Electroplating production whole process monitoring system and method

Abstract

The invention discloses an electroplating production full-process monitoring system and method, and the method comprises the steps: monitoring a current density signal and plating solution temperature distribution data in an electroplating bath in real time, and generating a process coupling response value through density gradient analysis and hot spot migration track extraction; screening an abnormal process component by utilizing a process coupling response value, positioning a plating solution circulation transmission path and constructing a sensor network arrangement matrix; identifying a quality risk area through a big data processing technology, and extracting a coating thickness gradient field to determine a key monitoring node; adjusting equipment monitoring frequency according to the key monitoring nodes, forming a self-adaptive acquisition sequence and extracting a core monitoring set; performing energy consumption analysis by using the power change component, determining a monitoring period through load matching, and forming a monitoring convergence field; and a process control window is determined by using the monitoring convergence field and is converted into an intelligent monitoring instruction, so that comprehensive intelligent analysis and optimal management and control of production operation and maintenance of the electroplating bath are realized, and the quality stability and operation and maintenance efficiency of electroplating production are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent operation and maintenance of electroplating equipment, and in particular to an electroplating production full-process monitoring system and method. BACKGROUND

[0002] As a core production equipment in modern manufacturing industry, the running state and the accurate control of process parameters of the electroplating tank directly determine the product quality and production efficiency. With the development of intelligent manufacturing, the massive operation and maintenance data generated by the electroplating tank contains rich process optimization information and equipment health state indication. However, these data often present the characteristics of multi-source heterogeneity and space-time dispersion, and the traditional operation and maintenance management method cannot fully tap and utilize these valuable data resources.

[0003] The existing electroplating tank operation and maintenance technology generally has the problem of data island, and each monitoring system operates independently, lacking a unified data fusion and analysis platform. The traditional monitoring method mainly relies on single-point parameter monitoring and regular manual inspection, and cannot realize comprehensive perception and dynamic optimization of the electroplating process. Especially in dealing with complex process phenomena such as current density distribution, temperature field change and plating solution circulation, the existing technology lacks effective multi-dimensional correlation analysis means, resulting in inaccurate abnormal process component identification, delayed quality risk early warning, unreasonable monitoring resource allocation and other problems. Therefore, a method is needed to solve at least one of the above problems. SUMMARY

[0004] The present application discloses an electroplating production full-process monitoring system and method, which establishes an intelligent monitoring and optimization control system of the electroplating process through deep fusion analysis of current density signal and temperature distribution data, realizes full-process automatic management and continuous optimization from abnormal identification, risk assessment to intelligent control through collaborative processing of multi-dimensional data and self-adaptive optimization mechanism, and provides a complete technical solution for the digital transformation of the electroplating industry.

[0005] The present application discloses an electroplating production full-process monitoring method, comprising the following steps:

[0006] Monitoring the current density signal and the plating solution temperature distribution data in the electroplating tank, performing density gradient analysis on the current density signal to obtain current distribution characteristics, extracting hot spot migration trajectories in the plating solution temperature distribution data, and generating process coupling response values by thermoelectric correlation between the current distribution characteristics and the hot spot migration trajectories;

[0007] Screening abnormal process components by using the process coupling response values, positioning the plating solution circulation transmission path through the abnormal process components, detecting flow field distribution data along the transmission path according to the flow rate change, and constructing a sensor network arrangement matrix according to the flow field distribution data;

[0008] The sensor network layout matrix is ​​processed with big data to identify quality risk areas, the coating thickness gradient field of the quality risk areas is extracted, the coating thickness gradient field is used to determine the key monitoring locations, and the key monitoring locations are compared with the process coupling response values ​​to determine the critical monitoring nodes.

[0009] The monitoring frequency of the equipment is adjusted according to the key monitoring nodes, and an adaptive acquisition sequence is formed using the monitoring frequency of the equipment. Data aggregation and analysis are performed on the adaptive acquisition sequence to extract the core monitoring set, and the monitoring process configuration is constructed based on the core monitoring set.

[0010] Energy consumption distribution data is obtained by using the power change component in the current density signal for energy consumption analysis. The monitoring cycle is determined by load matching between the energy distribution data and the monitoring process configuration. Time-division monitoring is performed according to the monitoring cycle to form a monitoring convergence field.

[0011] The process control window is determined using the monitoring convergence field, and the monitoring process configuration is converted into intelligent monitoring instructions through the process control window.

[0012] A second aspect of this invention provides a monitoring system for the entire electroplating production process, comprising:

[0013] The signal acquisition module is used to monitor the current density signal and the temperature distribution data of the plating bath in the electroplating tank, perform density gradient analysis on the current density signal to obtain the current distribution characteristics, extract the hot spot migration trajectory in the temperature distribution data of the plating bath, and perform thermoelectric correlation between the current distribution characteristics and the hot spot migration trajectory to generate process coupling response values.

[0014] The transfer path module is used to filter abnormal process components using the process coupling response value, locate the plating solution circulation transfer path through the abnormal process components, detect flow rate changes along the transfer path to obtain flow field distribution data, and construct a sensor network layout matrix based on the flow field distribution data.

[0015] The risk identification module is used to perform big data processing on the sensor network layout matrix to identify quality risk areas, extract the coating thickness gradient field of the quality risk areas, use the coating thickness gradient field to determine the key monitoring locations, and compare the key monitoring locations with the process coupling response values ​​to determine the critical monitoring nodes.

[0016] The frequency adjustment module is used to adjust the monitoring frequency of the equipment according to the key monitoring nodes, use the monitoring frequency of the equipment to form an adaptive acquisition sequence, perform data aggregation analysis on the adaptive acquisition sequence to extract the core monitoring set, and construct the monitoring process configuration according to the core monitoring set.

[0017] The energy management module is used to perform energy consumption analysis using the power change component in the current density signal to obtain energy consumption distribution data, determine the monitoring cycle by matching the energy consumption distribution data with the monitoring process configuration, and perform time-division monitoring operations according to the monitoring cycle to form a monitoring convergence field.

[0018] The intelligent control module is used to determine the process control window using the monitoring convergence field, and to convert the monitoring process configuration into intelligent monitoring instructions through the process control window.

[0019] The beneficial effects of this invention are reflected in the following aspects: On the one hand, by coupling current density gradient analysis with temperature hotspot migration trajectory, the problem of inaccurate anomaly identification caused by isolated analysis of current and temperature data in traditional monitoring is solved. The process coupling response value can quantify the correlation strength between current distribution and temperature field, realizing the source location of abnormal process components and transforming fault detection from passive discovery to proactive early warning. On the other hand, by analyzing the correlation between sensor network layout matrix and quality risk areas, the problems of blind monitoring point layout and unreasonable monitoring resource allocation are solved. The coating thickness gradient field can accurately identify weak links in quality, and the determination of key monitoring nodes avoids the waste of resources in comprehensive monitoring, achieving optimal matching between monitoring input and monitoring effect. Furthermore, by matching energy consumption distribution data with the load of the monitoring process, the problem that fixed monitoring cycles cannot adapt to process changes is solved. Time-sharing monitoring operation dynamically adjusts the monitoring density according to actual energy consumption needs, the automatic determination of the process control window eliminates the arbitrariness of manual setting, and intelligent monitoring commands realize closed-loop management from data analysis to control execution.

[0020] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0021] The accompanying drawings illustrate specific examples of the technical solutions described in this invention and, together with the detailed embodiments, form part of the specification, serving to explain the technical solutions, principles, and effects of this invention.

[0022] Unless otherwise specified or otherwise, the same reference numerals in different figures represent the same or similar technical features, and different reference numerals may be used to represent the same or similar technical features.

[0023] Figure 1 This is a flowchart illustrating a method for monitoring the entire electroplating production process according to the present invention.

[0024] Figure 2 This is a structural block diagram of a full-process monitoring system for electroplating production according to the present invention. Detailed Implementation

[0025] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0026] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0027] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0028] The technical solutions of the embodiments of this application will be described below.

[0029] like Figure 1 As shown, this embodiment of the invention provides a method for monitoring the entire electroplating production process, including the following steps S110-S160:

[0030] Step S110: Monitor the current density signal and the temperature distribution data of the plating bath in the electroplating tank, perform density gradient analysis on the current density signal to obtain the current distribution characteristics, extract the hot spot migration trajectory in the plating bath temperature distribution data, and perform thermoelectric correlation between the current distribution characteristics and the hot spot migration trajectory to generate the process coupling response value.

[0031] Specifically, high-precision current density sensors and temperature sensor arrays are deployed at key locations in the electroplating tank to collect current density signals and plating solution temperature distribution data in real time during the electroplating process. The current density sensors employ Hall effect probes, installed at multiple measuring points on the anode surface, covering a measurement range of 0-200 A / dm² with an accuracy of ±0.5 A / dm², and a sampling frequency of 50 Hz. The sensor layout uses a uniform grid distribution, with 25 measuring points on the anode surface, covering the entire effective electroplating area. The temperature sensors combine thermocouples and an infrared thermal imager. The thermocouples measure the temperature of the plating solution itself, while the infrared thermal imager monitors the temperature distribution on the anode surface. Thermocouples are arranged in three layers (upper, middle, and lower) of the plating solution, with 9 measuring points in each layer, achieving a measurement accuracy of ±0.1℃. The infrared thermal imager is installed above the electroplating tank, enabling non-contact continuous monitoring of the anode surface temperature, with a temperature resolution of 0.1℃ and an imaging frequency of 25 Hz. The data acquisition system adopts a distributed architecture, with each sensor connected to the data acquisition unit via a fieldbus, achieving multi-point synchronous acquisition and real-time data transmission. Establish a data preprocessing mechanism, including signal filtering, outlier removal, and data standardization, to ensure the accuracy and consistency of current density signals and temperature distribution data.

[0032] In some embodiments, the step of performing density gradient analysis on the current density signal to obtain current distribution characteristics includes: performing phase demodulation separation on the current density signal to obtain phase feature components; performing spatial mapping processing on the phase feature components to form a current density cloud map; performing gradient vector tracing on the current density cloud map to generate gradient streamlines; and performing feature encoding and summarization on the gradient streamlines to obtain current distribution characteristics.

[0033] Phase demodulation is performed on the current density signal to separate its phase characteristic components. The current density signal contains DC and AC components; the DC component reflects the average current level, while the AC component contains current fluctuations and pulsations. A phase demodulation algorithm is established, using the Hilbert transform to calculate the instantaneous phase φ(t) = arctan[H{I(t)} / I(t)], where φ(t) is the instantaneous phase, H{I(t)} is the Hilbert transform of the current signal I(t), and t is time. Phase demodulation separates the phase characteristic components, including phase jumps, phase drifts, and periodic phase changes. Phase jumps reflect abrupt changes in current density, typically occurring at anode edges and geometric abrupt changes. Phase drift reflects the gradual change trend of current density, related to electrolyte concentration gradients and temperature distribution. Periodic phase changes correspond to the effects of current pulsations and switching frequency. Analyze the phase characteristic components at each measuring point and calculate the phase difference Δφ = φᵢ - φⱼ, where φᵢ and φⱼ are the phase values ​​at different measuring points, and the phase difference reflects the relative relationship of current density between measuring points. Establish a spatial distribution matrix of the phase characteristic components, with rows and columns corresponding to the spatial coordinates of the sensor, and matrix elements representing the phase characteristic values ​​at the corresponding locations.

[0034] A current density cloud map is generated by spatially mapping the phase feature components. Based on the acquired phase feature components, a two-dimensional distribution cloud map of the current density within the electroplating tank is constructed using spatial interpolation and mapping techniques. The spatial distribution of the phase feature components is processed using a radial function interpolation method. The interpolation function adopts the Gaussian function form f(r) = exp(-r² / 2σ²), where f(r) is the radial function value, r is the distance between measurement points, and σ is the interpolation parameter. The smoothness and detail of the cloud map are controlled by adjusting the interpolation parameter. A color mapping scheme for the current density cloud map is established, mapping the numerical range of the phase feature components to a color space, using warm and cool hues to represent the high and low current density distribution. The color depth in the cloud map intuitively reflects the spatial variation of the current density, with red areas representing high current density and blue areas representing low current density. The distribution pattern of the current density cloud map is analyzed to identify high-density concentrated areas, low-density sparse areas, and areas with drastic gradient changes. The central area of ​​the electroplating tank usually exhibits a relatively uniform current distribution, while the edge areas show current concentration due to boundary effects. Establish quantitative descriptive parameters for the cloud map, including features such as distribution uniformity index, peak location, and gradient strength.

[0035] Gradient streamline bundles are generated by gradient vector tracing in the current density contour map. The current density gradient vector ∇I = (∂I / ∂x, ∂I / ∂y) represents the direction and magnitude of the fastest change in current density, where ∇I is the gradient vector, I is the current density, and x and y are the spatial coordinates. The gradient vector at each point in the contour map is calculated using numerical differentiation, and the central difference scheme is employed to improve computational accuracy. A gradient vector tracing algorithm is established, selecting a starting point from the current density contour map and tracing streamlines along the gradient direction. The fourth-order Runge-Kutta method is used for streamline tracing to ensure the accuracy and stability of the tracing path. Gradient streamline bundles are generated, with each streamline representing a path of current density change, and the streamline density reflecting the strength of the gradient change. Streamlines are denser in high-gradient regions and sparser in low-gradient regions. The geometric characteristics of the gradient streamline bundles are analyzed, including parameters such as streamline length, curvature, convergence, and divergence. Convergent streamline bundles indicate that current is concentrated in a certain region, while divergent streamline bundles indicate that current is diffused outward from a certain region.

[0036] The gradient streamline bundle is feature-encoded and summarized to obtain current distribution characteristics. Based on the generated gradient streamline bundle, key feature information of the current distribution is summarized using feature extraction and encoding techniques. Statistical characteristics of the streamline bundle, such as average length, length variance, principal orientation angle, and spatial distribution entropy, are calculated. Principal component analysis is used to reduce the dimensionality of the streamline bundle features and extract the most important feature components. A feature encoding algorithm is established to convert the multidimensional streamline bundle features into standardized feature vectors. Each component of the feature vector corresponds to different current distribution characteristics, such as distribution uniformity, concentration, directionality, and degree of change. A feature weight allocation strategy is designed, and the weight coefficients are determined according to the degree of influence of each feature on the electroplating quality. A comprehensive current distribution feature index F = Σᵢwᵢfᵢ is formed through weighted combination, where F is the current distribution feature index, wᵢ is the weight of the i-th feature, and fᵢ is the standardized value of the i-th feature. The obtained current distribution features contain complete information such as the spatial distribution pattern, trend, and abnormal regions of the current density.

[0037] Extract hotspot migration trajectories from plating bath temperature distribution data. Hotspots are localized high-temperature areas with temperatures significantly higher than the surrounding area, typically caused by factors such as concentrated heat from electrochemical reactions, resistance losses, and poor local heat transfer. A hotspot identification algorithm is established, determining hotspot locations through temperature threshold judgment and neighborhood comparison. Hotspot identification criteria are set, classifying a point as a hotspot when its temperature exceeds a specific multiple of the neighborhood average temperature. A moving window technique is used to scan and identify hotspots in the temperature distribution data; the window size is determined based on the plating bath size and sensor distribution. A hotspot tracking mechanism is established, tracking the spatial movement trajectory of hotspots through time series analysis. The spatial distance and movement speed of hotspot positions at adjacent times are calculated as v = Δr / Δt, where v is the hotspot movement speed, Δr is the position change, and Δt is the time interval. The driving mechanisms of hotspot migration are analyzed, including factors such as convective heat transfer, plating bath circulation, and changes in current distribution. Hotspot migration trajectories typically exhibit periodic or quasi-periodic characteristics, related to stirrer rotation and plating bath circulation patterns. Trajectory feature parameters are established, including geometric features such as migration distance, migration speed, migration direction, and trajectory shape.

[0038] The process coupling response value is generated by thermoelectrically correlating current distribution characteristics with hotspot migration trajectories. A thermoelectric correlation analysis algorithm is established to calculate the temporal correlation between current distribution characteristics and hotspot migration trajectories through time series analysis, determining the response time and intensity of current changes on hotspot migration. Spatial correlation analysis is performed to calculate the spatial overlap and proximity between high-current-density regions and hotspot locations, identifying the correspondence between current-concentrated regions and temperature hotspots. Intensity correlation characteristics are analyzed, comparing the correspondence between the magnitude of current density changes and the magnitude of hotspot temperature changes, establishing a correlation pattern between current intensity and temperature intensity. Based on the analysis results from the three dimensions of temporal, spatial, and intensity correlation, a method for calculating the process coupling response value is established. The process coupling response value is formed through a weighted combination, reflecting the tightness of coupling and mutual influence between the current distribution and the temperature field. A high coupling response value indicates a strong interaction between the current and temperature field, requiring coordinated control; a low coupling response value indicates that the two are relatively independent and can be optimized separately.

[0039] Step S120: Use the process coupling response value to screen abnormal process components, locate the plating solution circulation path through the abnormal process components, detect the flow rate change along the transmission path to obtain flow field distribution data, and construct a sensor network layout matrix based on the flow field distribution data.

[0040] Specifically, abnormal process components are screened using process coupling response values. Process coupling response values ​​reflect the strength of the interaction between current distribution and temperature field. When the coupling response value deviates significantly from the normal range, it indicates an abnormal process phenomenon. Anomaly judgment criteria are established by comparing the process coupling response value with historical normal values ​​and setting an abnormal threshold range. When the coupling response value exceeds the upper or lower limit of the normal range, it is determined to be an abnormal process component. Statistical analysis methods are used to calculate the mean and standard deviation of the process coupling response values, establishing a normal distribution interval. Data points deviating from three times the standard deviation are identified as abnormal components. The characteristics of abnormal process components are analyzed, including different types such as abnormal current density concentration, abnormal migration of temperature hotspots, and abnormal enhancement of thermoelectric coupling. Abnormal current density concentration is manifested as a local area current density much higher than the average level, usually caused by anode surface inhomogeneity or local changes in plating solution composition. Abnormal migration of temperature hotspots is manifested as irregular movement or stagnation of hotspot positions, possibly related to abnormal plating solution circulation or heat transfer obstacles. Abnormal enhancement of thermoelectric coupling is manifested as an abnormally increased correlation between current changes and temperature changes, reflecting an unstable state of the process. Establish a quantitative description of abnormal process components, recording key information such as the time, location, intensity, and duration of the abnormality. Design a classification system for abnormal components, categorizing abnormalities into three levels: minor, moderate, and severe, based on the degree of deviation.

[0041] In some embodiments, locating the plating solution circulation path by means of the abnormal process component includes: performing reverse propagation tracing on the abnormal process component to obtain the propagation source coordinates; establishing an energy diffusion field based on the propagation source coordinates; performing path aggregation analysis in the energy diffusion field to generate aggregated path bundles; and extracting the main path from the aggregated path bundles to determine the plating solution circulation path.

[0042] The anomalous process components are backpropagated to trace and obtain the coordinates of their propagation source. Anomalous process components are usually not isolated but originate from a source location and propagate to surrounding areas through plating solution circulation and diffusion mechanisms. A backpropagation tracing algorithm is established to analyze possible propagation paths and source locations backward from the observed position of the anomalous process component. A gradient back-tracking method is used to search for paths along the increasing direction of the anomalous intensity. The anomalous intensity gradient ∇A = (∂A / ∂x, ∂A / ∂y) is calculated, where A is the anomalous intensity, x and y are spatial coordinates, and ∇A is the anomalous intensity gradient vector. Back-integration is performed along the gradient direction to trace the location of the maximum anomalous intensity. Particle back-tracking technology is used to release virtual particles at the location of the anomalous process component, allowing the particles to move in the opposite direction of the flow field, and recording the particle trajectories. The particle back-motion equation is dx / dt = -v(x,t), where x is the particle position, v is the velocity vector, and the negative sign indicates back-motion. By tracking the backward trajectories of multiple particles, the convergence point of the trajectories is identified as a candidate propagation source. A confidence assessment of the source coordinates is established, and the reliability of the source location is determined based on the consistency and convergence of the backward trajectories. The three-dimensional position information (x0, y0, z0) of the propagation source coordinates is recorded, including its precise spatial location within the electroplating tank.

[0043] An energy diffusion field is established based on the coordinates of the propagation source. The propagation source serves as the starting point of the anomalous energy; the anomalous influence spreads to the surrounding area through a diffusion mechanism, forming an energy field with certain spatial distribution characteristics. A mathematical description of the energy diffusion field is established using the diffusion equation ∂E / ∂t=D∇²E+S, where E is the energy density, t is time, D is the diffusion coefficient, ∇² is the Laplace operator, and S is the source term. The coordinates of the propagation source are treated as a point source, and an energy injection boundary condition E(x0,y0,z0)=E0 is set at the source location, where E0 is the initial energy density at the source. The influence of the physical properties of the plating solution on the diffusion process is considered, including the effects of parameters such as density, viscosity, conductivity, and temperature on the diffusion coefficient. The anisotropic characteristics of energy diffusion are analyzed; due to the circulation and stirring of the plating solution, the diffusion process exhibits different diffusion rates in different directions. A numerical solution method for the diffusion field is established, using finite difference or finite element techniques to solve the diffusion equation and obtain the spatial distribution of energy density throughout the electroplating tank. Analyze the contour distribution of the energy diffusion field; the shape and density of the contour lines reflect the directionality and intensity changes of diffusion. Identify high-gradient and low-gradient regions in the energy diffusion field; high-gradient regions correspond to the main direction of the diffusion path, while low-gradient regions indicate areas where the diffusion effect is weaker.

[0044] Path aggregation analysis is performed in the energy diffusion field to generate aggregated path bundles. The energy diffusion field contains multiple possible propagation paths, and aggregation analysis is needed to group similar paths into representative path bundles. A path identification algorithm is established to trace propagation paths along the energy gradient direction in the energy diffusion field. Streamline tracing technology is used to generate streamlines starting from the propagation source and along the main direction of energy propagation. The streamline equation is dx / ds=∇E / |∇E|, where x is the position vector, s is the arc length parameter, and ∇E is the energy gradient. Through multi-startpoint streamline tracing, multiple starting points are set around the source to generate a family of streamlines covering the entire diffusion range. The geometric characteristics of the streamlines are analyzed, including parameters such as streamline length, curvature, orientation angle, and spatial distribution. Cluster analysis is used to group streamlines, grouping streamlines with similar geometric characteristics into the same category. The k-means clustering algorithm is used, with the geometric feature vector of the streamline as the clustering input, to divide the streamlines into several aggregated path bundles. Each aggregated path bundle represents a group of similar propagation paths, reflecting the main propagation pattern of energy in a specific direction. Calculate the centerline of the converged path bundle as its representative path. Analyze the statistical characteristics of the converged path bundle, including parameters such as the number of streamlines within the bundle, average length, directional consistency, and spatial coverage.

[0045] The main path of the polymerization path bundle is extracted to determine the plating solution circulation and transfer path. A main path evaluation criterion is established, comprehensively considering factors such as energy propagation intensity, spatial coverage, propagation speed, and stability. The importance weight of each polymerization path bundle is calculated as W = α·E_intensity + β·L_coverage + γ·V_speed, where W is the path weight, E_intensity is the energy propagation intensity, L_coverage is the spatial coverage length, V_speed is the propagation speed, and α, β, and γ are weight coefficients. The most important polymerization path bundles are selected as main path candidates based on their weight ranking. The spatial distribution pattern of the main path candidates is analyzed to identify the main circulation direction and propagation trend. Considering the physical constraints of the plating solution circulation system, including the location of the circulation pump, pipeline layout, and the influence range of the agitator, the physical rationality of the main path is verified. An optimization algorithm for the main path is established, extracting a concise and clear plating solution circulation and transfer path through path smoothing and simplification. The main path should accurately reflect the main propagation direction and circulation pattern of abnormal process components in the plating solution. Record the geometric description of the defined plating solution circulation path, including information such as the path start point, key nodes, end point, and total path length.

[0046] Flow field distribution data is obtained by detecting velocity changes along the transfer path. The circulation path of the plating solution reflects the main propagation direction of abnormal process components, and detecting velocity changes along the path can obtain accurate flow field distribution information. A velocity detection scheme is designed, using a combination of ultrasonic flowmeters and particle image velocimetry (PEV) technology. The ultrasonic flowmeter is installed at key nodes of the transfer path to measure the magnitude and direction of the velocity at the main cross-sections of the path. The PEV system measures the detailed flow field distribution near the path by projecting laser sheet light and using high-speed imaging technology. Tracer particles are added to the plating solution, with particle diameters selected to be on the order of micrometers and a density close to that of the plating solution, to ensure that the particles can accurately follow the liquid movement. A velocity measurement grid is established, and measurement cross-sections are set at certain intervals along the transfer path, with multiple points of velocity vector measured at each cross-section. The spatiotemporal characteristics of velocity changes are analyzed, and the magnitude, direction, and pulsation characteristics of the velocity are recorded. The velocity gradient ∂v / ∂n is calculated, where v is the velocity and n is the normal coordinate, and the gradient reflects the rate of change of velocity in the direction perpendicular to the path. Identify anomalous regions in the flow field distribution, including phenomena such as abnormally increased flow velocity, abrupt changes in flow direction, and eddy formation. Establish an organizational structure for the flow field distribution data, including complete information such as spatial coordinates, velocity vectors, pressure distribution, and turbulence intensity.

[0047] A sensor network layout matrix is ​​constructed based on flow field distribution data. A sensor layout optimization algorithm is established, with the optimization objectives of maximizing flow field monitoring accuracy and minimizing the number of sensors. Key monitoring areas in the flow field distribution data are analyzed, including high velocity gradient regions, regions with drastic flow direction changes, and regions with active eddies. Sensor density is increased in these key regions to ensure accurate capture of important flow field information. The topology of the sensor network is designed, establishing communication connections and data transmission paths between sensor nodes. A hierarchical layout strategy is adopted, dividing sensors into two levels: primary monitoring nodes and auxiliary monitoring nodes. Primary monitoring nodes are placed at locations with the most drastic flow field changes, providing high-precision continuous monitoring. Auxiliary monitoring nodes are distributed in relatively stable flow field regions, providing an overview of the global flow field. A sensor network layout matrix M is established, where the rows and columns correspond to the spatial coordinate grid of the electroplating tank, and matrix elements represent the sensor type and configuration parameters at the corresponding location. A matrix element of 1 indicates that a sensor is placed at that location, and 0 indicates that no sensor is placed.

[0048] Step S130: Perform big data processing on the sensor network layout matrix to identify quality risk areas, extract the coating thickness gradient field of the quality risk areas, use the coating thickness gradient field to determine the key monitoring locations, and compare the key monitoring locations with the process coupling response values ​​to determine the critical monitoring nodes.

[0049] In some embodiments, the step of performing big data processing on the sensor network layout matrix to identify quality risk areas includes: triggering electric field inductive coupling through the sensor network layout matrix to generate a collaborative monitoring field; capturing quality anomaly cluster centers in the collaborative monitoring field; performing risk density analysis on the quality anomaly cluster centers to form a risk density distribution; and identifying quality risk areas based on the risk density distribution.

[0050] For example, the step of generating a collaborative monitoring field by triggering electric field inductive coupling through the sensor network arrangement matrix includes: generating a coupling signal by triggering electric field inductive coupling using the sensor network arrangement matrix; extracting synchronization features by performing Faraday wave analysis on the coupling signal; mapping the synchronization features to a monitoring sensitivity distribution; and generating a collaborative monitoring field by gradient fusion of the sensitivity distribution.

[0051] A sensor network layout matrix is ​​used to trigger electric field induction coupling to generate coupling signals. The sensor network layout matrix defines the spatial position and electrical connection relationship of each sensor node, creating the physical conditions for electric field induction coupling. An excitation mechanism for electric field induction coupling is established: when a sensor detects a change in current within the electroplating tank, a changing electromagnetic field is generated around the sensor. Adjacent sensors are affected by the electromagnetic field, generating induced electromotive force, forming an electric field coupling effect between the sensors. The electric field induction intensity E = -∂A / ∂t is calculated, where E is the induced electric field intensity, A is the vector magnetic potential, and t is time. The relationship between coupling distance and coupling intensity between sensors is analyzed: sensors closer together generate stronger coupling signals, while sensors farther apart have weaker coupling effects. Frequency characteristic analysis of the coupling signals is established; current changes at different frequencies generate coupling signals of different frequencies. A circuit for acquiring and processing the coupling signals is designed, capturing weak coupling signals through a high-precision analog front-end circuit. The mutual coupling effect between multiple sensors is analyzed; when multiple sensors operate simultaneously, complex multi-source coupling signals are generated. Establish a description of the amplitude and phase characteristics of the coupled signal. The amplitude of the coupled signal reflects the coupling strength, and the phase reflects the timing relationship of the coupling.

[0052] Synchronization features are extracted from the coupled signals using Faraday wave analysis. Based on the generated coupled signals, Faraday wave analysis is employed to extract synchronization feature components. Faraday waves reflect potential fluctuations caused by changes in the electromagnetic field, containing important information about the coordinated response between sensors. A mathematical description of Faraday waves is established, based on Faraday's law of electromagnetic induction ε = -dΦ / dt, where ε is the induced electromotive force, Φ is the magnetic flux, and t is time. The wave components in the coupled signals are analyzed, including different types such as periodic, random, and transient waves. Spectral analysis is used to decompose the frequency components of the coupled signals, identifying the main wave frequencies and minor harmonic components. A synchronization feature extraction algorithm is established, and correlation analysis is used to identify the synchronization response characteristics between multiple sensors. The cross-correlation function of the coupled signals between sensors is calculated to determine the time delay and correlation strength between signals. The time-domain and frequency-domain characteristics of the synchronization features are analyzed; time-domain synchronization is reflected in the similarity of signal waveforms, while frequency-domain synchronization is reflected in the consistency of the spectrum. A quantitative index for the degree of synchronization is established, using correlation coefficients and phase differences to describe the synchronization level between sensors. Identify key factors affecting synchronization characteristics, including parameters such as sensor location, electroplating current distribution, and plating solution conductivity.

[0053] Synchronization characteristics are mapped to a monitoring sensitivity distribution. Synchronization characteristics reflect the coordinated response capability between sensors; high synchronization corresponds to high monitoring sensitivity, and low synchronization corresponds to low monitoring sensitivity. A mapping algorithm is established to convert the quantitative indicators of synchronization characteristics into sensitivity values. The sensitivity calculation formula is designed as S=f(C,Φ), where S is the monitoring sensitivity, C is the correlation coefficient, Φ is the phase difference, and f is the mapping function. The selection principle of the mapping function is analyzed: linear mapping is suitable for cases where synchronization characteristics and sensitivity are directly proportional, while nonlinear mapping is suitable for complex mapping relationships. A spatial interpolation method for the sensitivity distribution is established to extend the sensitivity values ​​of discrete sensor locations into a continuous spatial distribution. Radial function interpolation technology is used to process the spatial mapping of sensitivity, ensuring the smoothness and physical rationality of the interpolation results. The spatial characteristics of the sensitivity distribution are analyzed to identify high-sensitivity, medium-sensitivity, and low-sensitivity regions. High-sensitivity regions are usually located in areas with dense sensor density and good synchronization, capable of sensitively capturing minute anomalies.

[0054] A collaborative monitoring field is generated through gradient fusion of sensitivity distributions. Sensitivity distribution describes the monitoring capability of each point in space, and gradient fusion integrates monitoring information from different regions to form a unified collaborative monitoring field. A gradient calculation algorithm is established to calculate the spatial gradient of the sensitivity distribution ∇S=(∂S / ∂x,∂S / ∂y), where S is the sensitivity and x and y are the spatial coordinates. The physical meaning of the sensitivity gradient is analyzed: the gradient magnitude reflects the drastic change in sensitivity, and the gradient direction indicates the direction of the fastest increase in sensitivity. A gradient fusion strategy is designed to weight and fuse the sensitivity gradients at different locations to form a comprehensive monitoring field distribution. The fusion weights are calculated, and the fusion contribution of each location is determined based on the magnitude of the sensitivity and the strength of the gradient. A fusion algorithm F=Σᵢwᵢ∇Sᵢ is established, where F is the fused collaborative monitoring field, wᵢ is the fusion weight at the i-th location, and ∇Sᵢ is the sensitivity gradient at the i-th location. The spatial distribution characteristics of the collaborative monitoring field are analyzed to identify advantageous areas with strong monitoring capabilities and blind spots with weak monitoring capabilities. Establish a continuity analysis of the collaborative monitoring field to ensure a smooth spatial transition and logical coherence of the monitoring field.

[0055] This study identifies quality anomaly cluster centers within a collaborative monitoring field. The collaborative monitoring field provides a continuous spatial distribution of monitoring signals within the electroplating tank, with anomaly cluster centers corresponding to areas of concentrated abnormal signals. Identification criteria for anomaly cluster centers are established, determining them through a dual assessment of monitoring signal strength and anomaly severity. An anomaly signal threshold is set; when the signal strength in a region of the monitoring field exceeds the normal range, it is marked as an anomaly area. A clustering analysis algorithm is used to group anomaly areas, merging spatially adjacent anomaly points into anomaly clusters. The centroid position of each anomaly cluster is calculated as the coordinates of the cluster center. The intensity characteristics of the anomaly cluster centers are analyzed, including parameters such as the peak value, average value, and variability of the abnormal signals. An impact range assessment for cluster centers is established, determining the radius of influence and intensity distribution of each cluster center. A dynamic tracking mechanism for cluster centers is designed to monitor the temporal evolution and movement trajectory of their locations. A hierarchical system for cluster centers is established, classifying them into primary and secondary cluster centers based on anomaly intensity and impact range.

[0056] Risk density analysis is performed on the cluster centers of quality anomalies to form a risk density distribution. A risk density calculation method is established, taking the cluster centers as risk sources and calculating the risk density distribution in the surrounding space. Kernel density estimation is employed, using a Gaussian kernel function to calculate the risk density distribution. The risk propagation characteristics of different types of cluster centers are analyzed; current anomaly cluster centers mainly affect local areas, while temperature anomaly cluster centers have a larger diffusion range. The attenuation law of risk density is established, describing the attenuation characteristics of risk density with distance. A risk density superposition algorithm for multiple cluster centers is designed; when the influence ranges of multiple cluster centers overlap, the total risk density after superposition is calculated. The spatial characteristics of the risk density distribution are analyzed to identify high-risk density areas, medium-risk density areas, and low-risk density areas. Statistical parameters of the risk density distribution are calculated, including indicators such as maximum risk density, average risk density, and risk density variance.

[0057] Quality risk areas are identified based on risk density distribution. Risk area identification criteria are established, and risk density thresholds are set to divide continuous density distributions into discrete risk areas. A multi-threshold segmentation method is used to divide the risk density distribution into three levels: high-risk, medium-risk, and low-risk areas. A region growing algorithm is used to perform connectivity analysis on adjacent pixels of the same risk level to form continuous risk areas. The geometric characteristics of each risk area are calculated, including parameters such as area, perimeter, shape factor, and location center. The distribution patterns of risk areas are analyzed to identify their clustering characteristics and spatial correlations. A dynamic monitoring mechanism for risk areas is established to track their changing trends and evolution patterns over time. An alarm mechanism for risk areas is designed to trigger an alarm when the area of ​​a high-risk area exceeds a threshold or a new high-risk area appears. A priority ranking of risk areas is established, and processing priorities are determined based on risk level, impact range, and urgency.

[0058] Extract the coating thickness gradient field from quality risk areas. Quality risk areas are often accompanied by non-uniform distribution of coating thickness, and the thickness gradient field can reveal the spatial characteristics and severity of thickness variations. Establish a thickness measurement network, deploy high-precision thickness sensors within the quality risk areas, and employ a combination of eddy current and ultrasonic thickness measurement methods. Design a thickness data acquisition scheme, performing full-coverage measurements of the risk areas according to a regular grid to ensure the spatial continuity of thickness data. Establish a thickness gradient calculation algorithm, using numerical differentiation to calculate the spatial gradient of the thickness field ∇h=(∂h / ∂x,∂h / ∂y), where h is the coating thickness and x and y are spatial coordinates. Analyze the magnitude and direction characteristics of the thickness gradient; the gradient magnitude reflects the drasticness of thickness changes, and the gradient direction indicates the direction of the fastest thickness increase. Identify anomalous regions in the thickness gradient field, including regions with concentrated high gradients, regions with abrupt changes in gradient direction, and regions with zero gradient points. Analyze the correlation between thickness gradient and quality risk, establishing a mapping relationship between gradient intensity and the severity of quality problems.

[0059] The coating thickness gradient field is used to determine key monitoring locations. Selection criteria for key monitoring locations are established, prioritizing locations with large thickness gradients, drastic gradient changes, and significant impacts on quality. Statistical characteristics of the thickness gradient field are calculated, including gradient mean, variance, maximum value, and distribution skewness. A gradient threshold is set, and locations with gradients exceeding the threshold are marked as candidate monitoring locations. A spatial clustering algorithm is used to group candidate locations to avoid over-concentration of monitoring points in a single local area. An importance evaluation index for monitoring locations is established, comprehensively considering factors such as gradient strength, location representativeness, and monitoring operability. An optimization algorithm for monitoring locations is designed to minimize the number of monitoring points while meeting monitoring accuracy requirements. The spatial distribution characteristics of key monitoring locations are analyzed to ensure that these locations cover the main quality problem areas.

[0060] Key monitoring nodes are identified by comparing the monitored locations with the process coupling response values. A comparison analysis algorithm is established to spatially match the spatial coordinates of the monitored locations with the distribution of process coupling response values. The process coupling response value corresponding to each monitored location is calculated, establishing a correspondence between location and coupling value. Key evaluation indicators are designed, comprehensively considering both thickness gradient strength and process coupling response value. Monitoring locations are ranked according to key indicators, and the location with the highest indicator value is selected as the key monitoring node. The rationality of the distribution of key monitoring nodes is analyzed to ensure that the nodes can cover the main quality risk areas and process coupling anomaly areas. A configuration scheme for key monitoring nodes is established, including settings for node location, monitoring parameters, sampling frequency, and alarm thresholds. The communication topology of the node network is designed to ensure data sharing and collaborative operation among key monitoring nodes.

[0061] Step S140: Adjust the equipment monitoring frequency according to the key monitoring nodes, use the equipment monitoring frequency to form an adaptive acquisition sequence, perform data aggregation analysis on the adaptive acquisition sequence to extract the core monitoring set, and construct the monitoring process configuration based on the core monitoring set.

[0062] In some embodiments, adjusting the equipment monitoring frequency based on the key monitoring nodes includes: performing a monitoring load assessment on the key monitoring nodes to generate a load distribution spectrum; mining monitoring resource reserves from the load distribution spectrum to construct a reserve resource pool; performing frequency matching between the load distribution spectrum and the reserve resource pool to obtain an adjustment coefficient; and performing frequency adjustment based on the adjustment coefficient to determine the equipment monitoring frequency.

[0063] For example, the step of generating a load distribution spectrum by assessing the monitoring load of the key monitoring nodes includes: performing cross-time period load tracking analysis on the key monitoring nodes to obtain load change data; using the load change data to analyze the load contribution effect of each monitoring element to obtain load intensity assessment results, wherein each monitoring element includes current density change, plating solution temperature fluctuation and equipment operating status; and performing spectrum conversion processing on the load intensity assessment results to generate a load distribution spectrum.

[0064] Cross-period load tracking analysis is performed on key monitoring nodes to obtain load change data. A time window design for load tracking is established, dividing the monitoring period into multiple time periods. Each time period records load indicators such as the number of monitoring tasks, data processing volume, and computational resource consumption of each node. A load quantification method is designed to convert abstract monitoring tasks into calculable load values ​​L = αN + βD + γC, where L is the total load, N is the number of monitoring tasks, D is the data processing volume, C is the computational complexity, and α, β, and γ are weighting coefficients. Load characteristics in different time periods are analyzed to identify peak load periods, trough load periods, and stable load periods. A load change data acquisition mechanism is established to record the load status and trends of each key monitoring node in real time. The periodic characteristics of load changes are analyzed to identify regular change patterns such as daily and weekly cycles. A load data storage structure is designed to establish a time-series database to store a large amount of historical load change data. The suddenness of load changes is analyzed to identify sudden increases or decreases in load caused by abnormal events.

[0065] Based on the acquired load change data, this study analyzes the contribution of various monitoring elements, such as current density changes, plating solution temperature fluctuations, and equipment operating status, to the total system load. The impact of current density changes on the monitoring load is analyzed; drastic current density changes require high-frequency monitoring and complex data processing, leading to an increased monitoring load. A load contribution calculation for current density changes is established: L_current = k1 × |dI / dt|, where L_current is the load contribution of current density changes, k1 is the contribution coefficient, and |dI / dt| is the rate of change of current density. The contribution of plating solution temperature fluctuations to the monitoring load is analyzed; temperature fluctuations require real-time monitoring of the temperature field and heat conduction analysis, increasing the system's computational load. A load contribution calculation for temperature fluctuations is established: L_temp = k2 × σ_T, where L_temp is the load contribution of temperature fluctuations, k2 is the contribution coefficient, and σ_T is the temperature standard deviation. The impact of equipment operating status on the monitoring load is analyzed; changes in equipment status require additional status monitoring and fault diagnosis. The load contribution of equipment status is calculated as L_equipment = k3 × N_state, where L_equipment is the load contribution of the equipment status, k3 is the contribution coefficient, and N_state is the number of status changes. Combining the contribution effects of each monitoring element, the total load intensity assessment is established as L_total = L_current + L_temp + L_equipment. Through the above calculations, the complete load intensity assessment result L_total is obtained, which quantifies the overall load level of key monitoring nodes.

[0066] The load intensity assessment is processed by spectral transformation to generate a load distribution spectrum. The total load intensity assessment L_total provides complete time-domain information on load changes over time, and spectral transformation enables analysis of the frequency characteristics and periodic patterns of load changes. The Fast Fourier Transform (FFT) technique is used to convert the time series of the load intensity assessment result L_total into a frequency domain representation. The power spectral density of the load signal PSD(f) = |L_total(f)|² is calculated, where PSD is the power spectral density, L_total(f) is the Fourier transform of the load intensity assessment result, and f is the frequency. The frequency components of the load distribution spectrum are analyzed to identify the main frequency peaks and minor harmonic components. Low-frequency components correspond to long-term trends and slow changes in the load, while high-frequency components correspond to rapid fluctuations and sudden changes in the load. Characteristic parameters for the load distribution spectrum are extracted, including the dominant frequency, spectral width, peak power, and frequency centroid. The contributions of different monitoring elements to the load distribution spectrum are analyzed, and the characteristic frequency ranges of each element are identified.

[0067] This study aims to construct a surplus resource pool by mining monitoring resource reserves from the load distribution spectrum. A resource reserve identification algorithm is established, determining exploitable resource reserves through load threshold comparison and load gap analysis. A load threshold T_threshold is set; when the load of a frequency band falls below the threshold, there is available resource reserve R_available = T_threshold - L_actual for that band, where R_available is the available reserve and L_actual is the actual load. The reserve characteristics of different resource types are analyzed, including CPU computing reserve, memory storage reserve, network bandwidth reserve, and I / O processing reserve. A quantification method for surplus resources is established, converting abstract resource reserves into specific, allocable values. The organizational structure of the surplus resource pool is designed, classifying and managing surplus resources according to resource type, available time period, and available capacity. A dynamic update mechanism for surplus resources is established, dynamically adjusting the available reserve in the resource pool based on real-time load changes in the monitoring system. The spatiotemporal distribution characteristics of surplus resources are analyzed to identify the distribution patterns of resource reserves in time and space. Design a reserve strategy for surplus resources to provide a certain resource buffer for sudden high load demands.

[0068] The adjustment coefficient is obtained by frequency matching between the load distribution spectrum and the surplus resource pool. An objective function for frequency matching is established, with the optimization goals of maximizing monitoring accuracy and resource utilization. A matching algorithm is designed: M(f) = L(f) / R(f), where M(f) is the matching coefficient at frequency f, L(f) is the load value of the load distribution spectrum at frequency f, and R(f) is the available resources in the surplus resource pool at frequency f. The physical meaning of the matching coefficient is analyzed: a high matching coefficient indicates that the load demand in this frequency band exceeds the resource supply, requiring a reduction in monitoring frequency or an increase in resource allocation; a low matching coefficient indicates that there is sufficient resource surplus in this frequency band, allowing for an increase in monitoring frequency. A calculation method for the adjustment coefficient is established: K(f) = 1 / M(f). A high matching coefficient corresponds to a small adjustment coefficient, indicating frequency reduction; a low matching coefficient corresponds to a large adjustment coefficient, indicating frequency increase. Constraints on the adjustment coefficient are designed to ensure that the adjusted monitoring frequency remains within the system's allowable range. The frequency response characteristics of the adjustment coefficient are analyzed to identify frequency bands requiring focused adjustment and relatively stable frequency bands.

[0069] The monitoring frequency of the equipment is determined by performing frequency adjustment based on the adjustment coefficient. An execution algorithm for frequency adjustment is established: f_new = f_base × K(f), where f_new is the adjusted monitoring frequency, f_base is the base monitoring frequency, and K(f) is the adjustment coefficient. A constraint mechanism for frequency adjustment is designed to ensure that the adjusted frequency does not exceed the maximum operating frequency of the equipment, nor is it lower than the minimum frequency required to guarantee monitoring accuracy. A priority strategy for frequency adjustment is established, prioritizing the adjustment of the frequencies of critical monitoring nodes, followed by general monitoring nodes. The impact of frequency adjustment on monitoring performance is analyzed, and a relationship model is established between frequency, monitoring accuracy, response time, and resource consumption. A step-by-step execution mechanism for frequency adjustment is designed to avoid the impact of large frequency changes on the system through gradual adjustment. A feedback mechanism for frequency adjustment is established to evaluate the adjustment effect based on the adjusted system performance, and to perform secondary adjustments when necessary. Frequency adjustment strategies are designed for different equipment types: sensor equipment focuses on sampling frequency adjustment, data processing equipment focuses on calculation frequency adjustment, and communication equipment focuses on transmission frequency adjustment.

[0070] An adaptive acquisition sequence is constructed using equipment monitoring frequencies. The monitoring frequency determines the working rhythm of each device, and the adaptive acquisition sequence needs to coordinate the timing relationships between devices to avoid data acquisition conflicts and resource contention. Timing planning rules for the acquisition sequence are established, arranging the acquisition schedule according to the monitoring frequency and priority of each device. A time-slice allocation strategy is designed, dividing the monitoring period into multiple time slices and allocating appropriate time-slice resources to devices with different priorities. A conflict detection mechanism for the acquisition sequence is established to identify conflicts where multiple devices simultaneously request resources, resolving conflicts through timing adjustments and resource arbitration. Load balancing of the acquisition sequence is analyzed to ensure a relatively uniform distribution of acquisition load across time periods, avoiding overload in some periods while other periods are idle. An adaptive adjustment mechanism is designed to dynamically adjust the acquisition sequence arrangement based on actual acquisition results and system load. A synchronization mechanism for the acquisition sequence is established to ensure that data collected by different devices remains synchronized in time, facilitating subsequent data fusion and analysis. A fault-tolerance mechanism for the acquisition sequence is designed to quickly adjust the sequence arrangement when a device malfunctions or acquisition anomalies occur, ensuring the continuity of the overall acquisition task.

[0071] Data aggregation analysis is performed on adaptive acquisition sequences to extract the core monitoring set. A hierarchical architecture for data aggregation is established, proceeding step-by-step from the raw data layer, feature extraction layer, to the information fusion layer. A data importance assessment algorithm is designed, using methods such as information entropy, analysis of variance, and correlation analysis to evaluate the importance of each monitoring data point. Selection criteria for the core monitoring set are established, comprehensively considering factors such as data information content, sensitivity to change, quality indication capability, and real-time requirements. Principal component analysis is employed to reduce the dimensionality of the monitoring data and extract core feature components containing key information. A time window for data aggregation is designed, performing aggregation and statistical analysis on the monitoring data within an appropriate time frame to balance timeliness and stability requirements. A multi-source data fusion algorithm is established to effectively integrate monitoring data from different devices and of different types, forming comprehensive monitoring information. The correlations between monitoring data are analyzed to identify strongly correlated data combinations and relatively independent data categories.

[0072] Construct a monitoring process configuration based on the core monitoring set. Establish the overall architecture of the monitoring process, including key components such as data acquisition, data processing, anomaly detection, and response control. Design the parameter configurations for each process component, determining the specific values ​​of acquisition, processing, and control parameters based on the characteristics of the core monitoring set. Establish the execution sequence of the process, determining the start time, execution order, and completion deadline for each component. Design interface specifications between processes to ensure the accuracy of data transfer and information exchange between different process components. Establish an optimization mechanism for the process configuration, optimizing configuration parameters through process simulation and performance evaluation. Design a monitoring mechanism for process execution to monitor the execution status and performance indicators of each process component in real time. Establish version management for the process configuration, recording configuration change history and implementing a rollback mechanism to ensure the traceability and recoverability of the process configuration. Design an adaptive adjustment function for the process configuration to dynamically adjust configuration parameters based on monitoring results and process requirements.

[0073] Step S150: Energy consumption distribution data is obtained by using the power change component in the current density signal for energy consumption analysis. The monitoring cycle is determined by matching the load with the energy consumption distribution data and the monitoring process configuration. Time-sharing monitoring operation is performed according to the monitoring cycle to form a monitoring convergence field.

[0074] Specifically, energy consumption distribution data is obtained by analyzing the power change component in the current density signal. The current density signal contains rich information on power changes, which directly reflect the energy consumption characteristics and distribution patterns during the electroplating process. An algorithm for extracting the power change component is established, using signal decomposition techniques to separate the power change component from other components in the current density signal. Instantaneous power P(t) = I(t) × V(t) is calculated, where P(t) is the instantaneous power, I(t) is the current density signal, and V(t) is the corresponding voltage signal. The frequency characteristics of the power change component are analyzed to identify the main frequency bands and change patterns of power changes. The correlation between power change and energy consumption is established, and energy consumption E = ∫P(t)dt is calculated through power integration, where E is the cumulative energy consumption. The spatial distribution characteristics of energy consumption are analyzed, and the energy consumption corresponding to the current density signal at different locations is spatially mapped. An energy consumption density calculation method is established, distributing the total energy consumption according to spatial regions, forming an energy consumption density distribution ρ_E = E / A, where ρ_E is the energy consumption density and A is the area of ​​the region. Analyze the temporal variation characteristics of energy consumption distribution to identify peak, trough, and stable periods of energy consumption. Establish statistical characteristics of energy consumption distribution data, including parameters such as energy consumption mean, variance, skewness, and kurtosis.

[0075] In some embodiments, determining the monitoring period by load matching using the energy consumption distribution data and the monitoring process configuration includes: using the energy consumption distribution data to reconstruct energy density and obtain a density distribution field; decomposing the monitoring process configuration into monitoring load slices to establish a slice load library; performing spatial overlay matching between the density distribution field and the slice load library to obtain an overlay fusion point; and performing time-series convergence analysis based on the overlay fusion point to determine the monitoring period.

[0076] Energy density is reconstructed using energy consumption distribution data to obtain the density distribution field. Kriging interpolation is used to process the spatial reconstruction of energy density, considering spatial correlation and distance weights between energy consumption data points. An optimization strategy for the reconstruction parameters is designed, adjusting the interpolation parameters and weighting coefficients according to the distribution characteristics of the energy consumption data. Resolution control of the density distribution field is established, increasing spatial resolution in areas of drastic energy consumption changes and appropriately decreasing resolution in areas of relatively stable energy consumption. The spatial gradient characteristics of the density distribution field are analyzed, and the energy density gradient ∇ρ_E=(∂ρ_E / ∂x,∂ρ_E / ∂y) is calculated to identify the directions and regions of most drastic energy density changes. Boundary condition processing for the density distribution field is established to ensure the physical rationality of the reconstruction results in the boundary regions. A smoothing process for the density distribution field is designed to remove artifacts and discontinuities generated during the reconstruction process. The statistical characteristics of the density distribution field are analyzed, including characteristic parameters such as the density mean distribution, density variance distribution, and density skewness distribution.

[0077] The monitoring process is configured and decomposed into monitoring load slices to establish a slice load library. Criteria and methods for process decomposition are established, and slices are divided according to the functional characteristics, execution time, and resource requirements of the monitoring tasks. The definition structure of the monitoring load slices is designed, with each slice containing key information such as slice identifier, execution task, resource requirements, time constraints, and priority. The load characteristics of different types of monitoring tasks are analyzed: data acquisition slices mainly consume I / O resources, data processing slices mainly consume computing resources, and data transmission slices mainly consume network resources. A quantification method for slice load is established: L_slice = w1C + w2M + w3N, where L_slice is the slice load, C is the computing load, M is the memory load, N is the network load, and w1, w2, and w3 are weighting coefficients. The organizational structure of the slice load library is designed, classifying and storing slices according to load type, execution time, and resource requirements. Dependency analysis between slices is established to identify constraints such as pre- and post-dependencies, parallel execution, and mutual exclusion relationships between slices. A dynamic evaluation mechanism for slice load is designed to dynamically adjust the slice load evaluation value based on actual execution conditions. Establish a retrieval and matching mechanism for the slice load library to support fast slice search and combination based on load characteristics and constraints.

[0078] The density distribution field and the slice load library are spatially superimposed and matched to obtain superposition fusion points. A mathematical description of spatial superposition is established, and the density distribution field and slice load are superimposed and calculated in the same spatial coordinate system. An evaluation function for superposition matching is designed: M(x,y)=ρ_E(x,y)×L_slice(x,y), where M(x,y) is the matching degree at position (x,y), ρ_E(x,y) is the energy density, and L_slice(x,y) is the slice load density. The spatial distribution characteristics of the superposition results are analyzed to identify regions with high and low matching degrees. A fusion point identification criterion is established, marking locations with superposition matching degrees exceeding a threshold as superposition fusion points. A clustering algorithm is used to group the superposition fusion points, merging spatially adjacent fusion points into fusion regions. The centroid position of each fusion region is calculated as a representative fusion point. The load balance characteristics of the fusion points are analyzed to ensure relatively uniform load distribution among the fusion points and avoid local overload. A priority ranking of the fusion points is designed, determining the priority of fusion points based on factors such as matching degree, regional importance, and resource availability.

[0079] The monitoring period is determined based on time-series convergence analysis using overlay fusion points. A time-series convergence analysis framework is established, comprehensively considering factors such as monitoring accuracy requirements, system resource constraints, and response time constraints. The temporal characteristics of each fusion point are analyzed, including parameters such as task execution time, data update frequency, and response latency. A convergence criterion is designed, and the monitoring period that enables the system performance to reach a convergent state is found through iterative calculation. The convergence calculation formula is established as T_converge=argmin(∑ᵢ|f(T)-f_target|), where T_converge is the convergence period, f(T) is the system performance function corresponding to period T, and f_target is the target performance value. The impact of different monitoring periods on system performance is analyzed, including indicators such as monitoring accuracy, resource utilization, and response speed. Constraints for period optimization are designed to ensure that the determined monitoring period meets the system's physical limitations and performance requirements. A multi-objective optimization algorithm is established to find the optimal balance between monitoring accuracy and resource consumption. The stability of the monitoring period is analyzed to ensure that the determined period maintains good performance under different operating conditions.

[0080] A monitoring aggregation field is formed by performing time-sharing monitoring operations according to the monitoring cycle. A scheduling algorithm for time-sharing monitoring is established to arrange the execution sequence of monitoring activities based on the monitoring cycle and task priority. A time-slice allocation strategy is designed to divide the monitoring cycle into multiple time slices and allocate appropriate time resources to different types of monitoring tasks. A concurrency control mechanism for monitoring tasks is established to coordinate the simultaneous operation of multiple monitoring devices and avoid resource conflicts and data interference. The load distribution of time-sharing monitoring is analyzed to ensure that the monitoring load of each time slice is relatively balanced and to avoid overload in certain periods. A synchronization mechanism for monitoring data is designed to ensure that monitoring data from different time slices can be accurately correlated and fused. An algorithm for constructing the monitoring aggregation field is established to form a continuous monitoring field from the discrete data obtained from time-sharing monitoring through spatial interpolation and temporal fusion techniques. The coverage characteristics of the monitoring aggregation field are analyzed to ensure that the aggregation field can cover the key areas and important time periods of the electroplating tank.

[0081] Step S160: Determine the process control window using the monitoring convergence field, and convert the monitoring process configuration into intelligent monitoring instructions through the process control window.

[0082] In some embodiments, determining the process control window using the monitoring convergence field includes: constructing a multi-dimensional monitoring space using the monitoring convergence field; performing a control region search in the multi-dimensional monitoring space to obtain candidate control regions; performing boundary optimization on the candidate control regions to determine the optimal control boundary; and determining the process control window based on the optimal control boundary.

[0083] A multi-dimensional monitoring space is constructed using a monitoring convergence field. A coordinate system for this multi-dimensional space is established, with spatial coordinates (x, y) as the basic coordinate axis, and extended coordinate axes including time dimension t, parameter dimension p, and quality dimension q. A data mapping method for the multi-dimensional space is designed to map monitoring data from the monitoring convergence field to corresponding dimensional coordinates according to their temporal characteristics, parameter types, and quality levels. The relationships between the dimensions are analyzed, and a dimensional weight allocation strategy is established: W = w1X + w2Y + w3T + w4P + w5Q, where W is the comprehensive weight, X and Y are spatial weights, T is the time weight, P is the parameter weight, Q is the quality weight, and w1, w2, w3, w4, w5, w6, w7, w8, w9, w1, w1, w1, w2, w3, w1 ... 3、 w 4、 w5 represents the weighting coefficient. A distance calculation method for multidimensional space is established, defining a distance metric between two points in multidimensional space, supporting multidimensional similarity analysis and clustering. Resolution control for multidimensional space is established, increasing resolution in important dimensions and appropriately decreasing resolution in secondary dimensions to balance computational efficiency. The geometric features of multidimensional space are analyzed, identifying different characteristic regions such as dense regions, sparse regions, and boundary regions.

[0084] Candidate control areas are obtained through control area search in a multi-dimensional monitoring space. Based on the constructed multi-dimensional monitoring space, search criteria for control areas are established, setting search conditions such as monitoring information density threshold, process sensitivity threshold, and control responsiveness threshold. A scanning strategy for area search is designed, employing a sliding window technique to systematically scan the multi-dimensional space and evaluate the control potential of each candidate area one by one. An evaluation function for candidate areas is established: F = α·D + β·S + γ·R, where F is the area evaluation value, D is the monitoring information density, S is the process sensitivity, R is the control responsiveness, and α, β, and γ are weighting coefficients. The spatial distribution characteristics of candidate areas are analyzed, identifying basic features such as the geometric shape, spatial location, and coverage area of ​​the areas. A threshold mechanism for area selection is designed, marking areas with evaluation values ​​exceeding the threshold as valid candidate control areas. An overlap handling mechanism for candidate areas is established; when multiple candidate areas overlap, the overlap problem is handled through merging or selection strategies. A hierarchical system for candidate areas is designed, classifying candidate areas into different levels such as priority control areas, general control areas, and alternative control areas according to their evaluation values.

[0085] The optimal control boundary is determined through boundary optimization of candidate control areas. Based on the acquired candidate control areas, a boundary optimization algorithm is used to accurately determine the optimal boundary range of each control area. An objective function for boundary optimization is established, with the optimization goals of maximizing control effectiveness and minimizing boundary complexity. A search algorithm for boundary adjustment is designed, using optimization methods such as gradient search and genetic algorithms to find the optimal boundary configuration. Constraints on boundary optimization are analyzed, including regional connectivity constraints, minimum area constraints, and shape regularity constraints. An optimization strategy for boundary points is established, adjusting the positions of key boundary points one by one to improve the global boundary through local optimization. A boundary smoothing algorithm is designed to remove jagged protrusions and irregular bumps in the boundary, forming a smooth and continuous boundary curve. A convergence criterion for boundary optimization is established, and the convergence state of boundary adjustment is found through iterative calculation. The geometric characteristics of the optimal boundary are analyzed, including geometric parameters such as boundary length, enclosed area, convexity / concavity, and symmetry.

[0086] The process control window is determined based on the optimal control boundary. A standardized format for window parameters is designed, including basic parameters such as window center coordinates, window size, window shape, and window priority. The coverage characteristics of the control window are analyzed to ensure that it completely covers the area enclosed by the optimal control boundary. Resolution settings for the control window are established, determining the control point density and parameter resolution within the window based on control accuracy requirements. A multi-window organization and management mechanism is designed, establishing hierarchical relationships and coordination strategies among multiple control windows. Execution priorities for control windows are established, determining the execution order based on window importance and urgency. A dynamic adjustment interface for the control window is designed, supporting dynamic modification of window parameters and configurations according to process requirements. The effectiveness evaluation of the control window is analyzed, evaluating the rationality of the window configuration through indicators such as control performance and response time.

[0087] The monitoring process configuration is converted into intelligent monitoring instructions through the process control window. Based on the defined process control window and the previously obtained monitoring process configuration, an intelligent monitoring instruction generation mechanism is established to convert the abstract monitoring process into specific equipment execution instructions. The process control window defines the spatial range and parameter requirements for monitoring and control, while the monitoring process configuration provides the execution plan for monitoring activities. The combination of the two can generate targeted intelligent monitoring instructions. An instruction generation conversion algorithm is established to convert the abstract description in the monitoring process configuration into specific instructions that can be executed by the equipment. Standardized specifications for instruction formats are designed, including standard fields such as instruction identifier, execution equipment, control parameters, execution time, and priority. The instruction characteristics of different types of monitoring equipment are analyzed; sensor equipment requires sampling frequency and measurement range instructions, while actuator equipment requires action type and control amplitude instructions. A mapping relationship for instruction parameters is established to convert the spatial coordinates and control requirements in the control window into specific parameter settings for the equipment. The timing arrangement of the instruction sequence is designed, arranging the execution order and time interval of each instruction according to the timing requirements in the monitoring process configuration. A coordination mechanism for instruction execution is established to ensure that the execution of instructions between multiple devices does not conflict or interfere with each other. Establish a feedback mechanism for instruction execution status, monitor the execution progress and result status of instructions in real time, and ultimately monitor the entire electroplating production process.

[0088] To implement the electroplating production process monitoring method corresponding to the above method embodiments, in order to achieve the corresponding functions and technical effects. See also Figure 2 , Figure 2 This diagram illustrates a structural block diagram of an electroplating production process monitoring system 200 according to an embodiment of this application. For ease of explanation, only the parts relevant to this embodiment are shown. The electroplating production process monitoring system 200 provided in this embodiment includes:

[0089] The signal acquisition module 201 is used to monitor the current density signal and the temperature distribution data of the plating bath in the electroplating tank, perform density gradient analysis on the current density signal to obtain the current distribution characteristics, extract the hot spot migration trajectory in the temperature distribution data of the plating bath, and perform thermoelectric correlation between the current distribution characteristics and the hot spot migration trajectory to generate process coupling response values.

[0090] The transfer path module 202 is used to filter abnormal process components using the process coupling response value, locate the plating solution circulation transfer path through the abnormal process components, detect flow rate changes along the transfer path to obtain flow field distribution data, and construct a sensor network layout matrix based on the flow field distribution data.

[0091] Risk identification module 203 is used to perform big data processing on the sensor network layout matrix to identify quality risk areas, extract the coating thickness gradient field of the quality risk areas, use the coating thickness gradient field to determine the key monitoring locations, and compare the key monitoring locations with the process coupling response values ​​to determine the critical monitoring nodes.

[0092] Frequency adjustment module 204 is used to adjust the monitoring frequency of the equipment according to the key monitoring nodes, form an adaptive acquisition sequence using the monitoring frequency of the equipment, perform data aggregation analysis on the adaptive acquisition sequence to extract the core monitoring set, and construct the monitoring process configuration according to the core monitoring set;

[0093] Energy management module 205 is used to perform energy consumption analysis using the power change component in the current density signal to obtain energy consumption distribution data, determine the monitoring cycle by matching the energy consumption distribution data with the monitoring process configuration, and perform time-sharing monitoring operation according to the monitoring cycle to form a monitoring convergence field.

[0094] The intelligent control module 206 is used to determine the process control window using the monitoring convergence field, and to convert the monitoring process configuration into intelligent monitoring instructions through the process control window.

[0095] The electroplating production process monitoring system 200 described above can implement one of the electroplating production process monitoring methods described in the above method embodiments. The options in the above method embodiments are also applicable to this embodiment, and will not be detailed here. The remaining contents of this application embodiment can be referred to the contents of the above method embodiments, and will not be repeated in this embodiment.

[0096] The purpose of the above embodiments is to reproduce and derive the technical solution of the present invention by way of example, and to fully describe the technical solution, purpose and effect of the present invention. The purpose is to enable the public to have a more thorough and comprehensive understanding of the disclosure of the present invention, and not to limit the scope of protection of the present invention.

[0097] The above embodiments are not an exhaustive list based on the present invention, and there may be many other embodiments not listed. Any substitutions and improvements made without departing from the concept of the present invention are within the protection scope of the present invention.

Claims

1. A method for monitoring the entire electroplating production process, characterized in that, include: Monitor the current density signal and the temperature distribution data of the plating bath in the electroplating tank, perform density gradient analysis on the current density signal to obtain the current distribution characteristics, extract the hot spot migration trajectory in the temperature distribution data of the plating bath, and perform thermoelectric correlation between the current distribution characteristics and the hot spot migration trajectory to generate process coupling response values. Abnormal process components are screened using the process coupling response value. The abnormal process components are used to locate the plating solution circulation and transfer path. The flow rate change is detected along the transfer path to obtain flow field distribution data. A sensor network layout matrix is ​​constructed based on the flow field distribution data. The sensor network layout matrix is ​​processed with big data to identify quality risk areas. The coating thickness gradient field of the quality risk areas is extracted. The coating thickness gradient field is used to determine the key monitoring locations. The key monitoring locations are compared with the process coupling response values ​​to determine the critical monitoring nodes. The monitoring frequency of the equipment is adjusted according to the key monitoring nodes, and an adaptive acquisition sequence is formed using the monitoring frequency of the equipment. Data aggregation and analysis are performed on the adaptive acquisition sequence to extract the core monitoring set, and the monitoring process configuration is constructed based on the core monitoring set. Energy consumption distribution data is obtained by using the power change component in the current density signal for energy consumption analysis. The monitoring cycle is determined by load matching between the energy distribution data and the monitoring process configuration. Time-division monitoring is performed according to the monitoring cycle to form a monitoring convergence field. The process control window is determined using the monitoring convergence field, and the monitoring process configuration is converted into intelligent monitoring instructions through the process control window.

2. The method according to claim 1, characterized in that, The step of performing density gradient analysis on the current density signal to obtain current distribution characteristics includes: The current density signal is phase demodulated and separated to obtain phase feature components; The phase characteristic components are spatially mapped to form a current density cloud map; Gradient vector tracing is performed on the current density cloud map to generate gradient streamlines; The current distribution characteristics are obtained by summarizing the feature encoding of the gradient streamline bundle.

3. The method according to claim 1, characterized in that, The method of locating the plating solution circulation path through the abnormal process component includes: The abnormal process components are backpropagated to trace the source coordinates. An energy diffusion field is established based on the coordinates of the propagation source. Path aggregation analysis is performed in the energy diffusion field to generate aggregated path bundles; The main path of the polymerization path bundle is extracted to determine the plating solution circulation and transfer path.

4. The method according to claim 1, characterized in that, The step of performing big data processing on the sensor network layout matrix to identify quality risk areas includes: A collaborative monitoring field is generated by triggering electric field inductive coupling through the arrangement of the sensor network matrix. Capture the mass anomaly cluster center in the collaborative monitoring field; Risk density analysis is performed on the aforementioned quality anomaly cluster centers to form a risk density distribution; Quality risk areas are identified based on the aforementioned risk density distribution.

5. The method according to claim 1, characterized in that, The adjustment of the equipment monitoring frequency based on the key monitoring nodes includes: The load of the key monitoring nodes is assessed to generate a load distribution spectrum; The surplus resources are extracted from the load distribution spectrum to construct a surplus resource pool; The adjustment coefficient is obtained by frequency matching between the load distribution spectrum and the surplus resource pool; The monitoring frequency of the equipment is determined by performing frequency adjustment based on the adjustment coefficient.

6. The method according to claim 1, characterized in that, The step of determining the monitoring cycle by matching the energy consumption distribution data with the monitoring process configuration includes: The energy density distribution field is obtained by reconstructing the energy density using the energy consumption distribution data; The monitoring process configuration is decomposed into monitoring load slices to establish a slice load library; The density distribution field and the slice load library are spatially superimposed and matched to obtain the superposition and fusion point; The monitoring period is determined by performing time-series convergence analysis based on the superposition and fusion points.

7. The method according to claim 1, characterized in that, The process control window is determined using the monitoring convergence field, including: A multi-dimensional monitoring space is constructed using the aforementioned monitoring convergence field; Candidate control regions are obtained by searching the control region within the multidimensional monitoring space. The candidate control regions are subjected to boundary optimization to determine the optimal control boundary; The process control window is determined based on the optimal control boundary.

8. The method according to claim 4, characterized in that, The step of generating a collaborative monitoring field by arranging a matrix through the sensor network to trigger electric field inductive coupling includes: The sensor network is arranged in a matrix to trigger electric field inductive coupling to generate a coupling signal; Synchronization features were extracted by performing Faraday wave analysis on the coupled signals; The synchronization characteristics are mapped to a monitoring sensitivity distribution; A collaborative monitoring field is generated by gradient fusion of the sensitivity distribution.

9. The method according to claim 5, characterized in that, The process of assessing the monitoring load of the key monitoring nodes and generating a load distribution spectrum includes: Perform cross-time period load tracking analysis on the key monitoring nodes to obtain load change data; The load contribution effect of each monitoring element is analyzed using the load change data to obtain the load intensity assessment result. The monitoring elements include current density change, plating solution temperature fluctuation and equipment operating status. The load intensity assessment results are subjected to spectral conversion processing to generate a load distribution spectrum.

10. A monitoring system for the entire electroplating production process, characterized in that, include: The signal acquisition module is used to monitor the current density signal and the temperature distribution data of the plating bath in the electroplating tank, perform density gradient analysis on the current density signal to obtain the current distribution characteristics, extract the hot spot migration trajectory in the temperature distribution data of the plating bath, and perform thermoelectric correlation between the current distribution characteristics and the hot spot migration trajectory to generate process coupling response values. The transfer path module is used to filter abnormal process components using the process coupling response value, locate the plating solution circulation transfer path through the abnormal process components, detect flow rate changes along the transfer path to obtain flow field distribution data, and construct a sensor network layout matrix based on the flow field distribution data. The risk identification module is used to perform big data processing on the sensor network layout matrix to identify quality risk areas, extract the coating thickness gradient field of the quality risk areas, use the coating thickness gradient field to determine the key monitoring locations, and compare the key monitoring locations with the process coupling response values ​​to determine the critical monitoring nodes. The frequency adjustment module is used to adjust the monitoring frequency of the equipment according to the key monitoring nodes, use the monitoring frequency of the equipment to form an adaptive acquisition sequence, perform data aggregation analysis on the adaptive acquisition sequence to extract the core monitoring set, and construct the monitoring process configuration according to the core monitoring set. The energy management module is used to perform energy consumption analysis using the power change component in the current density signal to obtain energy consumption distribution data, determine the monitoring cycle by matching the energy consumption distribution data with the monitoring process configuration, and perform time-division monitoring operations according to the monitoring cycle to form a monitoring convergence field. The intelligent control module is used to determine the process control window using the monitoring convergence field, and to convert the monitoring process configuration into intelligent monitoring instructions through the process control window.

Images