature vector can be normalized and concatenated into a joint feature vector as model input. The first feedforward compensation amount is then predicted using a pre-trained SVR model (training data consists of historical process data, including the joint feature vector and corresponding process parameter adjustments. A radial basis function (RBF) kernel is used as the kernel function of the SVR model, and the SVR model is trained using the Sequential Minimum Optimization (SMO) algorithm. Combined with 5-fold cross-validation, the hyperparameters of the SVR model are tuned).

[0071] S712, obtain the silver layer thickness of the stainless steel mesh.

[0072] For example, non-destructive testing can be performed using X-ray fluorescence spectroscopy (XRF), where the thickness of the silver layer on the stainless steel mesh is calculated and recorded by measuring the characteristic X-ray intensity of silver. Figure 6 As shown.

[0073] S713, the first thickness deviation is obtained based on the silver layer thickness.

[0074] For example, the silver layer thickness can be compared with a preset first target thickness (e.g., 0.25 μm) to calculate the first thickness deviation.

[0075] S714, the process parameter adjustment amount is obtained based on the first thickness deviation and the first feedforward compensation amount.

[0076] For example, the process parameter adjustment amount can be obtained by weighted summation based on the first thickness deviation and the first feedforward compensation amount. For example, the process parameter adjustment amount ΔP = α ΔC+β K p Δd, where α and β are weighting coefficients (α+β=1), K p For feedback control gain.

[0077] Through steps S711 to S714 above, by fusing roughness and global features and utilizing SVR to capture nonlinear mapping relationships, the accuracy of compensation amounts is improved. Real-time thickness data acquisition via XRF allows for dynamic calculation of deviations, providing a foundation for feedforward-feedback composite control. This feedforward-feedback composite control utilizes feedforward compensation to quickly offset measurable disturbances (such as roughness changes), while feedback control corrects unmodeled errors (such as equipment aging), thus improving control precision.

[0078] Optionally, please refer to Figure 3 S710, based on the roughness feature vector and the global feature vector, obtains the process parameter adjustment amount for the next process step, and also includes: S715, when it is determined that the next process step is the blackening step, the second feedforward compensation amount is obtained based on the silver layer thickness, roughness feature vector and global feature vector.

[0079] For example, if the next process step is determined to be the blackening step, the silver layer thickness, roughness feature vector, and global feature vector can be input, and a support vector regression (SVR) model can be used to establish a nonlinear relationship between the features and the second feedforward compensation amount to obtain the second feedforward compensation amount.

[0080] S716, obtain the reflection spectrum and invert it to obtain the thickness of the silver sulfide layer.

[0081] For example, a fiber optic spectrometer (such as Ocean Optics USB2000+) can be used to acquire the surface reflectance spectrum of a sample, and a mathematical model of the reflectance spectrum and the thickness d of the silver sulfide layer can be established based on the Fresnel formula. Reflectance R(λ) = |(r... 12 +r 23 e −iδ ) / (1+r 12 r 23 e −iδ )| 2 , where r 12 r 23 Let λ be the interface reflection coefficient, δ = 4πn²dcosθ² / λ be the phase difference, n² be the refractive index of silver sulfide (obtainable by ellipsometer measurement or table lookup), and θ² be the refraction angle. The experimental reflection spectrum is fitted to the theoretical model using the least squares method, and the optimization objective function is: min d ∑ λ [R exp (λ)−R theory (λ,d)] 2 For example, the initial value is set to d = 0.1 μm, the iteration step size is set to 0.001 μm, and the termination condition is that the residual is less than 10. −4 Through iterative optimization, the residual was minimized when d=0.18μm, meaning the silver sulfide layer thickness was 0.18μm. The silver sulfide layer thickness was recorded as follows: Figure 6 As shown.

[0082] S717, the second thickness deviation is obtained based on the thickness of the silver sulfide layer.

[0083] For example, the thickness of the silver sulfide layer can be compared with a preset second target thickness (e.g., 0.15 μm) to calculate the second thickness deviation.

[0084] S718, the process parameter adjustment amount is obtained based on the second thickness deviation and the second feedforward compensation amount.

[0085] For example, the process parameter adjustment amount can be obtained by weighted summation based on the second thickness deviation and the second feedforward compensation amount. For example, the process parameter adjustment amount ΔP = α ΔC+β K p Δd, where α and β are weighting coefficients (α+β=1), K p For feedback control gain.

[0086] Through steps S715 to S718 above, multi-dimensional features are fused using SVR to capture nonlinear mapping relationships, which helps improve the accuracy of compensation. Real-time thickness data is acquired using reflectance spectroscopy, allowing for dynamic calculation of deviations and providing a foundation for feedforward-feedback composite control. Feedforward-feedback composite control utilizes feedforward compensation to quickly offset measurable disturbances (such as roughness changes), while feedback control corrects unmodeled errors (such as equipment aging), thus improving control accuracy.

[0087] Optionally, please refer to Figure 3 S710, based on the roughness feature vector and the global feature vector, obtains the process parameter adjustment amount for the next process step, and also includes: S719, when it is determined that the next process step is the grinding step, the process parameter adjustment amount is obtained based on the thickness of the silver sulfide layer and the roughness feature vector.

[0088] For example, if the next process step is determined to be a grinding step, the feature vectors of silver sulfide layer thickness and roughness can be normalized. A fuzzy logic reasoning system can be used to establish a nonlinear mapping relationship between the feature vectors of silver sulfide layer thickness and roughness and the process parameter adjustment amount. The normalized feature vectors of silver sulfide layer thickness and roughness are input into the fuzzy logic reasoning system, and reasoning is performed according to the preset fuzzy rule base to obtain the fuzzy set of output variables. The fuzzy set of output variables is then defuzzified (e.g., centroid method, maximum membership method, etc.) to obtain the process parameter adjustment amount.

[0089] Through the above step S719, the fuzzy logic reasoning system maps the silver sulfide layer thickness and roughness feature vectors to grinding process parameters. This helps to solve the problems of traditional methods relying on experience, ignoring multi-feature coupling, and having poor dynamic adaptability. It is conducive to achieving comprehensive optimization of surface quality, efficiency, and equipment life, and has innovative and engineering application value.

[0090] In one possible implementation, please refer to Figure 4 S700, based on the roughness eigenvector and the global eigenvector, obtains the optimal process parameters for the next process step, and also includes: S730 uses the partial derivative vector of the roughness feature vector with respect to the initial process parameters as the parameter sensitivity feature.

[0091] For example, a polynomial regression model can be used to fit each roughness parameter in the roughness feature vector to a polynomial model (historical process data can be divided into training, validation and test sets in a 7:2:1 ratio, and the polynomial coefficients can be fitted using the least squares method, with the mean square error (MSE) as the objective function). The initial process parameter values ​​are substituted into the polynomial model and the partial derivatives are calculated. All partial derivatives are then concatenated to obtain the parameter sensitivity feature vector.

[0092] For example, if the historical process data does not include some parameters (such as Rsk, Rku) in the roughness feature vector, then the roughness parameters can refer to the remaining parameters (such as Ra, Rz) included in the historical process data. S740, the risk confidence level is obtained based on the local feature vector and parameter sensitivity features.

[0093] For example, the deviation ΔR between the local feature vector and the preset target feature vector can be used as a basis. ij By combining the parameter sensitivity feature S with weighting, the risk index is obtained: Risk i =∑ 4 j=1 |ΔR ij S ij |, where S ij Let be the sensitivity of the j-th roughness parameter to the i-th process parameter. Statistical analysis of risk indicators for all regions (e.g., calculating the mean or maximum value) yields the risk confidence level.

[0094] S750 generates the optimal process parameters for the next process step based on the global feature vector and risk confidence level.

[0095] For example, the optimization objective can be defined as minimizing the risk confidence, while constraining the roughness parameter within the target range (e.g., Ra ∈ [0.8, 1.2] μm). Using process parameters (G, P, v) as variables and risk confidence as the optimization objective, a constrained optimization problem is constructed: min G,P,v Confidence(G,P,v), stRa min ≤f1(G,P,v)≤Ra max Rz min ≤f2(G,P,v)≤Rz max The Bayesian optimization algorithm is used to quickly converge and generate the optimal process parameters for the next process step through iterative sampling and model updates.

[0096] Through steps S730 to S750 above, by quantifying sensitivity, parameters that significantly affect the target roughness (such as high-sensitivity parameters) can be prioritized for adjustment, which helps reduce the number of trial and error attempts and improve optimization efficiency. By combining sensitivity, risks can be assessed more accurately, and potential defect areas can be identified in advance, which helps reduce the scrap rate. By combining global features and risk confidence, dynamic optimization can be achieved, which helps to shorten the grinding time per part and improve surface consistency.

[0097] In one possible implementation, please refer to Figure 2 S300 obtains initial process parameters based on the incoming material identification code, including: S310: Based on the preset target roughness and incoming material identification code, the initial process parameters are obtained by iteratively updating using the gradient descent method.

[0098] For example, the initial process parameter range can be obtained based on the preset target roughness and the incoming material identification code. A neural network model (such as a multilayer perceptron MLP, with ReLU activation function in the hidden layer, linear activation in the output layer (regression task), and mean squared error (MSE) as the loss function, trained using historical process data (including historical process parameters and corresponding surface roughness)) can be used as the conditional input to predict the corresponding surface roughness. The error between the predicted surface roughness and the preset target roughness can be used as the optimization objective. The initial process parameters can be iteratively updated using the gradient descent method.

[0099] By dynamically adjusting parameters using the incoming material feature vector through step S310, adaptability can be improved. Optimizing parameters via gradient descent, directly constrained by the target roughness, helps increase the achievement rate. Reducing the number of experiments through model prediction shortens debugging time.

[0100] In one possible implementation, please refer to Figure 2 S200, features of the stainless steel mesh are extracted and encoded to obtain an incoming material identification code, including: S210, multi-point sampling is performed on the surface of stainless steel mesh to obtain the element content vector and the surface roughness baseline value.

[0101] For example, elemental analysis can be performed on each sampling point using energy-dispersive X-ray spectroscopy (EDS) or inductively coupled plasma mass spectrometry (ICP-MS) to obtain an elemental content vector. The arithmetic mean roughness of all sampling points is measured using a surface roughness meter (such as a stylus or optical type) to obtain a baseline value for surface roughness.

[0102] S220 is an incoming material identification code based on the element content vector and the surface roughness baseline value.

[0103] For example, a hash algorithm (such as MD5, SHA-1) or an encoding rule (such as binary encoding) can be used to map the standardized element content vector and the surface roughness baseline value to obtain the incoming material identification code.

[0104] Through the above steps S210 to S220, the generation of a unique incoming material identification code from multi-point sampling is realized, which helps to solve the problems of incomplete sampling, low incoming material differentiation and low process parameter reuse rate of traditional methods, and is conducive to improving data representativeness, process adaptability and production efficiency.

[0105] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0106] Corresponding to the stainless steel mesh uniform roughening treatment method described in the above embodiments, this application also provides a stainless steel mesh uniform roughening treatment device, the various modules of which can realize the various steps of the stainless steel mesh uniform roughening treatment method. Figure 7 A structural block diagram of the stainless steel mesh uniform roughening treatment device provided in the embodiment of this application is shown. For ease of explanation, only the parts related to the embodiment of this application are shown.

[0107] Reference Figure 7 The device includes: The data acquisition module is used to acquire the contour point cloud data of stainless steel mesh in real time. The incoming material identification code module is used to extract features from the stainless steel mesh and encode them to obtain an incoming material identification code; The initial process parameter module is used to obtain initial process parameters based on the incoming material identification code; wherein, the initial process parameters include processing voltage, scanning speed, solution flow rate, evaporation power, sulfidation time, grinding pressure, and feed speed; The roughness and uniformity module is used to determine the roughness feature vector and uniformity index based on the contour point cloud data. A global feature vector module is used to obtain a global feature vector of the stainless steel mesh based on the initial process parameters and the roughness feature vector; wherein, the global feature vector is used to reflect the overall roughness distribution uniformity of the processed area of ​​the stainless steel mesh; The next process step module is used to perform the next process step according to the initial process parameters when it is determined that the uniformity index is greater than a preset threshold. The optimal process parameter module is used to obtain the optimal process parameters for the next process step based on the roughness feature vector and the global feature vector when the uniformity index is determined to be less than or equal to the preset threshold.

[0108] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0109] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above device can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0110] This application also provides a roughening treatment device. Figure 8 This is a schematic diagram of the roughening treatment apparatus provided in one embodiment of this application. Figure 8 As shown, the roughening processing apparatus 8 of this embodiment includes: at least one processor 80 ( Figure 8 Only one is shown in the image), at least one memory 81 ( Figure 8 (Only one is shown in the image) and a computer program 82 stored in the at least one memory 81 and executable on the at least one processor 80, wherein when the processor 80 executes the computer program 82, it causes the roughening treatment device 8 to perform the steps in any of the above embodiments of the uniform roughening treatment method for stainless steel mesh, or causes the roughening treatment device 8 to perform the functions of each module / unit in the above embodiments of the apparatus.

[0111] For example, the computer program 82 may be divided into one or more modules / units, which are stored in the memory 81 and executed by the processor 80 to complete this application. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program 82 in the roughening processing device 8.

[0112] The roughening processing device 8 can be a computing device such as an industrial computer, programmable logic controller, embedded control system, distributed control system, desktop computer, laptop, handheld computer, and cloud server. This roughening processing device may include, but is not limited to, a processor 80 and a memory 81. Those skilled in the art will understand that... Figure 8 This is merely an example of the roughening processing device 8 and does not constitute a limitation on the roughening processing device 8. It may include more or fewer components than shown, or combine certain components, or different components, such as input / output devices, network access devices, buses, etc.

[0113] The processor 80 can be a Central Processing Unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0114] In some embodiments, the memory 81 may be an internal storage unit of the roughening processing apparatus 8, such as a hard disk or memory of the roughening processing apparatus 8. In other embodiments, the memory 81 may be an external storage device of the roughening processing apparatus 8, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the roughening processing apparatus 8. Furthermore, the memory 81 may include both internal storage units and external storage devices of the roughening processing apparatus 8. The memory 81 is used to store operating systems, applications, bootloaders, data, and other programs, such as the program code of computer programs. The memory 81 can also be used to temporarily store data that has been output or will be output.

[0115] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0116] This application provides a computer program product that, when run on a roughening processing device, causes the roughening processing device to implement the steps in any of the above method embodiments.

[0117] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a roughening processing device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, such as a USB flash drive, a portable hard drive, a magnetic disk, or an optical disk.

[0118] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0119] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0120] In the embodiments provided in this application, it should be understood that the disclosed roughening processing apparatus and method can be implemented in other ways. For example, the roughening processing apparatus embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0121] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0122] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A stainless steel mesh uniform roughening treatment method characterized by, include: Real-time acquisition of the outline point cloud data of stainless steel mesh; The stainless steel mesh is feature extracted and encoded to obtain an incoming material identification code; Initial process parameters are obtained based on the incoming material identification code; wherein, the initial process parameters include processing voltage, scanning speed, solution flow rate, evaporation power, sulfidation time, grinding pressure, and feed speed; The roughness feature vector and uniformity index are determined based on the contour point cloud data; The global feature vector of the stainless steel mesh is obtained based on the initial process parameters and the roughness feature vector; wherein, the global feature vector is used to reflect the uniformity of the overall roughness distribution in the processing area of ​​the stainless steel mesh. If the uniformity index is determined to be greater than a preset threshold, the next process step is performed according to the initial process parameters. If the uniformity index is determined to be less than or equal to the preset threshold, the optimal process parameters for the next process step are obtained based on the roughness feature vector and the global feature vector.

2. The stainless steel mesh uniform roughening treatment method according to claim 1, characterized by, The step of obtaining the global feature vector of the stainless steel mesh based on the initial process parameters and the roughness feature vector includes: The processing area of ​​the stainless steel mesh is divided into spatially continuous grid units; Based on the initial process parameters, the roughness feature vector, and the adjacency relationship of the mesh cells, an adjacency matrix of a graph structure is constructed. Calculate the local feature vector of each grid cell based on the adjacency matrix, which incorporates neighborhood context information; A global pooling operation is performed based on the local feature vector to generate the global feature vector.

3. The method for uniformly roughening a stainless steel mesh according to claim 2, characterized by, The step of determining the roughness feature vector and uniformity index based on the contour point cloud data includes: The roughness feature vector is calculated based on the contour point cloud data; wherein, the roughness feature vector includes the arithmetic mean deviation, maximum height, skewness, and kurtosis of the contour; Calculate the local deviation value of each mesh element based on the roughness feature vector; The uniformity index is generated based on the coefficient of variation of each of the local deviation values.

4. The method for uniformly roughening stainless steel mesh according to claim 1, characterized by, The step of obtaining the optimal process parameters for the next process step based on the roughness feature vector and the global feature vector includes: The process parameter adjustment amount for the next process step is obtained based on the roughness feature vector and the global feature vector. The optimal process parameters for the next process step are obtained based on the adjustment amount of the process parameters.

5. The method for uniformly roughening a stainless steel mesh according to claim 4, characterized by, The step of obtaining the process parameter adjustment amount for the next process step based on the roughness feature vector and the global feature vector includes: If the next process step is determined to be the vapor deposition step, the first feedforward compensation amount is obtained based on the roughness feature vector and the global feature vector. Obtain the thickness of the silver layer on the stainless steel mesh; The first thickness deviation is obtained based on the silver layer thickness; The process parameter adjustment amount is obtained based on the first thickness deviation and the first feedforward compensation amount.

6. The method for uniformly roughening a stainless steel mesh according to claim 5, wherein The step of obtaining the process parameter adjustment amount for the next process step based on the roughness feature vector and the global feature vector further includes: If the next process step is determined to be the blackening step, the second feedforward compensation amount is obtained based on the silver layer thickness, the roughness feature vector, and the global feature vector. The thickness of the silver sulfide layer was obtained by acquiring the reflectance spectrum and inverting it. The second thickness deviation is obtained based on the thickness of the silver sulfide layer; The process parameter adjustment amount is obtained based on the second thickness deviation and the second feedforward compensation amount.

7. The method of uniform roughening of stainless steel mesh according to claim 6, characterized by, The step of obtaining the process parameter adjustment amount for the next process step based on the roughness feature vector and the global feature vector further includes: If the next process step is determined to be a polishing step, the process parameter adjustment amount is obtained based on the thickness of the silver sulfide layer and the roughness feature vector.

8. The method for uniformly roughening stainless steel mesh according to claim 2, wherein The step of obtaining the optimal process parameters for the next process step based on the roughness feature vector and the global feature vector further includes: The partial derivative vector of the roughness feature vector with respect to the initial process parameters is used as the parameter sensitivity feature. The risk confidence level is obtained based on the local feature vector and the parameter sensitivity feature; The optimal process parameters for the next process step are generated based on the global feature vector and the risk confidence level.

9. The method for uniformly roughening stainless steel mesh according to claim 1, wherein The process of obtaining initial process parameters based on the incoming material identification code includes: The initial process parameters are obtained by iteratively updating the parameters based on the preset target roughness and the incoming material identification code using the gradient descent method.

10. The method for uniformly roughening stainless steel mesh according to claim 1, wherein The step of extracting features from the stainless steel mesh and encoding them to obtain the incoming material identification code includes: Multi-point sampling was performed on the surface of the stainless steel mesh to obtain the element content vector and the surface roughness baseline value; The incoming material identification code is obtained based on the element content vector and the surface roughness baseline value.

Images