Coronary bionic adhesion array electrodeposition method

Abstract

The invention relates to a coronal bionic adhesion array electrodeposition method. Real-time current density, temperature and fluid flow rate sampling data in the electrodeposition process are collected in different regions of each key region of a coronal tail end bionic adhesion array electroforming core mold structure; a multi-field space coupling index is utilized to output parameter sets of local current, pulse waveform, micro-area stirring and the like with spatial resolution and dynamic response in real time, and closed-loop fine management of multi-channel power supply and micro-area stirring linkage control is realized.

Description

Technical Field

[0001] The present application relates to the field of "fine electrodeposition process control technology", and in particular to a method for electrodeposition of a crown-shaped bionic adhesion array. Background Art

[0002] Many organisms in nature possess remarkable adhesive abilities, such as beetles, geckos, and flies, whose adhesive functional units all feature terminal swellings. Inspired by this, the fabrication of biomimetic crown-shaped adhesive arrays offers a range of advantages, including high adhesion-to-pressure ratios, zero residue, and controllable adhesion and detachment. These arrays hold great promise in fields such as biomedicine, flexible grasping, climbing robotics, and space operations. In recent years, numerous research groups have dedicated themselves to practical industrial applications, and the efficient and controllable fabrication of biomimetic crown-shaped adhesive arrays is one of the most sought-after commercialization topics in biomimetics today.

[0003] Among numerous fabrication methods, molding technology is currently the mainstream approach for producing biomimetic adhesion arrays, due to its ability to efficiently and precisely replicate polymer surface microstructures. This method typically involves injecting a polymer into a mold. Depending on the mold cavity structure, the polymer is then cured and released, either directly or after a secondary dip, to produce the biomimetic adhesion array. In particular, nanoimprinting technology, a derivative of molding, has recently enabled the continuous fabrication of biomimetic adhesion arrays.

[0004] Currently, precise electrodeposition process control and multi-physics field real-time monitoring and feedback control technologies have garnered widespread attention in fields such as micro-nanofabrication, biomimetic materials, and electroforming of functional arrays. In particular, achieving precise, spatially distributed adaptive control of process parameters is a key technical challenge in mass-producing highly complex structures (such as biomimetic adhesion arrays of coronal tips) to ensure product consistency and scale-up. The mainstream existing approaches rely on static global parameter settings and feedback from a single physical field. For example, traditional electrodeposition processes typically employ single-point settings for full-bath uniform current, average temperature, and stirring rate, and rely on a small number of sensors for coarse monitoring and limited feedback of process parameters. Recent research has proposed approaches such as multi-channel power supply or zoned current regulation, but these approaches still primarily focus on empirical or zoned average parameter control. Some approaches attempt to control only a single physical field, such as current density or temperature, in real time, failing to account for the spatially non-uniform effects of coupled multi-physics fields (current, electrolyte fluid, temperature, etc.). Furthermore, for spatially distributed monitoring, the industry lacks a data-driven multi-parameter control system with high spatial resolution and precise acquisition and integration of in-situ data from multiple physical fields.

[0005] Typical representative technologies include: (1) a multi-channel power supply system based on limited partitions, which can only achieve block-based current adjustment at most, and is suitable for electrodeposition applications with a single structure or a large area, but has limited response capabilities to fine spatial differences of complex arrays at the micron / submicron level; (2) a small number of literature or patents propose the use of single-point temperature control, agitators or fluid field simulation to optimize the stirring position, which can only adjust a single parameter in the overall or local area, and cannot achieve coordinated real-time control of multiple points and multiple parameters in space; (3) some high-end equipment attempts to introduce measures such as online thickness measurement and total current negative feedback, but are mostly limited to static or simple process conditions, and lack effective response mechanisms for morphology mutations, process mutations and spatial parameter drifts during the strong coupling of multiple physical fields.

[0006] The applicability of these technologies is largely limited to applications with relatively uniform process parameter distribution, simple electrode structures, or relatively loose target consistency requirements. For the electroforming fabrication of biomimetic adhesion arrays of coronal tips, which require precise spatial control and highly consistent processes at the micro / submicron scale, there is a lack of systematic methods and equipment that can fully collect in-situ process information at multiple points in space and across multiple physical fields, and implement data-driven, full-process, closed-loop, intelligent spatial adaptive control.

[0007] The main problems of the existing technology are:

[0008] 1) Spatial resolution limitations: Most methods use global averaging or limited partitioning parameter settings, lacking the ability to perceive and respond to core sensitive areas (such as complex geometric shapes and microstructure arrays) with high spatial resolution in real time.

[0009] 2) Insufficient coupling of multiple physical fields: Mainstream process control is mostly limited to the independent adjustment of parameters of a single physical field (such as current, potential, temperature or flow rate), and it is difficult to deal with complex phenomena such as parameter coupling instability and spatial mutation caused by the interaction of multiple physical fields (current-temperature-fluid).

[0010] 3) Lack of data-driven and feedback closed loop: Existing control is mainly based on static settings, and there is insufficient deep mining of real-time monitoring data in the process and model prediction-feedback-adaptive iterative response, and there is a lack of early warning and automatic correction mechanism for process abnormalities.

[0011] 4) Batch consistency and process robustness are difficult to ensure: In multi-batch production and complex structure preparation, parameter fluctuations, regional segregation, micro-area process out of control and other phenomena are prone to occur, making it impossible to achieve self-learning optimization and version iterative upgrades of spatially distributed process parameters.

[0012] 5) Lack of a multi-scale and multi-physics field integrated scheduling system: There is a lack of a comprehensive solution that can combine the spatial heterogeneity of core mold structure, process disturbance response, parameter coupling of various regions and big data management capabilities.

[0013] Therefore, the industry urgently needs a method and system that can organically combine high-spatial-resolution multi-physical field in-situ acquisition, data-driven simulation and adaptive fuzzy control, multi-channel intelligent execution and closed-loop self-learning optimization to achieve precise collaborative management of multi-physical fields such as current, temperature, and fluid in the electroforming preparation of complex bionic structures. Summary of the Invention

[0014] In order to solve the problems existing in the above-mentioned prior art, the purpose of this application is to provide a crown-shaped bionic adhesion array electrodeposition method, aiming to solve the problem or one of the problems existing in the above-mentioned prior art.

[0015] The present invention provides a method for electrodeposition of a crown-shaped biomimetic adhesion array, comprising the following steps:

[0016] S1: Micro current density sensors, temperature sensors, and fluid flow rate sensors are arranged in each key area of ​​the crown-shaped end bionic adhesion array electroforming core mold structure and its corresponding electrolyte space, respectively, to collect real-time current density, temperature, and fluid flow rate sampling data during the electrodeposition process in different areas.

[0017] S2: performing denoising and spatial consistency correction on the current density, temperature and fluid flow rate sampling data collected from the multiple regions to obtain a high-precision in-situ physical parameter data set at each sampling location.

[0018] S3: Based on a high-precision in-situ physical parameter dataset and combined with the core mold morphology information of each region, the spatial electric field distribution, temperature distribution, and fluid distribution simulation values ​​of the electrolyte and the core mold surface in each region are calculated by inputting them into a multi-physics field coupling numerical analysis model.

[0019] S4: Based on the spatial distribution simulation values ​​of the multi-physics field coupling numerical analysis model and the real-time collected physical parameter data, the growth rate of the sedimentary layer in each region and the spatial distribution characteristics of the surface uniformity are calculated, and the surface uniformity is used as the regional performance evaluation indicator.

[0020] S5: Based on the regional performance evaluation index, the adaptive fuzzy control algorithm is applied to generate the corresponding local current amplitude adjustment parameters, pulse waveform setting parameters and micro-area stirring intensity adjustment parameters according to the spatial distribution differences of current density, temperature and fluid flow velocity in each region.

[0021] S6: Input the local current amplitude adjustment parameters, pulse waveform setting parameters and micro-area stirring intensity adjustment parameters into the multi-channel partitioned power supply equipment and the micro-area stirring device, perform targeted parameter modulation operations, and drive the electrodeposition equipment to perform regional process control according to the spatial distribution strategy.

[0022] S7: During the electrodeposition process, the current density, temperature and fluid flow rate feedback data of each area are cyclically collected and dynamically compared with the aforementioned simulation values ​​and performance evaluation indicators. Based on the comparison results, the output parameters of the adaptive fuzzy control algorithm are corrected in real time to achieve closed-loop spatial distribution process control.

[0023] S8: Perform multi-scale characterization measurements on the crown-shaped biomimetic adhesion arrays in each region after deposition to evaluate the consistency and spatial uniformity of the deposited layer growth. Input the characterization results into the self-learning optimization module to generate a new generation of electrodeposition parameter distribution correction factors.

[0024] S9: Based on the electrodeposition parameter distribution correction factor generated by the self-learning optimization module, batch parameter group version switching is executed, and the spatial distribution parameter rules for subsequent batches of core mold preparation are continuously optimized and iterated to improve the consistency and process robustness of large-scale preparation.

[0025] Beneficial effects of the present invention:

[0026] 1) Multi-physics field high spatial resolution in-situ sensing, breaking through monitoring blind spots

[0027] This invention, for the first time, meticulously arranges micro-current density, temperature, and fluid flow rate sensors on the core mold surface and within the electrolyte space based on three-dimensional process-sensitive areas, enabling submicron-level, high-spatial-resolution, in-situ data acquisition throughout the entire process. Through spatial model-driven sensor partitioning and placement, it covers key areas of variation in the electric, thermal, and flow fields during the process, effectively avoiding issues such as missed regional mutations and feedback lags that are common with traditional single-point or coarsely distributed monitoring. This significantly improves the comprehensiveness and representativeness of process monitoring, providing strong sensory support for subsequent regional precision control.

[0028] 2) Multi-physics field numerical simulation and actual closed-loop coupling greatly improves prediction and control accuracy

[0029] This method achieves precise mapping of simulated and measured data in spatial coordinates and process timing through spatially consistent high-precision physical parameter calibration, regional data registration, and coupled finite element simulation with regional growth mechanism analysis. A dynamic correction mechanism for regional growth rate and surface uniformity indicators effectively addresses the challenges of traditional numerical models, which suffer from weak predictions and lagging control due to actual perturbations, and enables dynamic compensation of regional parameter errors.

[0030] 3) Intelligent partitioning and adaptive control, significantly improving spatial consistency and preparation quality

[0031] This invention creates a fuzzy adaptive regionalized collaborative control system. This system utilizes multi-field spatial coupling indicators to output, in real time, a set of parameters, including local current, pulse waveform, and micro-region stirring, that combine spatial resolution and dynamic response. This allows for closed-loop, precise management of the coordinated control of multi-channel power supply and micro-region stirring. Compared to traditional holistic or manual adjustment methods, this system can dynamically balance local non-uniformities caused by multi-physical field interference, significantly reducing fluctuations in thickness, morphology, and performance between array elements, significantly reducing the standard deviation of regional surface uniformity, and achieving batch repeatability and stability superior to industry standards. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a flow chart of a crown-shaped biomimetic adhesion array electroplating method described in this application. DETAILED DESCRIPTION

[0033] like Figure 1 As shown, the present invention discloses a method for electro-deposition of a crown-shaped biomimetic adhesion array, comprising the following steps:

[0034] S1: Micro current density sensors, temperature sensors, and fluid flow rate sensors are arranged in each key area of ​​the crown-shaped end bionic adhesion array electroforming core mold structure and its corresponding electrolyte space, respectively, to collect real-time current density, temperature, and fluid flow rate sampling data during the electrodeposition process in different areas.

[0035] S2: performing denoising and spatial consistency correction on the current density, temperature and fluid flow rate sampling data collected from the multiple regions to obtain a high-precision in-situ physical parameter data set at each sampling location.

[0036] S3: Based on a high-precision in-situ physical parameter dataset and combined with the core mold morphology information of each region, the spatial electric field distribution, temperature distribution, and fluid distribution simulation values ​​of the electrolyte and the core mold surface in each region are calculated by inputting them into a multi-physics field coupling numerical analysis model.

[0037] S4: Based on the spatial distribution simulation values ​​of the multi-physics field coupling numerical analysis model and the real-time collected physical parameter data, the growth rate of the sedimentary layer in each region and the spatial distribution characteristics of the surface uniformity are calculated, and the surface uniformity is used as the regional performance evaluation indicator.

[0038] S5: Based on the regional performance evaluation index, the adaptive fuzzy control algorithm is applied to generate the corresponding local current amplitude adjustment parameters, pulse waveform setting parameters and micro-area stirring intensity adjustment parameters according to the spatial distribution differences of current density, temperature and fluid flow velocity in each region.

[0039] S6: Input the local current amplitude adjustment parameters, pulse waveform setting parameters and micro-area stirring intensity adjustment parameters into the multi-channel partitioned power supply equipment and the micro-area stirring device, perform targeted parameter modulation operations, and drive the electrodeposition equipment to perform regional process control according to the spatial distribution strategy.

[0040] S7: During the electrodeposition process, the current density, temperature and fluid flow rate feedback data of each area are cyclically collected and dynamically compared with the aforementioned simulation values ​​and performance evaluation indicators. Based on the comparison results, the output parameters of the adaptive fuzzy control algorithm are corrected in real time to achieve closed-loop spatial distribution process control.

[0041] S8: Perform multi-scale characterization measurements on the crown-shaped biomimetic adhesion arrays in each region after deposition to evaluate the consistency and spatial uniformity of the deposited layer growth. Input the characterization results into the self-learning optimization module to generate a new generation of electrodeposition parameter distribution correction factors.

[0042] S9: Based on the electrodeposition parameter distribution correction factor generated by the self-learning optimization module, batch parameter group version switching is executed, and the spatial distribution parameter rules for subsequent batches of core mold preparation are continuously optimized and iterated to improve the consistency and process robustness of large-scale preparation.

[0043] The step S1 specifically includes:

[0044] S1.1 Based on the three-dimensional layout model of the electroformed core mold structure of the bionic crown-shaped end adhesion array, the core mold surface and its corresponding electrolyte space are mapped and modeled in different regions to obtain the spatial partitioning information of the key distribution area of ​​current density, the temperature-sensitive area and the fluid disturbance-prone area.

[0045] The input conditions are the three-dimensional structural model of the bionic crown-shaped end adhesion array electroforming core mold, combined with the core mold surface morphology parameters and its corresponding two-dimensional or three-dimensional electrolyte spatial distribution model.

[0046] A method based on three-dimensional modeling and region division (parameters: three-dimensional geometric dimensions of the core mold, surface micro-nanostructure details, and electrolyte flow boundary conditions) is used to achieve spatial correlation modeling between the core mold surface and the electrolyte space.

[0047] Furthermore, through the spatial mapping function, each microstructure unit on the core mold surface is matched one-to-one with the electrolyte space display partition, realizing the spatial partition grid division based on structural differences and obtaining preliminary spatial partition units.

[0048] A parameter sensitivity analysis algorithm (input: theoretical current density distribution field, temperature field and initial distribution of fluid field) is used to realize the spatial sensitivity identification of the current density change mutation area, temperature significant area and fluid disturbance variable area on the core mold surface and electrolyte space, and form a regional sensitivity heat map.

[0049] Furthermore, the above sensitivity heat maps are clustered and graded through spatial clustering algorithm and threshold segmentation technology, and the specific spatial partition boundaries of the key current density distribution area, temperature sensitive area and fluid disturbance variable area are clearly defined.

[0050] Vectorized spatial coding methods (such as three-dimensional coordinate coding and grid cell numbering) are used to output partition identifiers with high spatial resolution, and a mapping table from core spatial parameters to partition codes is established to achieve quantitative calibration of the locations of spatially sensitive areas.

[0051] Through the above-mentioned multi-level chain data processing process, the three-dimensional physical field partition model of the core mold structure and its corresponding electrolyte space is converted into high-resolution spatial partition data for current, temperature, and fluid disturbance sensitivity, providing spatial resolution and layout strategy decision-making support for the subsequent precise arrangement of sensors, and realizing the basic spatial architecture for efficient perception of multiple physical fields.

[0052] For example, for a micron-scale crown cell array core mold, the three-dimensional structural dimensions were set to 2000 μm in length, 1000 μm in width, and 100 μm in height. The surface contained hemispherical (10 μm diameter) microstructured cells, modeled with centrosymmetry of the cell array. The electrolyte spatial grid was partitioned using a cubic grid with a cell step size of 5 μm. Finite-difference sensitivity analysis was performed, assuming that the initial theoretical current density distribution followed the boundary layer enhancement model. Different initial current boundary conditions were set for cells at the center and edges of the surface. The temperature field was initialized to 25°C, and fluid-structure interaction simulations were performed using the heating flow rate as a parameter. The fluid field inlet velocity was 1 mm / s, and a low-Reynolds number steady-state flow was used within a Reynolds number of 1. The simulation sensitivity heatmaps were clustered using a 5% threshold to delineate critical current density regions (18% of the total area), temperature-sensitive regions (8% of the area), and fluid disturbance regions (15% of the area). Each region was assigned a unique spatial code to achieve high-resolution spatial segmentation. This partitioning data provides a quantitative coordinate basis for subsequent sensor placement, supporting efficient detection of sensitive areas. During the actual placement process, micro-current density sensors, temperature sensors, and fluid flow rate sensors are precisely implanted in corresponding sensitive areas according to the spatial encoding described above, enabling high-spatial-resolution monitoring of the process. This process has been verified, and the spatial partitioning accuracy has been improved to the submicron level, achieving representative signal acquisition in each sensitive area and enhancing spatial discrimination capabilities.

[0053] S1.2 For the key current density areas output by the spatial partition model, the finite element field analysis method is used to calculate the expected intensity distribution of the electric field inside the partition and the coupling characteristics of the fluid field and thermal field. Based on this, the specific coordinate positions of the micro current density sensors on the core mold structure need to be analyzed to achieve targeted distribution positioning.

[0054] The input conditions are the high spatial resolution three-dimensional spatial partitioning model output in step S1.1, as well as the spatial encoding and grid coordinate data sets of the key distribution areas of current density on the core mold surface and electrolyte space.

[0055] The finite element field analysis method (parameters: spatial partition model, boundary potential setting, material conductivity, and initial current boundary conditions) is used to perform electric field distribution simulation on each key current density partition, achieving high-precision prediction of the potential gradient and current density vector field within the partition.

[0056] Furthermore, through the multi-physics field coupling finite element simulation method (parameters: local fluid velocity, boundary temperature, thermal conductivity coefficient, electrode surface overpotential), the fluid field and thermal field are coupled and solved to obtain the instantaneous temperature distribution, flow field disturbance pattern and its spatial coupling response with the current distribution in the key area during the electrodeposition process.

[0057] By using numerical extreme point extraction and streamline tracing algorithms, we conduct in-depth analysis of local extreme points, current density mutation zones, and maximum gradient areas in the electric field simulation results, and automatically screen out candidate monitoring nodes that are most sensitive to current density changes.

[0058] Through the spatial multi-point sensitivity ranking algorithm, sensitivity weight values ​​are assigned to all candidate coordinates according to the current density change amplitude, stability index and gradient coupling degree of each candidate point with the surrounding grid nodes, and the optimal layout coordinates of the current density sensor are screened out.

[0059] Based on the finite element simulation space grid coordinates, the spatial mapping relationship is used to realize the conversion of the simulation coordinate system to the actual physical size of the core mold, and the preferred coordinates are mapped to the actual layout points on the core mold surface to form a specific distribution coordinate list of the micro current density sensor.

[0060] Through the above-mentioned algorithm and spatial mapping, the key sensitive areas of current density are quantitatively converted into specific layout positions of microsensors, achieving targeted distribution positioning and ensuring that the monitoring signals are highly representative.

[0061] For example, for a crown-shaped end mandrel with three-dimensional dimensions of 2000 μm in length, 1000 μm in width, and 100 μm in height, a hemispherical array surface with a 20 μm array spacing and a cell diameter of 10 μm, the input spatial partitioning grid length step is 5 μm. The boundary potential in the central region is -0.7 V, and at the edge it is -0.9 V. The electrolyte conductivity is 1 S / m, the initial temperature is 25°C, and the fluid inlet velocity is 1 mm / s. Finite element method calculations of the electric field distribution reveal that the current density peaks at approximately 1.4 A / dm² in the central cell, decreasing to 0.9 A / dm² at the edge. The transition zone from the center to the edge is a highly sensitive region of current density gradient. Corresponding thermal field simulations show that the temperature of the local high overpotential region rises slightly to 28°C, and the flow field disturbance causes a local peak fluid velocity of 1.2 mm / s. Using extreme point extraction and sensitivity ranking algorithms, eight coordinates were selected as micro-current density sensor placement points, including the central array core (X=1000 μm, Y=500 μm), the edge transition zone (X=100 μm, Y=500 μm), and the four corner critical points (X=50 μm, Y=50 μm). These coordinates were then mapped to the specific locations on the actual mandrel surface. Simulation and field measurements demonstrated that this placement achieved over 90% spatial variation monitoring coverage of the mandrel surface current density distribution, while also providing real-time feedback with a high signal-to-noise ratio. This provides a foundation for high-precision process parameter monitoring for subsequent spatially distributed adaptive process control.

[0062] Based on the results of fluid field simulation and temperature gradient prediction, S1.3 adopts a weighted regional sensitivity evaluation algorithm to identify nodes in the electrolyte space that respond significantly to changes in fluid flow rate and temperature, and uses them as the preferred layout points for micro fluid flow rate sensors and temperature sensors to ensure that each physical parameter collection point is representative of process disturbances.

[0063] S1.4 uses a highly integrated packaging process according to the distribution coordinates obtained from the evaluation to implant micro current density sensors, temperature sensors, and fluid flow rate sensors in key areas of the core mold surface and corresponding positions of the electrolyte space. Micro-process interconnection cables are used to achieve regional signal output, providing a physical networking foundation for subsequent multi-channel acquisition system docking.

[0064] S1.5 performs interconnection integrity testing and static parameter consistency correction on all deployed micro current density sensors, temperature sensors, and fluid flow rate sensors to ensure the consistency and reliability of the physical parameter signals output by each acquisition channel under the same environment, and establish a high-precision basic data link for real-time in-situ physical parameter acquisition and data processing during the deposition process.

[0065] The step S2 specifically includes:

[0066] S2.1 uses a multi-channel time-series synchronous acquisition module to uniformly structure the data for all process original sampling data, targeting the current density measurement signals, temperature measurement signals, and fluid flow rate measurement signals output by multi-point sensors in the core mold key area and electrolyte space, so as to obtain a process original sampling data set that is fully associated with the spatial area number and acquisition time.

[0067] Based on the original sampling data set of the process in each spatial area, S2.2 uses an adaptive wavelet denoising algorithm to perform multi-scale noise separation and removal on the current density measurement signal, temperature measurement signal and fluid flow rate measurement signal, respectively, so as to retain the key physical change trends to the greatest extent and eliminate high-frequency random interference, thereby obtaining the process characteristic signal after spatial distribution denoising.

[0068] For the original process sampling data set (including the spatial region number, timestamp and original physical quantity sequence) collected by multi-region micro current density sensors, temperature sensors and fluid flow rate sensors, an adaptive wavelet denoising algorithm (parameters: signal length, number of decomposition levels, wavelet basis type, threshold selection rules) is used to perform multi-scale decomposition of the current density measurement signal, adaptively determine the optimal wavelet decomposition level and mother wavelet function, and realize the joint analysis of the current density signal at each spatial point in time and frequency domains.

[0069] Furthermore, the multi-scale threshold method is used to perform high-frequency noise identification and soft and hard threshold denoising operations on the decomposed detail coefficients and approximate coefficients respectively. According to the signal-to-noise situation of the spatial area and the actual noise spectrum distribution, the denoising intensity of each component is adaptively adjusted to separate and remove the high-frequency random interference components, and strictly retain the main process signal that reflects the dynamics of the actual physical process.

[0070] The same wavelet denoising process is used (parameter adjustment: optimizing the mother wavelet and decomposition level according to the temperature signal amplitude range and background noise type). Time-frequency multi-scale decomposition and adaptive threshold denoising are performed on the temperature measurement signal and fluid flow rate measurement signal, respectively. This effectively suppresses noise interference from the electromagnetic environment, human operation, and other factors, and enhances the low-frequency physical response characteristics of the temperature and fluid signals.

[0071] For spatially distributed sampling data, the wavelet multi-scale denoising process is executed in batches according to the spatial numbering to ensure the algorithm consistency and characteristic response homology of process signals in different spatial regions during the noise reduction process, and to achieve highly consistent noise reduction conversion of multi-point sampling signals within the partition.

[0072] Through the above-mentioned adaptive wavelet denoising algorithm, each set of spatial region current density, temperature and fluid flow rate signals outputted are respectively subjected to denoising process characteristic signals with high-frequency clutter fully removed, thereby achieving the goal of retaining the physical change trend to the greatest extent and significantly eliminating measurement noise, providing a high-quality signal foundation for subsequent spatial collaborative filtering and consistency correction.

[0073] For example, 600 sets of raw time-series process signals were obtained for each of eight typical sensitive coordinates on the array core mold surface (e.g., center, edge, and transition zone coordinates) using a 2 Hz sampling frequency and a single-point sampling duration of 600 s. The raw current density signal was decomposed using the db4 wavelet basis, with a five-level decomposition layer and the SureShrink adaptive threshold optimization rule as the threshold function. Soft thresholding was applied to the third and fourth layers of high-frequency detail coefficients, achieving a noise suppression rate of 85% and an increase in the signal correlation coefficient (compared with the actual process variation trend) to 0.97. The temperature and fluid flow rate signals were decomposed into four layers using the sym6 wavelet basis. The thresholds were adaptively selected based on the statistical model of the background noise at each point, achieving average noise suppression rates of 80% and 78%, respectively. After denoising, the physical fluctuation trends of the denoised current density, temperature, and fluid flow rate signals across all spatial regions were fully consistent with the experimentally investigated process disturbances. The principal component proportion of the spatially distributed characteristic signals increased to over 95%, effectively supporting subsequent spatial uniformization processing and multi-field coupling modeling applications.

[0074] Through adaptive wavelet noise reduction algorithm processing, the original process signal collected in the previous step is converted into a process characteristic signal after spatial distribution noise reduction, achieving significant denoising of high spatial resolution multi-physical field signals and accurate precipitation of physical intrinsic trends, laying the signal processing foundation for high-consistency parameter correction and intelligent control of bionic arrays.

[0075] S2.3 is based on the physical field spatial collaborative filtering algorithm for the process characteristic signals after noise reduction, and utilizes the spatial correlation between multi-point sensors and the consistency of sensor dynamic response to perform spatial interpolation correction and boundary normalization processing on the abnormal point data, thereby obtaining process-consistent physical characteristic data with spatial global coordination.

[0076] For the process characteristic signals of current density, temperature and fluid flow rate after spatial distribution noise reduction, a regional spatial correlation dataset is constructed based on the measurement signals of multiple point sensors in each region.

[0077] The physical field spatial collaborative filtering algorithm (parameters: spatial neighborhood size, signal sampling frequency, physical correlation threshold) is used to achieve synchronous processing of multi-point sensor signals in each spatial area and extract spatial collaborative response characteristics.

[0078] Furthermore, the spatial correlation calculation method is based on the following formula:

[0079]

[0080] in, is the spatial correlation coefficient of the sensor signals at position i and j, and is the regional characteristic signal at each sampling time t, and are their respective means, T is the number of samples, and the spatial correlation matrix of multi-point signals is constructed.

[0081] Through the outlier detection and identification algorithm (parameters: median absolute deviation threshold, spatial consistency criterion), the abnormal measurement points and mutation boundary points in the spatial data are calibrated, and the spatial interpolation correction algorithm is used to complete the data at the abnormal points. Spatial interpolation uses the weighted inverse distance method, and the specific formula is as follows:

[0082]

[0083] in, is the signal value after compensation of the kth abnormal point, is the signal value of the spatially adjacent valid point, is the spatial distance between point i and point k, p is the weighting coefficient, and N is the number of spatial neighboring points, which realizes the spatial physical consistency reconstruction of outliers.

[0084] Furthermore, the sensor signals at the region boundary are normalized using a boundary normalization processing algorithm (parameters: boundary determination threshold, normalization scale interval), using the following normalization formula:

[0085]

[0086] in, is the normalized signal, is the original signal, It is the signal set of all regions to be normalized, ensuring that the signal scales of the boundary regions are consistent.

[0087] Through the above-mentioned chain collaborative filtering, spatial correlation interpolation and normalization processing, the noise reduction signals of multi-point sensors are converted into spatially distributed and globally coordinated process-consistent physical characteristic data, achieving high-consistency modeling of the physical field information in the core area and the boundary area, providing a standardized physical input basis for subsequent spatially distributed process parameter calibration and intelligent process control.

[0088] For example, the current density, temperature, and fluid flow rate signals from eight micro-regions on the surface of the array core mold were spatially de-noised. A time series length of 600 and a spatial neighborhood radius of 20 μm were collected. Spatial correlation matrix calculations revealed correlation coefficients between central and peripheral units exceeding 0.92, with only a few points falling below 0.85. For some signal points detected as abnormal due to transient interference, weighted inverse distance spatial interpolation (weighted power exponent p=2) was used to correct the signal. The standard deviation of the interpolated signal was reduced to ±2% of the global mean. For all normalized signals in the edge regions, the distribution of physical quantities was uniformly normalized to the interval [0, 1], and the overall spatial signal principal component normalization was improved to 98%. The final output was a process-consistent physical characteristic dataset of current density, temperature, and fluid flow rate across the entire spatial domain, providing a unified physical signal benchmark for high-precision input into multi-physics numerical models and validation of spatially partitioned parameter modeling. The spatial distribution physical consistency was improved by over 12% compared to the pre-noise reduction period.

[0089] S2.4 is based on process-consistent physical characteristic data and applies a spatially distributed multi-sensor calibration method to perform sensor bias correction and calibration factor compensation on the measurement data of each region, eliminating the systematic drift between regions introduced by differences in micro-sensor manufacturing processes and assembly errors, and obtaining high-precision spatially distributed process parameters after calibration and correction.

[0090] The method uses spatially distributed and uniform process characteristic data as input, and covers the physical parameters of current density, temperature and fluid flow rate in each region after collaborative noise reduction and spatial interpolation normalization.

[0091] A spatially distributed multi-sensor calibration method (parameters: sensor initial position number, spatial region index, and time-synchronized sampling sequence) is used to achieve systematic bias diagnosis and parameter consistency modeling of measurement signals in each region.

[0092] Furthermore, through the regional sensor calibration factor extraction algorithm (parameters: historical reference signal, process stability period calibration sequence, theoretical physical field distribution model), a mapping relationship between each sensor measurement value and the standard process physical quantity is established, and the system bias introduced by manufacturing process differences and assembly errors is quantified.

[0093] Furthermore, bias correction is performed on the output data of each sensor channel through the least squares fitting method and the multi-region correction matrix solution (parameters: spatial neighborhood reference signal set, calibration reference point weight), and the correction factor is obtained as follows:

[0094]

[0095] in is the measured value after calibration in the i-th region, is the original output signal, is the sensor’s offset calibration factor, is the scale calibration factor, is the corresponding optimization weight, which is determined optimally by the calibration algorithm.

[0096] Furthermore, a spatial distribution calibration consistency assessment method (parameters: partition global standard deviation threshold, regional normalization coefficient) is used to perform global alignment and error verification on all calibrated measurement data with the standard theoretical physical field distribution, so as to achieve unified data scales between regions and eliminate the system drift effect.

[0097] Through the above-mentioned chain calibration and correction algorithm, the output signals of each multi-point micro-sensor are processed through bias correction, scale factor compensation and spatial consistency alignment, and a calibrated and corrected high-precision spatially distributed process parameter data set is output, thereby eliminating the systematic drift effect between regions caused by manufacturing process and assembly errors, and laying a reliable data foundation for multi-physical field coupling analysis and spatial adaptive control.

[0098] For example, for the eight spatially partitioned current density sensors arranged on the coronal bionic array core mold, three of them come from the same manufacturing batch and the rest are manually embedded. In actual measurements, the initial output has regional deviations compared with the theoretical model, with the maximum bias reaching 8% and the scale error reaching 5%. Through historical reference signal acquisition, the core mold is placed in a calibration environment with a uniform standard current of 300mA / cm², a temperature of 25°C, and a fluid velocity of 1 mm / s, and the standard output of each sensor during the stable process period is collected. Using the least squares regression method, the bias factor range of each channel is -0.18 to +0.10, and the scale factor is 0.92-1.07. Input formula calibration weight average result , After calibration, the average deviation between the current density signal and the theoretical value in each region was reduced to 0.6%, and the maximum drift of the temperature and fluid signals was less than 0.3°C and 0.02 mm / s. The global spatial consistency index was improved to 99.2%, and regional drift was completely eliminated. The output of the calibrated and corrected high-precision spatial distribution side parameters provides high-quality basic data for downstream multi-physics field modeling and feedback control, achieving the industry's highest level of inter-regional parameter accuracy.

[0099] S2.5 is based on the calibrated and corrected high-precision spatially distributed process parameters. It establishes a spatial parameter data set index table through unified coding of process physical quantities and regional mapping, ensuring that each set of physical parameters accurately corresponds to a specific core mold morphology spatial area, providing a standardized input basis for multi-physics field coupling numerical models and subsequent regionalized process control strategies.

[0100] The step S3 specifically includes:

[0101] S3.1 performs regional data segmentation operations on the high-precision in-situ physical parameter dataset to classify the current density, temperature, and fluid flow rate data of each spatial sampling point into the corresponding electroforming core mold structure area and its matching electrolyte space, ensuring that each physical parameter sampling point corresponds one-to-one to its spatial position information, and obtaining a regionalized physical parameter classification dataset.

[0102] S3.2 is based on the regionalized physical parameter classification data set, and the corresponding core model morphology information is retrieved for each spatial sampling area, including surface geometric features, dimensional parameters and crown structure microtopology. Data alignment and spatial registration algorithms are used to form a multidimensional input data set containing the coupling characteristics of regional physical parameters and core model morphology.

[0103] S3.3 takes as input a multi-dimensional input data set based on the coupling characteristics of regional physical parameters and core mold morphology, adopts a multi-field coupling finite element numerical analysis model, performs spatial electric field distribution simulation on each specified area, obtains the local electric field intensity distribution simulation value between the regional electrolyte and the core mold surface, and outputs it as a spatial electric field distribution prediction matrix.

[0104] S3.4 uses the spatial electric field distribution prediction matrix as the factor, combines the regional temperature parameter history, and based on the multi-field thermoelectric coupling solution algorithm, calculates the spatial temperature distribution simulation value of each core mold structure area, outputs the spatial temperature distribution prediction matrix, and realizes spatial distribution thermal field modeling.

[0105] S3.5 takes the spatial temperature distribution prediction matrix, current density parameters and fluid flow rate as input, applies the multi-field coordinated fluid dynamics and mass transfer numerical model, simulates the fluid flow field distribution in each specified area, obtains the flow velocity vector field and flow state distribution simulation values ​​of the electrolyte near each micro-area of ​​the core model, and forms a spatial fluid distribution prediction matrix.

[0106] S3.6 performs spatial data integration and regional result calibration operations on the above-mentioned spatial electric field distribution prediction matrix, spatial temperature distribution prediction matrix and spatial fluid distribution prediction matrix, outputs a multi-physical field collaborative distribution simulation value set, realizes a complete quantitative characterization of the multi-physical field spatial coupling response of each regional process, and provides multi-dimensional benchmark prediction data for subsequent partition process control parameter generation and deposition process data feedback comparison.

[0107] The step S4 specifically includes:

[0108] S4.1 aligns the time and space domain data of the high-precision in-situ physical parameter data set (including current density, temperature, and fluid flow rate data) that has undergone spatial consistency correction with the spatial distribution simulation values ​​of the multi-physical field coupling numerical analysis model in the same area to form a regional distribution physical parameter comparison matrix, providing the basic conditions for the subsequent extraction of deposition rate and surface uniformity characteristics.

[0109] A unified data numbering system and spatial labels are used to align the data structures of high-precision in-situ physical parameter data sets (including current density, temperature, and fluid flow rate data) that have undergone spatial consistency correction with the spatial distribution simulation values ​​of the multi-physical field coupling numerical analysis model in the same area, ensuring that each in-situ measured data point corresponds to the spatial coordinates, parameter type, and timestamp of its corresponding simulation prediction point.

[0110] A time-domain multi-channel synchronous alignment algorithm is used, based on time series interpolation and resampling technology, to achieve high-precision alignment of in-situ physical parameter data and simulation data on the time axis, eliminating timing errors caused by measurement delays, inconsistent sampling frequencies, and system response lags.

[0111] By combining spatial domain data with filtering and regional sampling grid interpolation methods, the spatial coordinate deviation caused by differences in sensor layout density and simulation grid scale is corrected for the complex spatial partitioning of the core model structure, and the measured-simulation data pairs after spatial consistency alignment are obtained.

[0112] The regional distribution parameter comparison matrix generation method is adopted. The measured parameters and simulated parameters that have been strictly aligned in the time domain and space domain are used to construct a two-dimensional parameter comparison matrix with the regional number as the index group. The matrix elements are filled with the measured values ​​and simulated values ​​of the same physical quantity respectively.

[0113] Furthermore, through the standardized data coding and outlier removal algorithm, the data outliers in the registration matrix are screened and corrected, and the data mismatches introduced by measurement errors or simulation boundary anomalies are eliminated or compensated, thereby improving the robustness and representativeness of the registration matrix.

[0114] Through the above-mentioned algorithms and processing methods, the high-precision in-situ measured parameter data after spatiotemporal synchronous registration and the multi-physical field coupling simulation values ​​are converted into a regional distribution physical parameter comparison matrix, thereby realizing the structuring of basic data for extracting characteristic quantities such as sedimentary layer growth rate and surface uniformity, and providing high-precision, traceable, and spatial data input for the implementation of subsequent growth rate, uniformity statistics, and performance evaluation algorithms.

[0115] For example, a batch of electroformed samples of coronal biomimetic adhesive array core molds were prepared. For example, in-situ current density microsensors were placed at a density of one per square millimeter, and temperature and fluid flow rate sensors were placed with a 10mm step size per spatial grid. The corresponding multiphysics simulation grid size was 2mm. The measured data were sampled at a 100Hz frequency, and the simulated data were output at a frequency of one frame per minute. A linear interpolation-based temporal alignment algorithm was used to resample all measured data to one frame per minute. The spatial numbers of the core mold surface and the electrolyte region were used as unique labels to achieve a one-to-one match with the simulated output. Bilinear spatial interpolation was used to compensate for spatial coordinate discrepancies caused by varying placement densities and offset simulation grid boundaries, and the data for each physical quantity were organized into a unified spatial region number matrix. Based on the spatially and temporally aligned dataset, a parameter comparison matrix consisting of 60×60 spatial nodes was constructed, with each node simultaneously recording current density, temperature, and fluid flow rate. Outliers were removed using the 3σ principle, and missing points were filled using the nearest neighbor method. The output registration matrix has been verified to have a spatial coordinate error of less than 0.5 mm, a timing error of less than 10 ms, and no significant abnormal data, meeting the data consistency requirements for subsequent statistical modeling of deposition growth rate and surface uniformity.

[0116] S4.2 is based on the regional distribution physical parameter comparison matrix and uses a physical-driven growth mechanism analysis model to calculate the sediment growth rate distribution per unit time in each region, and outputs a sediment growth rate matrix containing spatial coordinate resolution as an important process output to measure the consistency of spatial distribution.

[0117] The comparison matrix of regional distribution physical parameters (including measured and simulated data such as current density, temperature, and fluid flow rate) is processed using a physics-driven growth mechanism analysis model.

[0118] Collect the current density of each spatial area of ​​the core module ,temperature and fluid flow rate Parameters such as growth rate and growth factor serve as the core input of the growth mechanism analysis model.

[0119] The electrodeposition rate physics-driven model is used to correlate the metal ion reduction rate with the surface current density according to Faraday's law. The deposition growth rate of the spatial node is calculated using the following formula:

[0120]

[0121] in, In space coordinates The growth rate of the sediment layer per unit time at is the molar mass of the deposited metal, is the number of reaction electrons, is the Faraday constant, is the density of the deposited metal, is the current density.

[0122] Furthermore, by considering the temperature and deposition reaction rate constant The relationship between the two is to embed the Arrhenius equation into the rate model:

[0123]

[0124] in, is the prefactor, is the activation energy, is the gas constant.

[0125] Temperature-corrected metal ion mobility Corrected deposition rate, taking fluid flow rate into account Considering the influence of ion transport and interface concentration polarization, the following comprehensive growth rate model is established:

[0126]

[0127] in, is the fluid velocity correction function (for example, based on the Sherwood number model, reflecting the fluid shear enhancement mass transfer effect), which can be specifically set as:

[0128]

[0129] in is the fluid field influence coefficient.

[0130] Furthermore, the above model is applied to all spatial nodes in the registration matrix through spatial traversal and point-to-point operations, and the sediment growth rate matrix containing spatial coordinate resolution is output.

[0131] After regional accumulation and spatial grid statistics, the output ( A two-dimensional deposition layer growth rate matrix (where is the spatial number) is constructed to achieve quantitative analysis of the growth rate distribution in the entire array area.

[0132] Through the above-mentioned multi-physics field driven analytical model, the in-situ parameters and physical mechanisms in the micro-region are mapped into quantitative outputs of regional growth rates, providing core process output data for subsequent surface uniformity evaluation and adaptive process control.

[0133] For example, taking a batch of biomimetic crown end adhesion arrays as an example, the algorithm is set to input the first Parameters of spatial nodes: current density ,temperature , fluid velocity , taking copper deposition ( , , , ),set up (normalized), , , fluid influence coefficient .

[0134] First, calculate the current density contribution according to the Faraday formula:

[0135]

[0136] Calculate the Arrhenius temperature correction factor:

[0137]

[0138] Calculate fluid flow correction parameters:

[0139]

[0140] Comprehensive sedimentation growth rate:

[0141]

[0142] Calculate each of the 60\times60 spatial nodes point by point according to the above steps and output the sedimentary layer growth rate matrix All elements in the matrix are marked with spatial coordinates, physical units and process parameter sources. After multiple rounds of subsequent optimization, the growth rate matrix is ​​used to evaluate the spatial distribution uniformity and realize the closed-loop correction of the control parameters. Finally, the standard deviation of the coating growth rate between the core mold surface area and the array units is reduced from the initial Reduce to , which fully demonstrates the technical value of this step in improving the consistency of deposition space.

[0143] S4.3 Apply the surface morphology contour difference and spatial homogeneity analysis algorithm to perform two-dimensional or three-dimensional spatial statistical analysis on the deposition layer growth rate matrix, extract the surface uniformity characteristics of each key area (such as standard deviation, range and other statistical parameters), and form a uniformity index matrix that describes the consistency of the coating surface in each area.

[0144] Input data include: regional distribution of sediment growth rate matrix , the physical parameters of each spatial node (current density, temperature, fluid flow rate, etc.), as well as the region number and spatial distribution label.

[0145] The surface topography contour difference analysis method (parameters: spatial grid resolution and node coordinates) is used to realize the sediment growth rate matrix The two-dimensional or three-dimensional surface morphology is reconstructed for each key area in the image.

[0146] Furthermore, through the spatial homogeneity analysis algorithm (parameters: local sampling window size, global statistical division, spatial distance threshold), the spatial distribution statistics of the sedimentation rate values ​​in each area are realized, and a quantitative description of the surface uniformity at different spatial scales is obtained.

[0147] The spatial distribution fluctuation of sedimentation rate in each region was calculated by using the uniformity feature extraction method (parameters: statistical indicators such as mean, variance, range, and coefficient of variation).

[0148] Sedimentation rate data for each area , calculate the standard deviation as follows , extreme Key statistical indicators such as:

[0149]

[0150] in, is the number of spatial nodes in the region, is the deposition rate of each node in the region, is the average sedimentation rate in the area.

[0151]

[0152] in, and are the maximum and minimum values ​​of the node deposition rate in the region, respectively.

[0153] Furthermore, through the multi-region uniformity feature aggregation method, the uniformity statistical indicators of all key regions are combined to generate a complete uniformity index matrix , the matrix elements correspond to the standard deviation, range and other statistical characteristics of each region, achieving global quantification of spatial distribution consistency.

[0154] Through the above-mentioned surface morphology contour difference and spatial homogeneity analysis algorithm, the spatial distribution characteristics in the deposition layer growth rate matrix are converted into a uniformity index matrix that describes the surface consistency of the coating in each area, thereby achieving the goal of quantitative evaluation of the spatial consistency of the deposition process.

[0155] For example, in the actual electroforming batch of coronal end biomimetic adhesion arrays, it is assumed that the region for spatial nodes, input deposition rate data .

[0156] Using the above standard deviation formula,

[0157] ,

[0158]

[0159] Using the range formula,

[0160]

[0161] Will Equal values ​​are recorded in the uniformity index matrix Corresponding area element.

[0162] Calculate and compare in different areas or array units according to this method The statistical results show that before the implementation, the standard deviation of uniformity between regions was generally , after implementing the chain algorithm, the standard deviation is reduced to The following is the application of spatial statistical analysis and evaluation, which can significantly reflect the improvement of the surface consistency of the deposited layer and provide a quantifiable spatial uniformity index reference for subsequent process closed-loop optimization and adaptive adjustment. The final output is a structured spatial uniformity index matrix. , used to describe and compare the surface consistency of each deposition area and achieve the goal of process performance optimization.

[0163] S4.4 uses the error compensation algorithm to implement dynamic correction in the sedimentary layer growth rate matrix and the surface uniformity index matrix according to the error analysis results between the simulation values ​​output by the spatial distribution simulation model and the real-time high-precision in-situ physical parameter data, and outputs the compensated regional sedimentary layer growth rate matrix and surface uniformity index matrix to comprehensively improve the timeliness and accuracy of regional performance evaluation.

[0164] The input data is the comparison matrix of regional distribution physical parameters after spatial registration in the previous sub-step, which specifically includes the high-precision in-situ measured physical parameters (current density, temperature, fluid flow rate) of each spatial node and the spatial distribution simulation value of the multi-physics field coupling model, as well as the sediment layer growth rate matrix obtained by the previous homogeneity analysis. and surface uniformity index matrix .

[0165] The spatial error analysis method (parameters: spatial node number, physical quantity type, model output / measured data set) is used to achieve point-by-point comparison between measured values ​​and simulated values ​​in the regional distribution physical parameter comparison matrix to obtain the instantaneous error matrix at each spatial node. ,Right now:

[0166]

[0167] in, represents the model prediction value, is the measured value, Number the spatial nodes.

[0168] Furthermore, the error statistical feature extraction algorithm (parameters: sliding window, regional aggregation rule) is used to calculate the error statistical indicators in each spatial region, including mean, variance, range, etc., to form a spatial error statistical feature matrix , used to provide a parameter basis for subsequent dynamic error compensation.

[0169] Furthermore, a dynamic error compensation algorithm (parameters: regional error characteristics, correction coefficient adaptive adjustment rules) is used to calculate the sedimentation layer growth rate matrix. and surface uniformity index matrix The error adaptive correction is performed respectively. The growth rate matrix before correction is Taking θ as an example, the dynamic compensation of spatial nodes is completed through the following correction formula:

[0170]

[0171] in, To compensate for the corrected sediment growth rate, is the rate compensation coefficient, is the representative error of the region / node.

[0172] Correspondingly, the surface uniformity index matrix , apply the following consistency error compensation formula:

[0173]

[0174] in, To compensate for the corrected surface uniformity index, is the uniformity correction factor.

[0175] Furthermore, through the regional dynamic weight optimization algorithm (parameters: regional importance weight factor, sample historical iteration data), Adaptive adjustment can achieve targeted error compensation for different spatial partitions and areas with different stability levels.

[0176] Through the closed-loop incremental correction method (parameters: process cycle feedback window, historical accumulated error), the compensation results are updated in a rolling manner during the multi-cycle / batch process preparation process to ensure that dynamic error compensation can cover short-term fluctuations and long-term stability loss, and achieve drift suppression for long-term process evaluation.

[0177] Through the above-mentioned point-by-point error compensation algorithm and dynamic weight fusion correction, the deposition layer growth rate matrix and surface uniformity index matrix output in the previous step are converted into compensated regional index 、 , which significantly improved the timeliness and accuracy of regional performance evaluation, realized the dynamic management of spatial distribution deviations under the interaction of multiple physical fields, and provided a high-precision spatial performance input basis for subsequent process adaptive control and closed-loop optimization.

[0178] For example, in a batch of crown end biomimetic adhesion array space node grids are In the preparation process test, every 10 nodes are used as a local compensation area, and the corresponding spatial error statistical characteristics are: , compensation coefficient , . Initial deposition rate matrix center sub-region node , the node rate after compensation is:

[0179]

[0180] The surface uniformity index is initially , after compensation:

[0181]

[0182] After applying the periodic rolling error compensation, the difference between the regional growth rate and the surface uniformity index is less than , the standard deviation of growth rate was reduced to It can be seen that through the dynamic spatial error compensation processing in this step, the accuracy of the spatial performance evaluation of the deposition layer is significantly improved, providing reliable spatial basic data for subsequent partition adaptive control and process closed-loop optimization.

[0183] S4.5 takes the compensated and corrected regional deposition layer growth rate matrix and surface uniformity index matrix as input, and generates comprehensive regional performance evaluation indicators that reflect the deposition consistency and spatial distribution uniformity of the coating in each key area through normalization processing and multi-index weighted fusion algorithm, providing decision support output for the intelligent control system parameter adaptive adjustment and quality feedback evaluation.

[0184] The step S5 specifically includes:

[0185] S5.1 performs regional data mapping extraction based on the performance evaluation model for the regional performance evaluation indicators output in the previous step to obtain the spatial distribution characteristics of current density, temperature and fluid flow rate in each region for subsequent parameter optimization modeling.

[0186] S5.2 is based on the spatial distribution characteristics of regional current density, temperature and fluid flow rate, and uses the regional correlation measurement algorithm to calculate the coupling correlation between various physical parameters, form the spatial coupling correlation matrix of process parameters in each region, and construct the input feature vector for the fuzzy inference rule base.

[0187] S5.3 takes the spatial coupling association matrix of the process parameters of each area as input, calls the adaptive fuzzy control algorithm, and generates a set of target parameters for each area based on the multi-input multi-output (MIMO) fuzzy inference system, including the local current amplitude target, pulse waveform setting target and micro-area stirring intensity target, and preliminarily completes the distribution calculation of the adaptive control parameters.

[0188] S5.4 further implements a dynamic weight adjustment method for the target parameter set of each region generated by the adaptive fuzzy control algorithm, integrates the regional process sensitivity factor and the historical parameter correction factor, and optimizes the accuracy and response rate of the local current amplitude adjustment parameters, pulse waveform setting parameters and micro-area stirring intensity adjustment parameters of each region.

[0189] S5.5 outputs the weight-optimized local current amplitude adjustment parameters, pulse waveform setting parameters and micro-area stirring intensity adjustment parameters, forming a control parameter set with spatial resolution and process adaptability, which serves as input for the subsequent multi-channel partitioned power supply system and micro-area stirring device to perform targeted process modulation, ensuring the spatial uniformity and consistency of the electrodeposition process.

[0190] The step S6 specifically includes:

[0191] S6.1 performs digital signal verification and amplitude normalization on the local current amplitude adjustment parameters output by the adaptive fuzzy control algorithm to ensure that the multi-channel partitioned power supply equipment can accurately accept and execute the parameter settings, thereby laying the parameter foundation for the next step of spatial partitioned current drive.

[0192] S6.2 is based on the normalized local current amplitude adjustment parameters and adopts a real-time master-slave partition trigger mechanism to map the parameters of each area to the multi-channel partition power supply equipment, and perform partition synchronous current control operations, so that the core mold surface of each spatial area obtains a spatial current distribution consistent with the algorithm output, realizing closed-loop control of spatial current amplitude adjustment.

[0193] S6.3 performs waveform timing parameter analysis on the pulse waveform setting parameters output by the adaptive fuzzy control algorithm, and applies it to the multi-channel partitioned power supply equipment in combination with the spatial partitioning information, so that each area can load the required pulse waveform synchronously or asynchronously, realizing multi-zone timing fine pulse control, thereby further refining the spatial current driving effect.

[0194] S6.4 sets the output results of the local pulse waveform parameters, executes the pulse waveform transition compensation algorithm based on the current response time, waveform conversion rate and other hardware physical constraints of the actual electroplating equipment, and generates transition correction pulse parameters for different spatial regions to ensure the physical feasibility and real-time performance of the partitioned pulse drive.

[0195] S6.5 performs micro-area spatial mapping on the micro-area stirring intensity adjustment parameters output by the adaptive fuzzy control algorithm, and uses the spatial distribution mapping function to input it into the micro-area stirring device controller to achieve independent control of the stirring intensity, speed and direction of each spatial area, thereby forming a fluid flow field control mode consistent with the spatial current waveform.

[0196] Under the joint action of multi-channel partitioned power supply equipment and micro-area stirring device, S6.6 performs spatial distribution parameter modulation operations on the electroplating process of key areas on the core mold surface, drives the target area to achieve adaptive growth according to the parameter space mapping rules, and outputs an electroplated layer with uniform spatial distribution driven by process parameters, thereby achieving the goal of this round of spatial regulation.

[0197] The step S7 specifically includes:

[0198] S7.1 periodically collects current density, temperature, and fluid flow rate feedback data from the regionalized electrodeposition environment generated by the multi-channel partitioned power supply system and micro-area stirring device as input conditions for real-time physical process monitoring, ensuring comprehensive control of the spatial distribution conditions on the surface of the electroforming core mold.

[0199] S7.2 uses real-time time series filtering and spatial data synchronization algorithms to denoise and time-align the collected current density, temperature, and fluid flow rate feedback data, so that the obtained regional physical parameter feedback sequence meets high reliability standards, laying a data foundation for subsequent accurate comparison with the simulation values ​​of the multi-physics field coupling model.

[0200] S7.3 takes the denoised and calibrated regional physical parameter feedback sequence as input, and applies the model dynamic registration algorithm to correlate the spatial distribution simulation values ​​of the multi-physics field coupling model under the same coordinates and the regional performance evaluation indicators, thereby achieving high-precision spatial positioning pairing between the feedback data of each region and the simulation prediction.

[0201] S7.4 calculates the control deviation of each region based on the spatiotemporal differences between regional feedback data, model spatial distribution simulation values, and regional performance evaluation indicators, and forms parameter correction suggestions for the adaptive fuzzy control algorithm through a dynamic target vector generation mechanism, thereby realizing real-time dynamic adjustment of targets for spatial distribution control.

[0202] S7.5 inputs the dynamic target vector into the adaptive fuzzy control algorithm, combines the historical modulation parameters and equipment constraints, updates the local current amplitude adjustment parameters, pulse waveform setting parameters and micro-area stirring intensity adjustment parameters, and outputs a new round of spatial distribution process control instructions to achieve fine management of adaptive closed-loop electrodeposition process parameters for each area.

[0203] The step S8 specifically includes:

[0204] S8.1 Use multi-scale measurement instruments (such as three-dimensional morphology scanning, scanning electron microscope and regional performance testing system) to obtain the original spatial characterization data such as the spatial thickness distribution, surface roughness distribution and adhesion performance distribution of the deposition layer in each key area on the crown bionic adhesion array sample that has been completed by electrodeposition, so as to construct a multi-scale characterization measurement input data set.

[0205] S8.2 Based on the multi-scale characterization measurement input data set obtained in each region, a multi-scale feature extraction algorithm and spatial characterization data standardization processing are applied to quantify and standardize the spatial thickness distribution, surface roughness distribution and adhesion performance distribution data at different scales to obtain a set of regional characterization evaluation parameters with spatial comparability.

[0206] S8.3 Use spatial statistical analysis algorithms (such as local uniformity evaluation, spatial variation coefficient calculation, etc.) to perform data analysis on the standardized regional characterization evaluation parameter set to generate sedimentary layer spatial uniformity evaluation index parameters, including apparent consistency indicators between key areas and cross-scale hierarchical consistency indicators.

[0207] S8.4 uses the spatial uniformity evaluation index parameters as evaluation factors and inputs them into the self-learning optimization module. Based on the process-characterization data feedback mechanism, it adopts parameter attribution analysis technology and correction factor generation algorithm to associate the electroplating process parameter distribution data collected in the early stage to form a set of spatially resolved electroplating parameter distribution correction factors.

[0208] The S8.5 output spatially resolved electrodeposition parameter distribution correction factor serves as an important feedback variable in the automatic optimization closed loop, providing precise input for the adaptive distribution scheme of intelligent electrodeposition process parameters for subsequent preparation batches, thereby achieving continuous closed-loop iterative optimization of process parameter distribution and systematic improvement of the robustness of large-scale preparation performance.

[0209] The step S9 specifically includes:

[0210] S9.1 performs data analysis on the electrodeposition parameter distribution correction factor output by the self-learning optimization module to obtain a parameter group conversion plan that meets the spatial distribution characteristic requirements, ensuring that the correction factor can cover the spatial distribution control requirements of key areas.

[0211] S9.2 is based on the parameter group conversion scheme, performs standardized archiving and management of batch parameter group versions, and uses a version control system to implement hierarchical identification and traceable storage of different spatial distribution parameter rules, forming a batch parameter group version library to facilitate historical parameter call and comparative analysis.

[0212] For subsequent batch core mold preparation tasks, S9.3 extracts and distributes spatial distribution parameter rules from the batch parameter group version library according to the current optimal parameter group version. Through the regional parameter loading algorithm, the updated local current adjustment parameters, pulse waveform setting parameters, stirring intensity adjustment parameters, etc. are synchronously input into the multi-channel partitioned power supply equipment and micro-area stirring device to achieve precise initialization of spatial distribution parameters before batch preparation.

[0213] S9.4 In the new round of batch core mold electroforming preparation process, the spatial distribution parameters issued by the parameter loading algorithm are used to continuously collect and evaluate real-time process monitoring feedback data, analyze the improvement effect of spatial distribution consistency and preparation robustness after parameter iteration, and provide process data support for the closed-loop optimization of parameter correction factors.

[0214] Based on the spatial consistency evaluation and robustness analysis results after batch preparation, S9.5 models the correlation between the spatial distribution performance indicators of each region and the parameter correction factors of the previous round, optimizes and updates the self-learning adjustment factors, and realizes the adaptive iterative upgrade of the electrodeposition parameter distribution rules between subsequent batches, so as to ultimately achieve the optimal consistency of large-scale preparation and robustness of the process.

[0215] In the description of this application, it should be understood that the directions or positional relationships indicated by directional words such as "front, back, up, down, left, right", "horizontal, vertical, horizontal" and "top, bottom" are usually based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description. Unless otherwise specified, these directional words do not indicate or imply that the device or element referred to must have a specific direction or be constructed and operated in a specific direction. Therefore, they cannot be understood as limiting the scope of protection of this application.

[0216] Those skilled in the art can make various other corresponding changes and deformations based on the technical solutions and concepts described above, and all of these changes and deformations should fall within the scope of protection of the claims of this application.

Claims

1. A method for electrodeposition of a crown-shaped biomimetic adhesion array, comprising the following steps: S1: Micro current density sensors, temperature sensors, and fluid flow rate sensors are placed in each key area of ​​the crown-shaped end biomimetic adhesion array electroforming core mold structure and its corresponding electrolyte space to collect real-time current density, temperature, and fluid flow rate sampling data during the electrodeposition process in different areas; S2: De-noising and spatial consistency correction are performed on the collected current density, temperature and fluid flow rate sampling data to obtain a high-precision in-situ physical parameter data set for each sampling position; S3: Based on the high-precision in-situ physical parameter data set and combined with the core mold morphology information of each region, the spatial electric field distribution, temperature distribution and fluid distribution simulation values ​​of the electrolyte and the core mold surface in each region are calculated; S4: Calculate the spatial distribution characteristics of the sediment growth rate and surface uniformity in each region, and use surface uniformity as a regional performance evaluation indicator; S5: Based on the regional performance evaluation index, the adaptive fuzzy control algorithm is applied to generate the corresponding local current amplitude adjustment parameters, pulse waveform setting parameters and micro-area stirring intensity adjustment parameters for the spatial distribution differences of current density, temperature and fluid flow rate in each region; S6: Inputting the local current amplitude adjustment parameter, pulse waveform setting parameter and micro-area stirring intensity adjustment parameter into the multi-channel partition power supply device and the micro-area stirring device, performing targeted parameter modulation operation, and driving the electrodeposition equipment to perform regional process control according to the spatial distribution strategy; S7: Circularly collect the current density, temperature and fluid flow rate feedback data of each area, dynamically compare them with the aforementioned simulation values ​​and performance evaluation indicators, and modify the output parameters of the adaptive fuzzy control algorithm in real time based on the comparison results to achieve closed-loop spatial distribution process control.

2. The method for electrodeposition of a crown-shaped biomimetic adhesion array according to claim 1, characterized in that: After step S7, the following steps are also included: S8: Perform multi-scale characterization measurements on the crown-shaped biomimetic adhesion arrays in each region after deposition to evaluate the consistency and spatial uniformity of the deposited layer growth. Input the characterization results into the self-learning optimization module to generate a new generation of electrodeposition parameter distribution correction factors; S9: Based on the electrodeposition parameter distribution correction factor, execute the batch parameter group version switching strategy.

3. The method for electro-deposition of a crown-shaped biomimetic adhesion array according to claim 1, wherein: The step S2 specifically includes: A multi-channel time-series synchronous acquisition module is used to uniformly structure all the original process sampling data for the current density measurement signals, temperature measurement signals, and fluid flow rate measurement signals output by multi-point sensors in the core mold key area and electrolyte space, in order to obtain a process original sampling data set that is fully associated with the spatial area number and acquisition time. Based on the original sampling data set of the process, obtaining a process characteristic signal after spatial distribution noise reduction; Obtaining process-consistent physical feature data with spatial global coordination according to the de-noised process feature signal; Based on the process-consistent physical characteristic data, sensor bias correction and calibration factor compensation are performed on the measurement data of each area to eliminate the systematic drift between areas introduced by microsensor manufacturing process differences and assembly errors, thereby obtaining high-precision spatial distribution process parameters after calibration correction. Based on the calibrated and corrected high-precision spatially distributed process parameters, an index table of spatial parameter datasets is established through unified coding of process physical quantities and regional mapping.

4. The method for electrodeposition of a crown-shaped biomimetic adhesion array according to claim 1, wherein: The step S3 specifically includes: Perform regional data segmentation on the high-precision in-situ physical parameter dataset to classify the current density, temperature, and fluid flow rate data at each spatial sampling point into the corresponding electroforming core mold structure region and its matching electrolyte space, ensuring a one-to-one correspondence between each physical parameter sampling point and its spatial location information, thereby obtaining a regionalized physical parameter classification dataset. Based on the regionalized physical parameter classification dataset, the corresponding core model morphology information is retrieved for each spatial sampling area, including surface geometric features, dimensional parameters and crown structure microtopology. Data alignment and spatial registration algorithms are used to form a multidimensional input dataset containing the coupling characteristics of regional physical parameters and core model morphology.

5. The method for electrodeposition of a crown-shaped biomimetic adhesion array according to claim 4, characterized in that: After forming the multidimensional input data set containing the coupling characteristics of regional physical parameters and core model morphology in step S3, the following steps are further included: Taking a multi-dimensional input data set based on the coupling characteristics of regional physical parameters and core mold morphology as input, a spatial electric field distribution simulation is performed on each specified area to obtain the simulated value of the local electric field intensity distribution between the regional electrolyte and the core mold surface, and the output is a spatial electric field distribution prediction matrix; Using the spatial electric field distribution prediction matrix as the factor and combining the regional temperature parameter history, the spatial temperature distribution simulation value of each core mold structure area is calculated, and the spatial temperature distribution prediction matrix is ​​output. The spatial temperature distribution prediction matrix, current density parameters and fluid flow rate are used as input to simulate the fluid flow field distribution in each specified area, and the flow velocity vector field and flow state distribution simulation values ​​of the electrolyte near each micro-area of ​​the core mold are obtained to form a spatial fluid distribution prediction matrix.

6. The method for electro-deposition of a crown-shaped biomimetic adhesion array according to claim 1, wherein: The step S4 specifically includes: The high-precision in-situ physical parameter data set that has undergone spatial consistency correction is aligned with the spatial distribution simulation values ​​of the multi-physics field coupling numerical analysis model in the same area in time and space domains to form a regional distribution physical parameter comparison matrix; According to the regional distribution physical parameter control matrix, the sediment growth rate matrix containing the spatial coordinate resolution is output; Perform two-dimensional or three-dimensional spatial statistical analysis on the growth rate matrix of the deposited layer, extract the surface uniformity characteristics of each key area to form a uniformity index matrix that describes the surface consistency of the coating in each area; Based on the error analysis results between the simulated values ​​output by the spatial distribution simulation model and the real-time high-precision in-situ physical parameter data, the compensated and corrected regional sediment layer growth rate matrix and surface uniformity index matrix are output; Taking the compensated and corrected regional deposition layer growth rate matrix and surface uniformity index matrix as input, a comprehensive regional performance evaluation index reflecting the deposition consistency and spatial distribution uniformity of the coating in each key area is generated through normalization processing and multi-index weighted fusion algorithm.

Images