Titanium alloy surface electroplating control method

Abstract

The invention belongs to the technical field of electroplating process optimization, and particularly discloses a titanium alloy surface electroplating control method which comprises the following steps: dynamically generating a hanger posture configuration scheme based on workpiece three-dimensional contour data acquired in real time; a bubble distribution image and electroplating liquid flow field data are synchronously collected, and a bubble-flow field coupling feature set is generated through multi-modal fusion; performing real-time linkage adjustment on the rotation speed and the swing amplitude of the workpiece according to the coupling feature set; after electroplating is completed, plating thickness, edge integrity and surface defect data are collected, and a multi-source plating data set is formed by combining workpiece position information; historical data and real-time data are integrated to carry out hanger layout adjustment judgment, and hanger spacing dynamic reconstruction optimization is started when adjustment is needed. According to the method, multi-parameter cooperative control and closed-loop quality optimization in the electroplating process are achieved, the uniformity and integrity of a titanium alloy workpiece coating are remarkably improved, and generation of bubble related defects is effectively restrained.

Description

Technical Field

[0001] This invention belongs to the field of electroplating process optimization technology, and specifically relates to a method for controlling electroplating on titanium alloy surfaces. Background Technology

[0002] Titanium alloys are widely used in aerospace, marine engineering, medical devices, and high-end equipment manufacturing due to their high strength, low density, excellent corrosion resistance, and high-temperature mechanical properties. However, titanium alloys also have some inherent defects on their surface, so electroplating is often used to improve them. The control of the electroplating process directly affects the quality of the final product.

[0003] Prior art 1, such as the electroplating equipment and method disclosed in Chinese invention patent application No. 201810487847.5, involves forming an independently controllable electric field on the wafer surface using multiple electrodes. By applying independently adjustable electric field strengths to different regions, especially at wafer notches, the total charge received by the notch region during the electroplating process is directly reduced, thereby precisely controlling the electroplating height in that region. Compared to the traditional method that relies on adjusting the wafer rotation speed, this improves the accuracy and reliability of control while increasing electroplating efficiency.

[0004] Prior art 2, such as the chip electroplating system and chip electroplating control method disclosed in Chinese invention patent application No. 202011615619.5, uses a pneumatic cylinder to drive a pusher plate to move in the electroplating solution to adjust the liquid level. At the same time, it controls the chip clamping device to move in coordination with the pusher plate, so that when the chip to be plated is immersed in the electroplating solution, opposing pressure is formed on the chip surface, which enhances the expulsion of attached air bubbles, thereby reducing the air residue on the chip surface during the electroplating process and improving the defects of copper pits in electroplating.

[0005] The two existing technical solutions mentioned above still have the following shortcomings in the control of the electroplating process: 1. Existing technology 1 focuses on adjusting the electroplating height of the wafer notch area by independently controlling the electric field through multiple electrodes, but its control is mainly based on the preset electric field distribution, which is a feedforward control. It lacks real-time monitoring and feedback adjustment of the workpiece surface state and bubbles, making it difficult to achieve the expected adjustment effect.

[0006] 2. Existing technology 1 improves the deposition uniformity at the wafer notch by adjusting the electric field, but its focus is limited to specific areas, such as the electroplating height of the notch, and does not fully consider the problem of coating thickness uniformity of workpieces at different positions on the same fixture.

[0007] 3. Existing technology 2 uses mechanical means, namely, the push plate and clamping to generate opposing pressure to expel bubbles. However, this method is relatively crude and cannot accurately locate the distribution characteristics and size of bubbles, making it difficult to guarantee the effectiveness of the elimination method. Its bubble removal effect and adaptability may be limited.

[0008] 4. Both existing patents focus on a single control dimension, such as electric field and current, and lack multi-parameter collaborative optimization. The single control dimension cannot comprehensively improve the electroplating quality. Summary of the Invention

[0009] In view of this, in order to solve the above problems, a method for controlling electroplating on titanium alloy surfaces is proposed.

[0010] The objective of this invention can be achieved through the following technical solution: This invention provides a method for controlling electroplating on the surface of titanium alloys, the method comprising: S1, dynamically generating a fixture posture configuration scheme that matches the geometric features of the workpiece based on the collected three-dimensional contour data of the titanium alloy workpiece.

[0011] S2. Immerse the workpiece in the electroplating solution according to the fixture posture configuration scheme, and simultaneously collect bubble distribution images and electroplating solution flow field data on the workpiece surface during the electroplating process. Generate a bubble-flow field coupling feature set through multimodal data fusion.

[0012] S3. Based on the bubble-flow field coupling feature set, adjust the rotation speed and oscillation amplitude of the workpiece in real time.

[0013] S4. When electroplating is completed, simultaneously collect coating thickness distribution data, edge integrity image data, and surface defect distribution data, and associate them with the spatial coordinates of the workpiece on the hanger to form a multi-source coating dataset for the current workpiece.

[0014] S5. Based on the historical workpiece coating data sequence of the fixture and the multi-source coating dataset of the current workpiece, determine the need for fixture layout adjustment.

[0015] S6. If it is determined that adjustment is needed, then start the dynamic reconstruction and optimization of the hanger spacing and then control the spacing.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention realizes dynamic monitoring and feedback control of bubble behavior during electroplating by collecting bubble distribution images and electroplating liquid flow field data on the workpiece surface in real time and performing multimodal fusion analysis. This overcomes the shortcomings of traditional feedforward control methods in responding to real-time changes in working conditions and significantly improves the control accuracy and reliability of the electroplating process.

[0017] (2) Based on three-dimensional contour data, this invention identifies high-risk electroplating areas and generates matching fixture posture configuration schemes. Combined with simulation verification methods, it ensures that workpieces with different geometric features can obtain stable electrolyte coverage and current distribution, effectively improving the problem of coating uniformity among multiple workpieces on the same fixture.

[0018] (3) By introducing a bubble morphology feature extraction and sub-region division mechanism, this invention achieves accurate quantitative evaluation of bubble size, distribution and aggregation tendency. Then, by adjusting the rotation speed and swing amplitude of the workpiece in a coordinated manner, a targeted bubble elimination strategy is implemented, overcoming the limitations of poor adaptability and crude effect of mechanical bubble removal methods.

[0019] (4) By comprehensively utilizing multi-source information fusion technologies such as flow field simulation, image processing, and machine learning, this invention constructs a multi-dimensional collaborative optimization system covering bubble behavior, flow field state, and coating quality, realizing the leap from single-parameter control to multi-parameter coupled optimization in the electroplating process, and comprehensively improving coating consistency and defect suppression capabilities.

[0020] (5) By establishing a correlation analysis mechanism between historical plating data and real-time production data, this invention enables intelligent judgment of the fixture layout adjustment requirements and dynamic optimization of the spacing scheme, forming a complete closed-loop quality control, thereby significantly enhancing the adaptability and continuous optimization capability of the electroplating process. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. 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 diagram of the overall implementation process of electroplating on the titanium alloy surface of the present invention.

[0023] Figure 2 This is a schematic diagram of the process for generating the attitude configuration scheme of the mounting device according to the present invention.

[0024] Figure 3 This is a simplified schematic diagram of the overall multimodal data fusion process of the present invention.

[0025] Figure 4 This is a schematic diagram of the linkage adjustment process of rotation speed and swing amplitude of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0027] Please see Figure 1As shown, the present invention provides a method for controlling electroplating on titanium alloy surfaces. The method includes: S1, dynamically generating a fixture posture configuration scheme that matches the geometric features of the workpiece based on the collected three-dimensional contour data of the titanium alloy workpiece.

[0028] S2. Immerse the workpiece in the electroplating solution according to the fixture posture configuration scheme, and simultaneously collect bubble distribution images and electroplating solution flow field data on the workpiece surface during the electroplating process. Generate a bubble-flow field coupling feature set through multimodal data fusion.

[0029] S3. Based on the bubble-flow field coupling feature set, adjust the rotation speed and oscillation amplitude of the workpiece in real time.

[0030] S4. When electroplating is completed, simultaneously collect coating thickness distribution data, edge integrity image data, and surface defect distribution data, and associate them with the spatial coordinates of the workpiece on the hanger to form a multi-source coating dataset for the current workpiece.

[0031] S5. Based on the historical workpiece coating data sequence of the fixture and the multi-source coating dataset of the current workpiece, determine the need for fixture layout adjustment.

[0032] S6. If it is determined that adjustment is needed, then start the dynamic reconstruction and optimization of the hanger spacing and then control the spacing.

[0033] Understandably, the acquisition of three-dimensional contour data in step S1 is achieved by using a combination of a laser three-dimensional scanner and a structured light imaging system to obtain data including surface curvature distribution, edge contour features, and geometric dimensional parameters.

[0034] Specifically, please refer to Figure 2 As shown, the specific generation process of the fixture posture configuration scheme in step S1 is as follows: S11, obtain the coordinate information and geometric feature parameters of each point on the workpiece surface, the geometric feature parameters including curvature, surface area and distribution of concave and convex regions.

[0035] S12. Construct a three-dimensional model based on the obtained workpiece surface coordinate information and geometric feature parameters to identify high-risk electroplating areas.

[0036] Understandably, high-risk electroplating areas include convex or concave areas with an absolute curvature greater than a preset threshold, as well as complex structural areas with a surface area ratio exceeding a preset proportion.

[0037] For example, the preset absolute value threshold of curvature is determined based on the material properties of the titanium alloy workpiece and the requirements of the electroplating process, and typically ranges from [value missing]. The preset surface area ratio threshold is 20%-30%. Complex structural regions of titanium alloy workpieces can be defined based on the workpiece's geometric complexity, number of features, and connection methods. If a region contains three or more different types of geometric features, such as planes, cylindrical surfaces, and conical surfaces, and the transition radius between these features is less than a certain value, such as 0.5mm, and at least two features within this region have dimensional tolerances less than a specific range, such as ±0.05mm, then this region can be defined as a complex structural region. Alternatively, when a region contains intersecting and nested structures, and the number of intersecting or nested layers reaches three or more, it can also be considered a complex structural region.

[0038] S13. Generate an initial fixture posture configuration scheme based on the three-dimensional model, the high-risk electroplating area, and the preset fixture constraints.

[0039] Understandably, the constraints on the hanger include the number of contact points between the hanger and the workpiece, the load-bearing range of the contact points, and the dimensional limitations of the hanger itself. Among these constraints, the number of contact points between the hanger and the workpiece is determined based on the workpiece weight. For example, there should be no fewer than two contact points per 1 kg of workpiece, and these contact points should be evenly distributed in non-high-risk areas. The load-bearing range of the contact points is 5-20 N to prevent workpiece deformation due to excessive load.

[0040] S14. Verify the electrolyte coverage and current density uniformity of the initial scheme in high-risk areas through fluid dynamics simulation.

[0041] Understandably, in the simulation verification step, computational fluid dynamics is used to simulate the electrolyte flow state, and finite element analysis is used to simulate the current distribution. The preset coverage standard is not less than 95%, and the current density uniformity standard is that the ratio of the maximum current density to the minimum current density does not exceed 1.5. All the methods mentioned above use existing technical means, and will not be described in detail here.

[0042] Electrolyte coverage is quantified by calculating the percentage of the effective area in contact with the electrolyte in high-risk areas relative to the total area of ​​that area. Current density uniformity is quantified by the coefficient of variation, which is the ratio of the standard deviation to the mean of the current density at each measurement point in the high-risk area.

[0043] S15. When the verification result meets the preset threshold, the initial fixture posture configuration scheme is output as the final scheme. Otherwise, the contact parameters between the fixture and the workpiece are adjusted and the verification is repeated until a qualified fixture posture configuration scheme is obtained. The contact parameters include the contact point position and angle.

[0044] It should be noted that in the step of adjusting the position and angle of the contact point between the fixture and the workpiece, the contact point near the high-risk area should be adjusted first, so that the distance between the contact point and the high-risk area is not less than 5mm, and the angle of the fixture posture adjustment does not exceed 10° each time, so as to ensure the stability of the adjustment process.

[0045] It should be added that, addressing the issue of uneven coating in high-risk areas due to the complex geometry of titanium alloy workpieces, a combination of 3D modeling and simulation verification is used to pre-assess and optimize the impact of the mounting fixture posture on electrolyte flow and current distribution in a virtual environment. This ensures that the generated mounting fixture configuration significantly improves electrolyte coverage and current density uniformity in high-risk areas, reducing coating defects caused by workpiece geometry at the source, and enhancing the stability of the electroplating process and the consistency of coating quality.

[0046] Specifically, please refer to Figure 3 As shown, the specific process of multimodal data fusion in step S2 includes: S21, acquiring bubble distribution images by using multi-angle industrial cameras set around the electroplating tank, and preprocessing and binarizing the images.

[0047] Understandably, the bubble distribution image undergoes preprocessing, including but not limited to noise reduction, contrast enhancement, and brightness adjustment, to improve the distinction between the bubbles and the background. The preprocessing and binarization conversion utilize existing, mature technologies, which will not be described in detail here.

[0048] S22. Based on the prior marker constraints of bubble morphology, the watershed algorithm is used to segment the adhered bubbles and extract the morphological feature parameters of each independent bubble. The morphological feature parameters include equivalent diameter, morphology factor and three-dimensional spatial coordinates.

[0049] It should be noted that the adhesive state is a state in which the edges of two or more air bubbles overlap or contact each other.

[0050] Understandably, the marker point constraint based on bubble morphology priors specifically includes: obtaining potential bubble regions through distance transformation, generating marker points by combining bubble roundness priors and size priors, and constraining the starting position of the watershed transformation.

[0051] Understandably, the watershed algorithm specifically includes: initially extracting bubble regions through Otsu thresholding, performing distance transformation on the binary image, and obtaining a distance mapping map.

[0052] Based on bubble morphology priors, roundness thresholds and size range thresholds are set, and regions that meet the conditions are selected as marker points.

[0053] By inputting the marker points into the watershed algorithm as constraints, accurate bubble segmentation is achieved.

[0054] The Otsu threshold segmentation mentioned above is an existing segmentation method, and the segmentation process will not be described in detail again.

[0055] It should be added that the threshold settings are based on a large amount of experimental data statistics and process requirements. By analyzing the differences in morphological characteristics between normal and abnormal bubbles, the roundness threshold is determined to distinguish between near-circular bubbles and deformed bubbles, and the size range threshold is used to filter out invalid bubble signals that are too large or too small.

[0056] Understandably, a two-dimensional image of the bubbles in a sub-region can be captured by using a high-precision optical imaging system, such as a laser confocal microscope. The outline of each bubble can be extracted using image segmentation algorithms such as threshold segmentation or edge detection. The area S of each bubble outline can be calculated, and the equivalent diameter can be calculated according to the equivalent diameter formula.

[0057] Understandably, the shape factor is quantified using a circularity measure, and the circularity formula is: , Indicates roundness. The area representing the bubble's outline. The circumference of the bubble outline is denoted by 0. The value of roundness ranges from 0 to 1. When it is 1, it is a standard circle. The smaller the roundness, the more irregular the bubble shape.

[0058] Understandably, 3D scanning techniques such as structured light 3D scanners can be used to obtain the 3D coordinates of each bubble within a sub-region.

[0059] S23. Divide the electroplating solution container into three-dimensional meshes according to the mesh size that matches the minimum diameter of the bubbles, and obtain multiple unit sub-regions.

[0060] Understandably, the grid size setting needs to ensure that a single grid can fully accommodate the smallest bubble while avoiding insufficient feature accuracy due to an excessively large grid. For example, for a 100L container, a cubic grid with a side length of 5-10mm can be divided.

[0061] It should be noted that the grid size is set according to the volume of the electroplating solution in the container and the required accuracy of bubble detection, and the volume of a single grid cell should not exceed 0.5% of the total volume of the container.

[0062] For example, if the surface area of ​​the workpiece is small, such as a micro-precision part, the area is less than Furthermore, the average bubble diameter is 0.1-1mm. It is generally recommended to set the grid size to 5-10mm to ensure that each sub-region contains at least 3-5 bubbles, avoiding excessively high data randomness due to excessively small areas. If the workpiece surface area is large, such as a large sheet metal, the area should be larger than... Furthermore, the average bubble diameter is 1-5mm, and the grid size can be set to 20-50mm, which reduces the amount of subsequent calculations while ensuring data representativeness.

[0063] S24. Spatial matching is performed between the binary image of each sub-region and the 3D mesh, and the bubble volume ratio of each sub-region is calculated to obtain the bubble coverage rate.

[0064] S25. The flow velocity vector field of the electroplating solution is obtained by using particle image velocimetry technology, and the flow velocity and direction of each bubble are measured as the flow field parameters of each bubble.

[0065] S26. Based on the extracted bubble morphological parameters and flow field parameters, calculate the real-time R value of each bubble, and calculate the aggregation tendency index of each sub-region accordingly. The R value is the ratio of drag to buoyancy.

[0066] S27. After spatiotemporally aligning the bubble coverage, aggregation tendency index and flow field parameters, feature-level fusion is performed using a feature fusion algorithm to generate a bubble-flow field coupled feature set.

[0067] It should be noted that data from a single type of sensor cannot fully reflect the complex bubble-flow field coupling state during electroplating. By fusing visual images and flow field measurement data, a multi-dimensional feature characterization system can be constructed to establish a quantitative correlation between bubble morphology distribution, motion characteristics, and flow field parameters. This provides a real-time and comprehensive data foundation for subsequent precise control of workpiece motion parameters, thereby enabling proactive intervention and elimination of bubble behavior and significantly improving coating quality.

[0068] Furthermore, the specific calculation process of the aggregation tendency index in step S26 includes: 1) statistically analyzing the distribution characteristics of the morphological parameters of each bubble in the sub-region, normalizing the distribution characteristics, and obtaining the bubble dynamic instability index by dynamically weighting and fusing them through the first fusion function based on the titanium alloy-electrolyte interface characteristics.

[0069] Specifically, the statistical process for the distribution characteristics includes: calculating the coefficient of variation of the equivalent diameter of bubbles in a sub-region based on the equivalent diameter of each bubble, and calculating the variance and skewness of the R values ​​of all bubbles in the sub-region.

[0070] The morphology factor values ​​are divided into several discrete intervals and their frequencies are statistically analyzed. Based on the frequencies, the distribution entropy value of the bubble morphology factor within the sub-region is calculated.

[0071] The dynamic weighted fusion strategy of the first fusion function specifically includes: B1, determining the composition, thickness and porosity of the surface oxide film of titanium alloy samples under different pretreatment processes through experiments.

[0072] Understandably, X-ray photoelectron spectroscopy or Auger electron spectroscopy are commonly used for composition determination to accurately identify the types and contents of relevant oxides in the oxide film, while thickness determination is performed using an ellipsometry or scanning electron microscope cross-section observation. Porosity determination is achieved through image analysis or gas adsorption methods. Both image analysis and gas adsorption methods are existing technologies and will not be described further.

[0073] It should be noted that the characteristics of the oxide film on the surface of titanium alloy directly affect the adhesion and uniformity of the subsequent electroplating layer. This step provides experimental basis for screening the optimal pretreatment process and controlling the quality of the oxide film.

[0074] B2. Under the same electroplating process, the initial nucleation rate of bubbles, average residence time, and final coating porosity of the sample surface are collected simultaneously.

[0075] Understandably, the initial bubble nucleation rate is determined by photographing the sample surface with a high-speed industrial camera and counting the number of newly generated bubbles per unit time. The average residence time is recorded using image tracking technology, taking the time from nucleation to detachment of a single bubble, and the average value of multiple bubbles is calculated. The final coating porosity is determined by observing the coating cross-section using a metallographic microscope or indirectly reflecting the porosity by measuring the corrosion current through a salt water immersion test.

[0076] Understandably, the initial nucleation rate of bubbles reflects how fast bubbles are generated, the average residence time reflects the stability of bubble adsorption on the surface, and the porosity of the coating directly reflects the core indicator of the anti-corrosion performance of electroplated products.

[0077] It should be added that if a large number of bubbles remain during the electroplating process, it can lead to defects such as pinholes and blistering in the coating. This step, by quantifying the relationship between bubble behavior and coating quality, provides direction for subsequent optimization of bubble control.

[0078] B3. Through causal correlation analysis based on partial least squares path model, the set of oxide film characteristic parameters that have a dominant influence on bubble behavior is identified.

[0079] It should be added that the partial least squares path model is an existing model. First, the oxide film characteristic parameters are set as exogenous latent variables and the bubble behavior parameters are set as endogenous latent variables. The path coefficients of each oxide film parameter on bubble behavior are calculated through the model. The larger the absolute value of the path coefficient, the stronger the influence. The set of dominant parameters is determined by combining cross-validation.

[0080] It should be noted that this step enables a focus on core parameters rather than multiple parameters. Subsequently, only a few dominant parameters need to be controlled specifically to efficiently regulate bubble behavior and reduce the complexity of process control.

[0081] B4. Real-time acquisition of microscopic temporal images of the current workpiece surface, and extraction of oxide film feature parameter estimates using a trained deep learning model.

[0082] Understandably, deep learning models typically employ a combination of image segmentation and feature regression architecture. For example, a semantic segmentation model based on U-Net first segments the oxide film region from a microscopic temporal image, and then extracts features such as texture and grayscale within the region using convolutional neural networks such as ResNet, outputting estimated values ​​corresponding to the dominant parameters, such as porosity percentage and thickness value.

[0083] It should be added that this step enables real-time perception of oxide film characteristics, which is a prerequisite for subsequent dynamic adjustment of process parameters and early intervention of bubble defects, breaking the traditional passive mode of post-event inspection and batch scrapping.

[0084] B5. Calculate the expected detachment time of the bubble based on the bubble force balance mechanism, and obtain the adsorption stability deviation index by combining the actual average residence time.

[0085] Understandably, the actual average residence time of bubbles can be statistically analyzed by analyzing time-series images, and the standardized residual between the expected detachment time and the actual average residence time can be calculated as an adsorption stability deviation index.

[0086] It should be added that this step transforms the bubble behavior from a qualitative description to a quantitative description, providing a clear numerical basis for subsequent judgment on whether process adjustments are needed.

[0087] B6. The final bubble dynamic instability index is obtained by weighted summing of the adsorption stability deviation index and the normalized value of the distribution characteristics.

[0088] Understandably, when performing weighted summation, the weights are determined through the analytic hierarchy process or actual process experience. For example, since adsorption stability has a more direct impact on coating defects, the weight of the adsorption stability deviation index is 0.6, and the weight of the distribution characteristic normalized value is 0.4.

[0089] 2) By analyzing the fluctuation characteristics of bubble coverage within a sub-region through time series analysis, the bubble coverage instability index is obtained.

[0090] Understandably, the sliding window method and time-domain analysis method can be used to extract fluctuation characteristics. For example, the window size can be set to 10 time points, and the coverage variance, the maximum value of the coverage change rate and the fluctuation period within the sliding window can be calculated. The coverage variance reflects the intensity of short-term fluctuations, the maximum value of the coverage change rate reflects the degree of coverage mutation, and the fluctuation period is extracted by Fourier transform to reflect the periodic fluctuation law of the coverage.

[0091] After normalizing the above fluctuation characteristics, the bubble coverage instability index is obtained by weighted summation. The weights can be set empirically, for example, because short-term fluctuations have a more direct impact on instantaneous aggregation. The short-term fluctuation variance is weighted at 0.5, the maximum value of the coverage rate change is weighted at 0.3, and the fluctuation period is weighted at 0.2, because short-term fluctuations have a more direct impact on instantaneous aggregation.

[0092] It should be added that this step compensates for the shortcomings of static morphological analysis by using the fluctuation characteristics in the time dimension. It can capture the dynamic change risk of bubble coverage. For example, if the coverage of a certain sub-region suddenly increases from 10% to 40% within 1 second, its coverage instability index is close to 1, indicating that bubbles may rapidly accumulate in this area and early intervention is required.

[0093] 3) Generate an adsorption energy density field based on the interface characteristics of titanium alloy and electrolyte.

[0094] Understandably, the interface characteristics of titanium alloys and electrolytes include the surface energy of the titanium alloy, the composition of the oxide film, the micro-roughness, and the composition of the electrolyte.

[0095] Specifically, based on the oxide film composition, micro-roughness, and electrolyte composition of the titanium alloy workpiece surface, an adsorption energy density field representing the distribution of adsorption intensity of bubbles at different surface locations is generated through molecular dynamics simulation or experimental calibration.

[0096] It should be added that the adsorption energy density field is a key bridge connecting the physical properties of the interface and the behavior of bubbles. For example, the roughness of the scratches on the surface of titanium alloy is high, and the adsorption energy density is significantly higher than that of other areas. Bubbles tend to accumulate in this area, and the instability index weight of this area should be given special attention when correcting the index.

[0097] 4) The bubble dynamic instability index and bubble coverage instability index are corrected by the adsorption energy density field. The corrected index and the adsorption energy density field are then input into a pre-trained physical information neural network, and the final aggregation tendency index of the sub-region is output.

[0098] Understandably, the physical information neural network adopts a combination architecture of fully connected layers and physical constraint layers. The input layer is the corrected double exponent and adsorption energy density field data. The hidden layer extracts features through the ReLU activation function. The physical constraint layer embeds the core physical laws of bubble aggregation, such as the higher the adsorption energy, the more positively correlated the bubble aggregation probability with the adsorption time. The output layer is the aggregation tendency index.

[0099] It should be added that this step incorporates the physical essence through interface modification, and integrates data-driven approaches with physical laws through PINN, solving the problem of traditional pure data models being detached from actual processes or pure physical models lacking accuracy. The final output index can directly guide the dynamic control of the electroplating production line. For example, adjusting the plating solution flow rate and optimizing the rack posture in high-risk sub-regions can reduce plating defects caused by bubble aggregation.

[0100] It should be added that, by integrating multi-dimensional information such as bubble morphology and distribution characteristics, temporal fluctuation patterns, and interfacial adsorption energy fields, and by introducing a physical information neural network for coupled analysis, a quantitative assessment model for bubble aggregation tendency that conforms to the physical laws of electrochemical interfaces was established. This model can accurately predict and intervene in bubble aggregation behavior, thereby effectively suppressing coating defects caused by bubble merging and adhesion.

[0101] Furthermore, the specific execution process of the correction includes: multiplying the bubble dynamic instability index of each sub-region by the average adsorption energy density of the corresponding region to obtain the bubble dynamic instability index after interface correction.

[0102] The bubble dynamic instability index of each sub-region is multiplied by the adsorption energy gradient of the corresponding region to obtain the bubble coverage instability index after interface correction.

[0103] Understandably, the bubble dynamic instability index is used to assess the unevenness of forces acting on bubbles and the instability of their motion trends. The bubble coverage instability index is used to assess the volatility and severity of bubble adhesion distribution on surfaces.

[0104] Specifically, please refer to Figure 4 As shown, the specific execution process of the linkage adjustment in step S3 is as follows: S31, calculate the overall bubble coverage rate, bubble coverage uniformity and bubble coverage concentration based on the bubble coverage rate of each sub-region, and form a bubble coverage feature vector.

[0105] Understandably, the overall bubble coverage rate is calculated using a weighted average method, which involves summing and quantifying the product of the bubble coverage rate of all sub-regions and the ratio of the area of ​​the corresponding sub-region to the total area of ​​the workpiece. The bubble coverage uniformity is quantified by the coefficient of variation method to determine the dispersion of the coverage rate of each sub-region. The bubble coverage concentration is determined by spatial clustering analysis, which uses the DBSCAN clustering algorithm to identify sub-regions with coverage rates higher than a threshold, such as 30%, that form continuous clusters. The concentration is calculated by multiplying the ratio of the total area of ​​the sub-regions within the cluster to the total area of ​​the workpiece by the ratio of the average coverage rate within the cluster to the overall bubble coverage rate. A larger value indicates that the bubbles are more concentrated in a local area.

[0106] It should be added that a single overall bubble coverage rate cannot fully reflect the spatial distribution characteristics of bubbles on the workpiece surface. It is necessary to assess the uniformity of distribution through coverage uniformity and identify local clustering areas through coverage concentration. This allows for a comprehensive characterization of the multi-dimensional feature vector of the bubble spatial distribution morphology, providing precise input for subsequent motion parameter adjustments and enabling differentiated and precise control of different bubble distribution patterns, such as uniform distribution and local clustering.

[0107] S32. Extract the flow velocity and flow direction parameters of each sub-region to form a flow field feature vector, and form a bubble aggregation feature vector based on the aggregation tendency index of each sub-region.

[0108] Understandably, flow field parameters directly affect the migration and detachment of bubbles. Low flow velocity can easily lead to bubble retention, and chaotic flow direction can easily cause local bubble accumulation. The aggregation tendency index can provide early warning of potential aggregation risks. The combination of the two extends process evaluation from static status quo to dynamic trend, which can significantly improve the foresight of adjustment decisions.

[0109] S33. Based on the workpiece material thickness and electroplating process parameters, query the corresponding reference motion parameter group and corresponding feature tolerance range from the preset process knowledge base.

[0110] Understandably, the data in the process knowledge base can be comprehensively constructed through a large number of experiments and combined with empirical data.

[0111] It should be added that allowable fluctuation ranges are set for the three types of feature vectors: bubble coverage, flow field, and aggregation tendency. For example, for a titanium alloy thickness of 3mm and a current density of 2A / When the bubble coverage feature vector is within the range of: overall coverage 15%-30%, uniformity ≥0.6, and concentration ≤0.5.

[0112] S34. When all eigenvectors are within the corresponding tolerance range, the reference motion parameter set is used as the control output.

[0113] S35. Otherwise, start the motion parameter adjustment mechanism and output the adjustment motion parameter category and adjustment ratio. After correcting the baseline motion parameters based on the motion parameter category and adjustment ratio, output the corrected parameters.

[0114] Furthermore, the specific execution process of the motion parameter adjustment mechanism is as follows: F1. Determine the category of motion parameter to be adjusted based on the situation where the feature vector exceeds the tolerance range. The category includes rotational speed and swing amplitude.

[0115] It should be added that if any of the following conditions are met: the bubble aggregation tendency index exceeds the upper limit of the tolerance range, the overall bubble coverage exceeds the upper limit of the tolerance range, or the flow field velocity is lower than the lower limit of the tolerance range, then the adjustment parameter is determined to be the rotation speed.

[0116] If any of the bubble coverage concentration, bubble coverage uniformity, or overall bubble coverage exceeds the upper limit of the corresponding tolerance range, then the adjustment parameter is determined to be the swing amplitude.

[0117] F2. If it is determined that the rotation speed should be adjusted, the rotation speed adjustment ratio is calculated and output through the Sigmoid function based on the bubble aggregation tendency index, the degree of excess of the overall bubble coverage, and the degree of deficiency of the flow field velocity. The degree is quantified by the relative deviation value.

[0118] F3. If the adjustment swing amplitude is determined, the swing amplitude adjustment ratio is calculated and output through the Sigmoid function based on the bubble coverage concentration, bubble coverage uniformity, and the degree of excess of the overall bubble coverage rate.

[0119] Understandably, rotational speed primarily affects fluid shear force and bubble detachment, making it suitable for addressing macroscopic bubble problems caused by high aggregation tendency, excessive overall coverage, or insufficient flow field velocity. Oscillation amplitude, on the other hand, mainly alters the local flow regime and contact frequency on the workpiece surface, making it suitable for improving bubble distribution uniformity and eliminating localized concentrated areas. Its function is to establish a mapping relationship between multi-parameter abnormal states and optimal control methods, achieving smooth proportional adjustments through the Sigmoid function. This avoids interference from parameter abrupt changes in the electroplating process and enables precise and stable dynamic control for different bubble defect modes.

[0120] Specifically, in step S4, an X-ray fluorescence thickness gauge can be used to collect the coating thickness distribution data on the workpiece surface, with a sampling point density of no less than 5 points per square centimeter. The thickness value of each sampling point is obtained, and the thickness variation coefficient, i.e., the ratio of the standard deviation to the mean of the thickness data, is calculated. Images of the workpiece edge area are acquired using a high-resolution industrial camera with a resolution of no less than 2048×2048. The Canny operator is used for edge detection to identify edge integrity, and the edge defect length ratio is calculated, i.e., the ratio of the total length of defects to the total length of the workpiece edge. Surface defect data is collected using a laser scanning microscope with a scanning accuracy of no more than 1 micrometer. Defect types such as pinholes, pitting, and peeling are identified, and the defect density is calculated. The defect density is the ratio of the total number of defects in the defect density detection area to the detection area.

[0121] Specifically, the determination of the need for adjusting the fixture layout in step S5 includes: S51, calculating the thickness variation coefficient, defect density, and edge defect length ratio based on thickness distribution data, surface defect distribution data, and edge integrity image data, respectively, as the characteristics of each defect.

[0122] S52. Traverse the historical batch data. If the thickness variation coefficient or defect density of a certain batch of workpieces exceeds a preset multiple of the median of the historical data, then remove the batch data and record the remaining batches after removal as valid historical batches.

[0123] Understandably, the preset multiplier is usually set to 1.5 times, but it can be adjusted according to production stability.

[0124] S53. Based on the statistical benchmark values ​​of each defect feature of all valid historical batches of workpieces, calculate the deviation rate of each defect feature of the current workpiece from its benchmark value, and output the comprehensive defect degree after linear weighting.

[0125] Understandably, the weights can be determined by process expert scoring, such as a thickness variation coefficient weight of 0.4, a defect density weight of 0.4, and an edge defect ratio weight of 0.2.

[0126] In one specific embodiment, the specific statistical process of the benchmark values ​​of each defect feature includes: filtering outliers of each defect feature for each group of data using the quartile method, calculating the arithmetic mean of each defect feature for each group of data after filtering, and obtaining the thickness variation coefficient benchmark, the edge defect length ratio benchmark, and the defect density benchmark.

[0127] If there are fewer than three valid batches of a certain type of feature data in the historical data sequence, the industry standard value of the same type of workpiece under the same process conditions will be used as the initial benchmark. After enough historical data is accumulated, the benchmark value will be updated through the above steps.

[0128] It is important to note that after every 3-5 batches of qualified workpieces have their coating data updated, the outlier filtering and average value calculation steps should be repeated. The thickness variation coefficient benchmark, edge defect length ratio benchmark, and defect density benchmark should be iteratively updated to ensure that the benchmark values ​​always reflect the stable state of the current process. When process parameters such as electroplating current or electrolyte concentration change, the benchmark values ​​are reset. Starting with the coating data of the first qualified workpiece after the change, historical data is re-accumulated and a new benchmark is calculated.

[0129] S54. Select historical workpieces whose production time falls within a preset time window as a reference set.

[0130] Understandably, the preset time window setting can determine the length of the time window according to the production rhythm, usually 7 days or 20 production batches, to ensure that the sample size of the reference set is sufficient, while avoiding excessively long time that would lead to large differences in production conditions.

[0131] S55. When the overall defect rate of the current workpiece exceeds the preset threshold, and the proportion of workpieces with continuously exceeding the defect rate in the reference set reaches the preset threshold, it is determined that the fixture layout needs to be adjusted; otherwise, it is determined that no adjustment is needed.

[0132] Understandably, the preset threshold for the overall defect rate can be set based on historical qualified batch data, usually 0.3. That is, when the overall defect rate is greater than 0.3, the current workpiece quality deviates significantly from the normal level. The preset threshold for the proportion of consecutive exceedances can be set in conjunction with production stability, usually 30%. That is, in the reference set, when the proportion of batches with three or more consecutive workpieces exceeding the defect rate is greater than or equal to 30%, it is defined as consecutive exceedances.

[0133] It should be added that the dual threshold judgment avoids misadjustment, reduces unnecessary fixture adjustments, lowers downtime and labor costs, and also prevents missed adjustments. It can detect fixture layout problems in a timely manner and avoid scrap losses caused by continuous batch defects.

[0134] Specifically, the specific steps for initiating the dynamic reconstruction optimization of the spacing between the hangers in step S6 are as follows: S61, assign a unique identifier to each hanger in the electroplating tank and record its initial horizontal position.

[0135] S62. Calculate the horizontal spacing between adjacent hangers and evaluate the correlation between each spacing value and the corresponding workpiece coating quality index using the Pearson correlation coefficient, and identify hanger pairs whose spacing values ​​are significantly correlated with the quality index.

[0136] Understandably, the correlation coefficient ranges from [-1, 1]. The larger the absolute value, the stronger the linear association. Significant correlation is determined by conducting a significance test. A significance p-value of less than 0.05 is set as significant correlation. For example, if the correlation coefficient between the spacing of the M01-M02 fixture pair and the thickness uniformity of the corresponding workpiece is -0.72 and the significance p-value is 0.02, it means that the smaller the spacing, the worse the thickness uniformity, and the association is significant.

[0137] S63. Based on the preset adjustment constraints of the horizontal spacing of the hangers, design multiple horizontal spacing adjustment schemes, and select candidate schemes that meet the process requirements from the adjustment schemes through flow field and current density distribution simulation analysis.

[0138] It should be added that, before designing the scheme, it is also necessary to determine the concentrated area of ​​quality defects based on the plating quality data of the corresponding workpieces of the fixture, and to determine whether the defects are caused by unreasonable spacing of the fixtures.

[0139] For example, the coating quality data of the suspended workpieces, such as thickness inspection reports and surface defect images, are analyzed for significantly relevant fixtures. The workpiece surface is divided into 10×10 grid units using a grid division method. The number of defects in each unit, such as pinholes and excessive thickness deviations, is counted. The defect distribution is visualized through a heat map to determine the grid areas where defects are concentrated.

[0140] It should be added that, through the correlation analysis of the fixture with the defect area, non-spacing factors, such as fixture contact corrosion and workpiece morphology, were first ruled out. This ensured that subsequent spacing optimization schemes were designed only for quality problems truly caused by flow field or electric field interference. This avoided ineffective spacing adjustments and significantly improved optimization efficiency and the relevance of the solutions.

[0141] It should also be added that the defect distribution of the workpieces can be measured by selecting hangers with different spacings under the same batch, current, temperature and plating solution. If the defect density in the defect concentration area increases by more than 30% when the spacing is reduced, such as from 120mm to 100mm, and the defect density decreases by more than 20% when the spacing is adjusted to 140mm, it indicates that the defects are strongly correlated with the spacing.

[0142] If the defect concentration area coincides with the flow field / current field area affected by the spacing between the fixtures, such as the dead angle of the flow field between the two fixtures corresponding to the defect area of ​​the workpiece, and the defect still changes with the spacing after excluding other factors, then the defect is determined to be caused by the unreasonable spacing between the fixtures.

[0143] Understandably, the adjustment constraints of the horizontal spacing of the hangers include spatial constraints and process constraints. The spatial constraints are: the minimum distance from the inner wall of the electroplating tank to the edge hangers is greater than or equal to 50mm, and the minimum spacing between adjacent hangers is greater than or equal to 80mm, to avoid hanger collisions or excessive flow field disturbance. The process constraints are: the spacing adjustment range is less than or equal to ±20mm at a time, to avoid large adjustments causing sudden changes in the current field and affecting the workpiece quality of other hangers.

[0144] Specifically, simulation analysis can simulate the electrolyte flow rate and update frequency between adjacent workpieces under different adjustment schemes through computational fluid dynamics, and combine it with finite element analysis to simulate the current density distribution, and screen out the schemes that meet the requirements of flow field uniformity greater than or equal to a preset threshold and current density distribution deviation less than or equal to a preset deviation as candidate schemes.

[0145] Understandably, the flow field uniformity threshold is determined by analyzing the coefficient of variation of the flow velocity distribution in the flow field simulation data corresponding to historical high-quality coatings, ensuring sufficient electrolyte renewal and the absence of dead zones. The current density distribution deviation threshold is set based on the correlation analysis between finite element simulation results and measured coating thickness, limited to the maximum relative deviation of current density corresponding to the allowable thickness tolerance. Both types of thresholds must meet process specification requirements and their effectiveness must be verified through actual production.

[0146] S64. Based on the comprehensive defect rate of historical workpieces, score the candidate solutions and select the solution with the highest score as the target adjustment solution.

[0147] For example, the overall defect rate of fixture pairs with similar spacing values ​​to candidate solutions is selected. For instance, if the candidate solution spacing is 120mm, then historical spacing values ​​within the range of 115-125mm are matched. The average overall defect rate of each candidate solution group is calculated. The complement of the average overall defect rate is multiplied by a set rated score, such as 100, to obtain the overall score of each candidate solution group. The solution with the highest overall score and no individual indicator score below 60 is selected as the target adjustment solution. The complement is the value obtained by subtracting the average overall defect rate from 1.

[0148] S65. Adjust the position of the fixture according to the target adjustment plan, and verify and correct the spacing value through iterative optimization after adjustment until the coating quality of multiple consecutive workpieces in the same batch is qualified.

[0149] Understandably, after the first workpiece is electroplated according to the target adjustment scheme, if its comprehensive score does not reach the preset threshold, it will be iteratively adjusted within the preset margin of the target spacing according to the preset step size until the plating quality of multiple consecutive workpieces in the same batch is qualified.

[0150] It should be added that the preset allowance range is determined based on the adjustment accuracy of the fixture's mechanical structure and the tolerance of the electroplating process, and is usually set to ±5%-10% of the target spacing value to ensure that the adjustment process is carried out safely within the allowable fluctuation range of the process. The preset step size is set comprehensively according to the spacing adjustment accuracy requirements and production efficiency, and is usually 0.5-2mm / step, which ensures adjustment accuracy while avoiding excessive iteration that may affect the production cycle.

[0151] It should be noted that by using correlation analysis to identify key fixtures that are significantly related to quality defects, the blindness of global adjustments can be avoided. By using fluid and electric field simulation in a virtual environment to pre-screen candidate solutions that can significantly improve the uniformity of flow field and current distribution, the cost of trial and error can be greatly reduced. Combining historical data scoring and production iteration verification, an effective closed loop of optimization process is formed, thereby ensuring that the spacing adjustment is not only based on theoretical prediction but also has practical feasibility, and ultimately achieves a stable improvement in the coating quality of the entire batch of workpieces.

[0152] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for controlling electroplating on titanium alloy surfaces, characterized in that, The method includes: S1. Based on the collected three-dimensional contour data of the titanium alloy workpiece, dynamically generate a fixture posture configuration scheme that matches the geometric features of the workpiece. S2. Immerse the workpiece in the electroplating solution according to the fixture posture configuration scheme, and simultaneously collect bubble distribution images and electroplating solution flow field data on the workpiece surface during the electroplating process. Generate a bubble-flow field coupling feature set through multimodal data fusion. S3. Based on the bubble-flow field coupling feature set, adjust the rotation speed and oscillation amplitude of the workpiece in real time. S4. When electroplating is completed, simultaneously collect coating thickness distribution data, edge integrity image data and surface defect distribution data, and associate them with the spatial position coordinates of the workpiece on the hanger to form a multi-source coating dataset of the current workpiece. S5. Based on the historical workpiece coating data sequence of the fixture and the multi-source coating dataset of the current workpiece, determine the need for fixture layout adjustment. S6. If it is determined that adjustment is needed, then start the dynamic reconstruction and optimization of the hanger spacing and then control the spacing.

2. The method for controlling electroplating on titanium alloy surfaces as described in claim 1, characterized in that: The specific process for generating the mounting attitude configuration scheme is as follows: The coordinate information and geometric feature parameters of each point on the workpiece surface are obtained. The geometric feature parameters include curvature, surface area, and distribution of concave and convex regions. A three-dimensional model is constructed based on the obtained workpiece surface coordinate information and geometric feature parameters to identify high-risk electroplating areas; An initial fixture posture configuration scheme is generated based on the aforementioned 3D model, the high-risk electroplating area, and the preset fixture constraints. The electrolyte coverage and current density uniformity of the initial scheme in high-risk areas were verified through fluid dynamics simulation. When the verification result meets the preset threshold, the initial fixture posture configuration scheme is output as the final scheme; otherwise, the contact parameters between the fixture and the workpiece are adjusted and the verification is repeated until a qualified fixture posture configuration scheme is obtained.

3. The method for controlling electroplating on titanium alloy surfaces as described in claim 1, characterized in that: The specific process of multimodal data fusion includes: Bubble distribution images are acquired by multi-angle industrial cameras set around the electroplating tank, and the images are preprocessed and binarized. Based on the marker point constraint of bubble morphology prior, the watershed algorithm is used to segment the adhering bubbles and extract the morphological feature parameters of each independent bubble, including equivalent diameter, morphology factor and three-dimensional spatial coordinates. The electroplating solution container is divided into three-dimensional meshes according to the mesh size that matches the minimum diameter of the bubbles, resulting in multiple unit sub-regions; The binary images of each sub-region are spatially matched with the 3D mesh, and the bubble volume ratio of each sub-region is calculated to obtain the bubble coverage rate. Particle image velocimetry is used to obtain the flow velocity vector field of the electroplating solution, and the flow velocity and direction at the location of each bubble are measured as the flow field parameters of each bubble. Based on the extracted morphological and flow field parameters of the bubbles, the real-time R value of each bubble is calculated, and the aggregation tendency index of each sub-region is calculated accordingly. After spatiotemporal alignment of the bubble coverage, aggregation tendency index, and flow field parameters, a bubble-flow field coupling feature set is generated by feature-level fusion through a feature fusion algorithm.

4. The method for controlling electroplating on titanium alloy surfaces as described in claim 3, characterized in that: The specific calculation process for the clustering tendency index includes: The distribution characteristics of the morphological features of each bubble in the statistical sub-region are analyzed, the distribution characteristics are normalized, and the bubble dynamic instability index is obtained by dynamic weighted fusion based on the titanium alloy-electrolyte interface characteristics through a first fusion function. By analyzing the fluctuation characteristics of bubble coverage within a sub-region over time, a bubble coverage instability index is obtained. The adsorption energy density field is generated based on the interface characteristic parameters of titanium alloy and electrolyte. The bubble dynamic instability index and bubble coverage instability index are corrected by the adsorption energy density field, and the corrected index and adsorption energy density field are input into a pre-trained physical information neural network to output the final aggregation tendency index of the sub-region.

5. The method for controlling electroplating on titanium alloy surfaces as described in claim 4, characterized in that: The specific execution process of the correction includes: The bubble dynamic instability index after interface correction is obtained by multiplying the dynamic instability index of each sub-region by the average adsorption energy density of the corresponding region. The bubble dynamic instability index of each sub-region is multiplied by the adsorption energy gradient of the corresponding region to obtain the bubble coverage instability index after interface correction.

6. The method for controlling electroplating on a titanium alloy surface as described in claim 4, characterized in that: The dynamic weighted fusion strategy of the first fusion function specifically includes: The composition, thickness and porosity of the surface oxide film of titanium alloy samples under different pretreatment processes were determined by experiments. Under the same electroplating process, the initial nucleation rate of bubbles, average residence time, and final coating porosity on the sample surface were collected simultaneously. By using causal correlation analysis based on partial least squares path model, a set of oxide film characteristic parameters that have a dominant influence on bubble behavior was identified. Real-time acquisition of microscopic temporal images of the current workpiece surface, and extraction of oxide film feature parameter estimates using a trained deep learning model; The expected detachment time of the bubble is calculated based on the bubble force balance mechanism, and the adsorption stability deviation index is obtained by combining the actual average residence time. The final bubble dynamic instability index is obtained by weighted summing of the adsorption stability deviation index and the normalized value of the distribution characteristics.

7. The method for controlling electroplating on a titanium alloy surface as described in claim 3, characterized in that: The specific execution process of the coordinated adjustment is as follows: Based on the bubble coverage rate of each sub-region, the overall bubble coverage rate, bubble coverage uniformity, and bubble coverage concentration are calculated to form a bubble coverage feature vector; The flow velocity and flow direction parameters of each sub-region are extracted to form a flow field feature vector, and a bubble aggregation feature vector is formed based on the aggregation tendency index of each sub-region. Based on the workpiece material thickness and electroplating process parameters, the corresponding reference motion parameter set and corresponding feature tolerance range are queried from the preset process knowledge base; When all feature vectors are within the corresponding tolerance range, the reference motion parameter set is used as the control output; Otherwise, the motion parameter adjustment mechanism is activated and the adjustment type and adjustment ratio of the motion parameters are output. The baseline motion parameters are then corrected based on the motion parameter type and adjustment ratio before being output.

8. The method for controlling electroplating on titanium alloy surfaces as described in claim 7, characterized in that: The specific execution process of the motion parameter adjustment mechanism is as follows: The category of motion parameter adjustment is determined based on the case where the feature vector exceeds the tolerance range. The category includes rotational speed and oscillation amplitude. If the rotation speed adjustment is determined, the output rotation speed adjustment ratio is calculated using the Sigmoid function based on the bubble aggregation tendency index, the degree of excess of the overall bubble coverage, and the degree of deficiency of the flow field velocity. If the adjustment swing amplitude is determined, the swing amplitude adjustment ratio is calculated and output using the Sigmoid function based on the bubble coverage concentration, bubble coverage uniformity, and the extent to which the overall bubble coverage exceeds the limit.

9. The method for controlling electroplating on a titanium alloy surface as described in claim 1, characterized in that: The determination of the need to adjust the layout of the hanging fixtures includes: Based on thickness distribution data, surface defect distribution data, and edge integrity image data, the thickness variation coefficient, defect density, and edge defect length ratio are calculated as each defect feature. Traverse the historical batch data. If the thickness variation coefficient or defect density of a certain batch of workpieces exceeds a preset multiple of the median of the historical data, then remove the batch data and record the remaining batches after removal as valid historical batches. Based on the baseline values ​​of each defect feature of all valid historical batches of workpieces, the deviation rate of each defect feature of the current workpiece from its baseline value is calculated and then linearly weighted to output the comprehensive defect degree. Select historical workpieces whose production time falls within a preset time window as a reference set; If the overall defect rate of the current workpiece exceeds the preset threshold, and the proportion of workpieces with continuously exceeding the defect rate in the reference set reaches the preset threshold, then it is determined that the fixture layout needs to be adjusted; otherwise, it is determined that no adjustment is needed.

10. The method for controlling electroplating on a titanium alloy surface as described in claim 9, characterized in that: The specific steps for the dynamic reconstruction and optimization of the starting fixture spacing are as follows: Assign a unique identifier to each fixture in the electroplating tank and record its initial horizontal position; Calculate the horizontal spacing between adjacent hangers, and evaluate the correlation between each spacing value and the corresponding workpiece coating quality index using the Pearson correlation coefficient to identify hanger pairs whose spacing values ​​are significantly correlated with the quality index. Based on the preset adjustment constraints of the horizontal spacing of the hangers, multiple horizontal spacing adjustment schemes are designed. Through flow field and current density distribution simulation analysis, candidate schemes that meet the process requirements are selected from the adjustment schemes. Candidate solutions are scored based on the comprehensive defect rate of historical workpieces, and the solution with the highest score is selected as the target adjustment solution. Adjust the position of the fixture according to the target adjustment plan, and verify and correct the spacing value through iterative optimization after adjustment until the coating quality of multiple consecutive workpieces in the same batch is qualified.

Images