Intelligent electroplating parameter adaptive regulation and control system based on big data analysis

Abstract

The invention relates to the technical field of adaptive regulation and control, in particular to an intelligent electroplating parameter adaptive regulation and control system based on big data analysis, and the system comprises a data collection and edge sensing module which is used for arranging a sensor, collecting multi-parameter data, and outputting an edge multi-parameter data stream after time sequence calibration and noise suppression; the edge modeling and feature extraction module is used for performing sliding window modeling and parameter coupling feature extraction on the edge multi-parameter data stream to construct an edge feature set; the big data collaborative reasoning module is used for sending the edge feature set to the central server, performing matching analysis on the edge feature set and a historical big data sample library, and generating a regulation and control reference strategy; and the self-adaptive regulation and control execution module is used for generating a regulation and control decision instruction in real time, executing closed-loop regulation and feeding back a regulation and control result to the data acquisition and edge sensing module so as to update the edge multi-parameter data flow. According to the method, an adaptive gain control mechanism and parameter partial derivative matrix modeling are introduced, and the method is suitable for modern manufacturing scenes with multi-variety, precise and high-consistency requirements.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of adaptive regulation, and particularly relates to an intelligent electroplating parameter adaptive regulation system based on big data analysis. BACKGROUND

[0002] The electroplating process is widely used in the fields of electronics, electricity, machinery and precision manufacturing, and its stability and consistency directly affect the functional reliability and appearance quality of the final product. The traditional electroplating control system usually relies on experience to set parameters, and cooperates with a central server to perform periodic sampling and back-end analysis, so it is difficult to achieve dynamic linkage adjustment of high frequency and multiple parameters. Especially when facing the coupling changes of voltage, current, temperature, pH value and flow, the existing system lacks effective real-time identification and response mechanism, and the regulation precision and response speed cannot meet the modern high-density and high-consistency electroplating requirements. At the same time, there is a lack of state coordination mechanism among electroplating equipment, resulting in late abnormal early warning and rigid regulation strategy, which cannot adapt to the flexible manufacturing scene of multiple categories and small batches.

[0003] In recent years, although some researches have tried to introduce big data analysis and machine learning methods to optimize electroplating control, there are still the following deficiencies: the data processing highly depends on the central server, the edge computing capability is missing, and the low-latency response of on-site high-frequency data cannot be achieved; the control strategy is only based on historical sample prediction, and lacks the closed-loop regulation capability of integrating real-time process state and mechanism knowledge; the parameter coupling relationship is not systematically modeled, and interference amplification and control conflicts are easily generated in the regulation process. At present, there is still a lack of an integrated system with edge computing preprocessing capability, parameter coupling decoupling mechanism and closed-loop adaptive regulation capability to realize intelligent, flexible and fine control of the electroplating process. SUMMARY

[0004] The present application provides an intelligent electroplating parameter adaptive regulation system based on big data analysis.

[0005] An intelligent electroplating parameter adaptive regulation system based on big data analysis comprises the following modules: A data acquisition and edge perception module: sensors are arranged in the electroplating site to collect multi-parameter data including voltage, current, temperature, pH value and flow, and the multi-parameter data are preliminarily preprocessed based on the edge node, and the edge multi-parameter data stream after time series calibration and noise suppression is output; An edge modeling and feature extraction module: the edge multi-parameter data stream is modeled by sliding window and the parameter coupling features are extracted, and an edge feature set including process state trend, local disturbance mode and abnormal sign index is constructed; Big data collaborative reasoning module: send the edge feature set to the center server, match and analyze with the historical big data sample library, call the multi-source case library and process mechanism knowledge base to generate a control reference strategy, the control reference strategy includes a target parameter adjustment range, a control decoupling suggestion set and an electroplating beat optimization suggestion; Adaptive control execution module: based on the control reference strategy, in combination with the current edge feature set, real-time generation of control decision instructions, closed-loop adjustment of electroplating power supply, electrolyte flow and stirring frequency control unit, and feedback of control results to the data acquisition and edge sensing module to update the edge multi-parameter data stream.

[0006] Optionally, the data acquisition and edge sensing module comprises: Sensor layout and acquisition: layout multiple types of industrial sensors in the electroplating field, including voltage sensors, current sensors, temperature sensors, pH sensors and flow sensors, real-time acquisition of multi-parameter data, and local reception and synchronization through edge nodes; Data alignment and timing calibration: time aligning asynchronous data collected by different sensors to construct unified multi-parameter time series data; Edge pre-processing output: filtering and abnormal correction on the aligned multi-parameter time series data, removing noise interference, and outputting stable edge multi-parameter data stream.

[0007] Optionally, the data alignment and timing calibration comprises: Multi-source data unified modeling: to solve the problem of different sampling times of different types of sensors, the system unifies the parameter acquisition results into a time series matrix form; Asynchronous data linear interpolation calibration: to eliminate the influence of sampling time difference on parameter comparison and analysis, the system uses linear interpolation algorithm for time normalization of asynchronous data.

[0008] Optionally, the edge pre-processing output comprises: Weight function definition and filtering control: using Gaussian kernel function form, calculating the different weights given to adjacent time points when using local weighted regression filter algorithm, i.e. neighborhood weight function; Abnormal point detection and filtering modeling: applying local weighted regression filter algorithm to the data matrix after time series calibration, identifying and smoothing the short-term fluctuations and outliers in the time series calibrated data, and obtaining the filtered data as the edge multi-parameter data stream.

[0009] Optionally, the edge modeling and feature extraction module comprises: Sliding window modeling and local state aggregation: applying a sliding time window mechanism to the input edge multi-parameter data stream, combining multi-parameter data in the current and previous period of time, and constructing a local state set within the time window; Parameter coupling feature extraction and index construction: the system extracts three types of key features from the edge multi-parameter data stream within each sliding window, including the average change rate of each parameter, the fluctuation degree of each parameter, and the deviation of the current parameter value from the historical average value, and forms an edge feature set.

[0010] Optionally, the parameter coupling feature extraction and index construction includes: Parameter coupling feature extraction: extracting state trend indicators, local disturbance scores, and abnormal symptom indices; Index construction: combining the extracted state trend indicators, local disturbance scores, and abnormal symptom indices to form an edge feature set.

[0011] Optionally, the big data collaborative reasoning module includes: Feature set upload and matching analysis: uploading the edge feature set output by the edge modeling and feature extraction module to the central server, and the server performs similarity matching analysis in the historical big data sample library to form a candidate feature set; Multi-source knowledge fusion and strategy generation: the system performs fusion analysis on the matched candidate feature set and historical cases and process knowledge to generate a control reference strategy.

[0012] Optionally, the multi-source knowledge fusion and strategy generation includes: Recommended adjustment range generation: based on the historical sample data in the candidate feature set, the upper and lower limits of the key control parameters are calculated to construct the recommended adjustment interval under the current operating state; Coupling relationship analysis and decoupling suggestion: based on the edge feature set, a parameter partial derivative matrix is constructed to obtain a control decoupling suggestion set based on coupling relationship analysis; Electroplating beat optimization suggestion: calling the beat feedback information in the historical cases and combining the process model to infer and generate a beat optimization suggestion; Control reference strategy generation: based on the generated recommended adjustment interval, control decoupling suggestion set, and beat optimization suggestion result, a control reference strategy is constructed.

[0013] Optionally, the adaptive control execution module includes: Control instruction generation and control execution: combining the control reference strategy and the current edge feature set, real-time control instructions are generated according to the parameter deviation to adjust the gain to generate control instructions and execute; Control feedback and data update: after the regulation execution is completed, the measured response values and execution states returned by each execution unit are collected and fused with the original edge data to update the edge multi-parameter data stream for the next round.

[0014] Optionally, the regulation instruction generation and control execution comprises: Decision fusion and regulation instruction generation: receiving the regulation reference strategy and the current edge feature set, and making adaptive regulation decisions based on the strategy interval and the current state; Control execution and parameter adjustment: the system distributes each parameter control instruction to the corresponding physical control unit.

[0015] The beneficial effects of the present application are: The present application realizes intelligent management of the whole process from data acquisition, real-time modeling, strategy reasoning to closed-loop control by constructing an integrated electroplating parameter regulation system based on edge computing and big data collaborative analysis. The system introduces edge nodes in the electroplating field, pre-processes, time-aligns and sliding window models the multi-source data of voltage, current, temperature, pH value and flow, effectively reduces the central processing delay and communication bandwidth pressure, and improves the local recognition ability of the system to abnormal fluctuations and disturbance patterns. Through edge feature extraction and feature uploading, the system realizes similarity matching with historical big data samples, generates regulation reference strategies in combination with the case library and process knowledge base, and has stronger context understanding and reasoning ability.

[0016] The present application introduces an adaptive gain control mechanism and a parameter partial derivative matrix modeling in the regulation execution layer, can dynamically adjust the regulation strength of key parameters according to the feature change degree, and automatically generates control decoupling suggestions through coupling analysis, effectively avoiding interference conflicts between parameters; at the same time, the regulation results are fed back to the perception module in real time, forming a complete closed-loop data update path; the overall system can support dynamic beat optimization, abnormal trend prediction and multi-parameter collaborative control in flexible electroplating lines. Compared with traditional control systems, the present application has significant improvement in real-time performance, accuracy, adaptability and scalability, and is suitable for modern manufacturing scenarios with multiple varieties, precision and high consistency requirements. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only illustrate the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0018] Fig. 1 The system module diagram of the embodiment of the present application is shown in the figure. Fig. 2 The data acquisition and perception diagram of the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0019] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0020] It should be noted that references in the specification to "one embodiment," "an embodiment," "exemplary embodiments," "some embodiments," etc. indicate that the described embodiments may include specific features, structures, or characteristics, but not necessarily every embodiment will include such specific features, structures, or characteristics. Furthermore, when specific features, structures, or characteristics are described in conjunction with an embodiment, it is within the knowledge of persons skilled in the relevant art to implement such features, structures, or characteristics in conjunction with other embodiments (whether or not explicitly described).

[0021] In general, terms can be understood, at least in part, from their use in context. For example, depending at least in part on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending at least in part on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0022] like Figs. 1-2 As shown in the figure, an intelligent electroplating parameter adaptive control system based on big data analysis includes the following modules: Data acquisition and edge perception module: Sensors are deployed at the electroplating site to collect multi-parameter data, including voltage, current, temperature, pH value, and flow rate. The multi-parameter data is pre-processed based on edge nodes and outputs edge multi-parameter data streams that have been time-calibrated and noise-suppressed. Edge modeling and feature extraction module: This module performs sliding window modeling and parameter coupling feature extraction on edge multi-parameter data streams, constructing an edge feature set including process state trends, local disturbance patterns, and abnormal sign indicators; Big data collaborative reasoning module: sends edge feature sets to the central server, performs matching analysis with the historical big data sample library, and calls the multi-source case library and process mechanism knowledge library to generate control reference strategies. The control reference strategies include the target parameter adjustment range, the control decoupling suggestion set, and the electroplating beat optimization suggestions. Adaptive regulation execution module: based on the regulation reference strategy, combined with the current edge feature set, real-time generation of regulation decision instruction, the closed loop adjustment of the electroplating power supply, electrolyte flow and stirring frequency control unit is executed, and the regulation result is fed back to the data acquisition and edge perception module to update the edge multi-parameter data stream.

[0023] The data acquisition and edge perception module comprises: Sensor layout and acquisition: a plurality of industrial-grade sensors are laid out in the electroplating site, including voltage sensors, current sensors, temperature sensors, pH sensors and flow sensors, real-time multi-parameter data are collected, and local reception and synchronization are performed through edge nodes; specifically comprising: (1) Voltage sensor and current sensor: deployed at the power supply control point between the anode and the cathode, used for measuring the real-time voltage and current signals of the electroplating loop; (2) Temperature sensor and pH sensor, respectively installed in the electrolyte flow path and the tank side wall, collecting chemical environment fluctuations; (3) Flow sensor, laid in the inlet and circulation branch, collecting electrolyte flow rate and flow; Data alignment and timing calibration: time alignment is performed on the asynchronous data collected by different sensors, unified multi-parameter time series data is constructed, and comparability of various parameters at the same time point is ensured; Edge preprocessing output: filtering and abnormal correction are performed on the aligned multi-parameter time series data, noise interference is removed, and stable edge multi-parameter data stream is output for subsequent feature analysis.

[0024] Data alignment and timing calibration comprises: Multi-source data unified modeling: in view of the problem that different types of sensors have different sampling times, the system unifies the parameter collection results into a time series matrix form, which is expressed as: ; Among them, is a unified multi-parameter time series data vector composed of various electroplating process parameters at time , which is used for subsequent edge modeling and analysis, is the voltage value at time , is the current value at time , is the electrolyte temperature at time , is the solution pH value at time , is the flow value at time , and the matrix serves as the basic structure of multi-parameter data fusion, ensuring that various parameters are processed in a unified framework. Asynchronous data linear interpolation calibration: in order to eliminate the influence of sampling time difference on parameter comparison and analysis, the system adopts linear interpolation algorithm to carry out time normalization processing on asynchronous data, and the specific filling method is represented as: ; Among them, represents the value of a certain type of sensor at time (such as , etc.), is the standard sampling time point, and are its adjacent known time points before and after, and the interpolation processing ensures that the data collected by different sensors can be aligned at the same time.

[0025] The edge preprocessing output includes: Weight function definition and filter control: using Gaussian kernel function form, the different weights given to adjacent time points when calculating the local weighted regression filter (LOWESS) algorithm, that is, the neighborhood weight function , is represented as: ; Among them, is the current time point, is the reference time point in the neighborhood, is the time window bandwidth, which determines the range and sensitivity of the weight distribution. The larger the window bandwidth, the stronger the smoothing effect, the steeper the weight function, and the more sensitive the filtering to local changes. Abnormal point detection and filtering modeling: for the data matrix that has completed time series calibration, the local weighted regression filter (LOWESS) algorithm is applied to identify and smooth the short-term fluctuations and outliers in the time series calibrated data, and the filtered data is the edge multi-parameter data stream , represented as: ; Among them, is the neighborhood weight function, is the time window bandwidth, which controls the smoothing degree, and the edge multi-parameter data stream output after filtering is used as the input of the subsequent edge modeling and feature extraction module, which can improve the stability of the parameter data.

[0026] The edge modeling and feature extraction module includes: Sliding window modeling and local state aggregation: the input edge multi-parameter data stream is modeled by applying a sliding time window mechanism, combining the multi-parameter data in the current and previous time period to construct the local state set within the time window , represented as: ; wherein, is the local state set within the time window constructed by the sliding window data block ending at the current time and with length , is the filtered edge multi-parameter data vector at the current time, is the sliding window length in time steps, by which the time series segment is extracted for the subsequent analysis of the coupling relationship between parameters; Parameter coupling feature extraction and index construction: In each sliding window, the system extracts three types of key features from the edge multi-parameter data stream, including the average change rate of each parameter, the fluctuation degree of each parameter, the weighted combination to generate the disturbance score, and the deviation of the current parameter value from the historical average value, to identify potential abnormal signs and form the edge feature set as the input for subsequent intelligent analysis.

[0027] Parameter coupling feature extraction and index construction include: Parameter coupling feature extraction: Extract state trend indicators, local disturbance scores, and abnormal sign indexes, including: (1) State trend indicators: Calculate the average change rate of each parameter in the window, which is used to depict the process trend. Taking voltage as an example, it is expressed as: wherein, μ v is the average change rate of voltage; (2) Local disturbance score: Weighted combination of the standard deviation of parameter changes in the window, define disturbance degree Ψ t , expressed as: Ψ t = α1·σ v + α2·σ i + α3·σ T + α4·σ pH + α5·σ q ; wherein, σ v , σ i , σ T , σ pH , σ q represent the standard deviation of voltage, current, temperature, pH, and flow in the sliding window, which is used to reflect the fluctuation degree, α1, α2, α3, α4, α5 represent the disturbance sensitivity weight coefficients of voltage, current, temperature, pH, and flow, which are empirically set values, satisfying α1+α2+α3+α4+α5=1, reflecting the sensitivity of each parameter to disturbance, and Ψ t is the disturbance score index, which is weighted by the standard deviation of multiple parameters at time t, representing the stability of the current process state; (3) Abnormal sign index: Calculate the deviation of each parameter at the end of the sliding window relative to the historical mean Expressed as: in, is the historical average value of a parameter, which serves as a reference value for judging the current degree of deviation. x (t) is the relative deviation of a parameter at time t relative to its historical mean, which is used to identify potential anomalies; When ε x (t) Exceeding the set parameter threshold is considered a potential abnormality sign; The value ranges of each parameter threshold include: 1) Voltage θ v : 0.03-0.08; 2) Current θ i : 0.05-0.10; 3) Temperature θ T : 0.02-0.05; 4) pH valueθ pH : 0.01-0.03; 5) Flow rate θ q : 0.05-0.12; Indicator construction: The extracted state trend indicators, local disturbance scores, and abnormal sign indexes are combined to form an edge feature set, which is expressed as: F t ={μ v ,μ i ,μ T ,μ pH ,μ q ,Ψ t ,ε v ,ε i ,...,ε q}; Among them, F t The edge feature set constructed at time t is sent as input to the big data collaborative reasoning module for further analysis.

[0028] The big data collaborative reasoning module includes: Feature set upload and matching analysis: The edge feature set output by the edge modeling and feature extraction module Upload to the central server, the server in the historical big data sample library Similarity matching analysis is performed in , forming a candidate feature set, and the matching degree is expressed as: ; in, It is the current moment The extracted edge feature set, It is the first in the historical sample library feature samples, is the Euclidean distance, is the feature similarity score (the value range is ), the higher the matching degree, the closer the current running state is to the historical sample; The system selects the similarity higher than the set matching threshold Several samples of , where the matching threshold , which means that the system only retains historical samples that match the current edge feature set by more than 80%, to ensure the accuracy and reliability of the control strategy generation; Multi-source knowledge fusion and strategy generation: The system integrates and analyzes the matched candidate feature sets with historical cases and process knowledge to generate a reference control strategy.

[0029] Multi-source knowledge fusion and strategy generation include: Recommended adjustment range generation: Based on the historical sample data in the candidate feature set, the upper and lower limits of key control parameters are counted to construct the recommended adjustment range under the current operating status ; For each parameter , and its recommended adjustment range is expressed as: ; in, is a parameter The recommended adjustment range is It is the edge feature set in the historical big data sample The matching degree is not less than The candidate sample set, Is a candidate sample The value of this parameter in; Combine the recommended adjustment range of each key parameter and the recommended adjustment range under the current operating status , expressed as: ; Each sub-item They are composed of the minimum and maximum values ​​of the parameter in the candidate samples; Coupling relationship analysis and decoupling suggestions: based on edge feature sets , construct the parameter partial derivative matrix , and obtain a set of control decoupling suggestions based on coupling relationship analysis , parameter partial derivative matrix Expressed as: ; in, is the partial derivative matrix of parameters, if a parameter has high sensitivity to another parameter (i.e. If the sensitivity is high, the system suggests decoupling by adjusting the priority or using an independent control channel to avoid the cumulative interference effect and ensure the stability of the regulation and control; Based on the parameter partial derivative matrix , a logical judgment matrix is constructed, which is expressed as: ; Wherein, is the set coupling threshold, , is the constructed logical judgment matrix, which indicates whether there is a strong coupling relationship between parameter and parameter ; According to the value of the logical judgment matrix , a decoupling suggestion function is defined, and the control suggestion is output, which is expressed as: ; Based on the output control suggestion, a control decoupling suggestion set is obtained, which is expressed as: ; Wherein, is the parameter set to be analyzed; Electroplating beat optimization suggestion: call the beat feedback information in the historical case, and combine the process model (such as cathode current density model, diffusion layer thickness change model, etc.) to infer and generate beat optimization suggestion , the beat optimization suggestion specifically includes: (1) Shorten the electroplating period to improve the production capacity; (2) Add intermediate stirring time to improve the uniformity of the plated layer; (3) Delay the slot change to reduce the disturbance frequency; The optimization suggestion of the beat is output as part of the regulation and control strategy, which is used to guide the operation of the next execution module; Regulation and control reference strategy generation: based on the generated recommended adjustment interval, the control decoupling suggestion set and the beat optimization suggestion result, the regulation and control reference strategy is constructed, which is expressed as: ; Wherein, is the regulation and control reference strategy, is the generated recommended adjustment interval, is the control decoupling suggestion set based on the coupling relationship analysis, is the beat optimization suggestion result.

[0030] The adaptive regulation and control execution module includes: Control instruction generation and control execution: combined with the regulation reference strategy and the current edge feature set, real-time regulation instruction is generated according to the parameter deviation to adjust the gain to generate control instruction and execute; Control feedback and data update: after the regulation execution is completed, the measured response value returned by each execution unit is collected and the execution state, and is fused with the original edge data to update the edge multi-parameter data stream for the next round , which is represented as: ; Wherein, is the original edge data, is the measured feedback value after regulation, indicates the fusion function, and the updated data will be re-input to the perception module to complete the closed loop of the whole adaptive regulation, providing the latest basic data for subsequent real-time monitoring and analysis.

[0031] Control instruction generation and control execution include: Decision fusion and control instruction generation: receive the regulation reference strategy and the current edge feature set , make adaptive regulation decision based on the strategy interval and the current state, and for each parameter , let the current value be , the recommended interval is , and the target value of each parameter control instruction is represented as: ; Wherein, is the target value, which is the midpoint of the recommended interval, is the adaptive adjustment gain, , which is dynamically adjusted according to the deviation degree, is the instruction value output to the control execution unit; Control execution and parameter adjustment: the system distributes each parameter control instruction to the corresponding physical control unit, which specifically includes: (1) electroplating power supply controller: receives , , adjusts voltage and current; (2) electrolyte flow controller: receives , adjusts pump speed and valve opening; (3) stirring frequency adjustment unit: combined with the periodic optimization results in , adjust the stirring rhythm and intensity.

[0032] The present application encompasses any alternatives, modifications, equivalent methods and solutions made to the essence and scope of the present application. In order to make the public have a thorough understanding of the present application, specific details are described in the following preferred embodiments of the present application, and the present application can also be fully understood without the description of these details to those skilled in the art. In addition, in order to avoid unnecessary confusion to the essence of the present application, well-known methods, processes, procedures, elements and circuits, etc. are not described in detail.

[0033] The above is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can also be made, which should be considered as the protection scope of the present application.

Claims

1. An intelligent electroplating parameter adaptive control system based on big data analysis, characterized in that: Includes the following modules: Data acquisition and edge perception module: Sensors are deployed at the electroplating site to collect multi-parameter data, including voltage, current, temperature, pH value, and flow rate. The multi-parameter data is pre-processed based on edge nodes and outputs a time-calibrated and noise-suppressed edge multi-parameter data stream. Edge modeling and feature extraction module: This module performs sliding window modeling and parameter coupling feature extraction on edge multi-parameter data streams to construct an edge feature set including process state trends, local disturbance patterns, and abnormal sign indicators; Big data collaborative reasoning module: Sends edge feature sets to the central server for matching analysis with the historical big data sample library, calls the multi-source case library and process mechanism knowledge library to generate control reference strategies, which include the target parameter adjustment range, control decoupling suggestion set, and electroplating beat optimization suggestions; Adaptive control execution module: Based on the control reference strategy and combined with the current edge feature set, it generates control decision instructions in real time, executes closed-loop adjustments to the electroplating power supply, electrolyte flow and stirring frequency control units, and feeds back the control results to the data acquisition and edge perception module to update the edge multi-parameter data stream.

2. The intelligent electroplating parameter adaptive control system based on big data analysis according to claim 1 is characterized in that: The data acquisition and edge perception module includes: Sensor deployment and data collection: Various industrial-grade sensors, including voltage sensors, current sensors, temperature sensors, pH sensors, and flow sensors, are deployed at the electroplating site to collect multi-parameter data in real time, which is then received and synchronized locally through edge nodes. Data alignment and timing calibration: Time-align asynchronous data collected by different sensors to construct unified multi-parameter time series data; Edge preprocessing output: Filter and correct anomalies on the aligned multi-parameter time series data, remove noise interference, and output a stable edge multi-parameter data stream.

3. The intelligent electroplating parameter adaptive control system based on big data analysis according to claim 2 is characterized in that: The data alignment and timing calibration include: Unified modeling of multi-source data: To address the issue of asynchronous sampling times among different types of sensors, the system uniformly models the acquisition results of each parameter into a time series matrix format. Asynchronous data linear interpolation calibration: In order to eliminate the influence of sampling time difference on parameter comparison and analysis, the system uses linear interpolation algorithm to perform time normalization processing on asynchronous data.

4. The intelligent electroplating parameter adaptive control system based on big data analysis according to claim 2 is characterized in that: The edge preprocessing output includes: Weight function definition and filter control: Using the Gaussian kernel function form, calculate the different weights given to adjacent time points when applying the local weighted regression filter algorithm, that is, the neighborhood weight function; Outlier detection and filtering modeling: Apply the local weighted regression filter algorithm to the data matrix that has completed time series calibration, identify and smooth the short-term fluctuations and outliers in the time series calibration data, and obtain the filtered data as an edge multi-parameter data stream.

5. The intelligent electroplating parameter adaptive control system based on big data analysis according to claim 4 is characterized in that: The edge modeling and feature extraction module includes: Sliding window modeling and local state aggregation: Apply a sliding time window mechanism to the input edge multi-parameter data stream for modeling, combining the multi-parameter data from the current and previous periods to construct a local state set within the time window; Parameter coupling feature extraction and indicator construction: Within each sliding window, the system extracts three types of key features from the edge multi-parameter data stream, including the average change rate of each parameter, the degree of fluctuation of each parameter, and the deviation between the current parameter value and the historical average value, and constitutes the edge feature set.

6. The intelligent electroplating parameter adaptive control system based on big data analysis according to claim 5 is characterized in that: The parameter coupling feature extraction and index construction include: Parameter coupling feature extraction: extract state trend indicators, local disturbance scores and abnormal sign indexes; Indicator construction: The extracted state trend indicators, local disturbance scores, and abnormal sign indexes are combined to form an edge feature set.

7. The intelligent electroplating parameter adaptive control system based on big data analysis according to claim 6 is characterized in that: The big data collaborative reasoning module includes: Feature set upload and matching analysis: The edge feature set output by the edge modeling and feature extraction module is uploaded to the central server. The server performs similarity matching analysis in the historical big data sample library to form a candidate feature set; Multi-source knowledge fusion and strategy generation: The system integrates and analyzes the matched candidate feature sets with historical cases and process knowledge to generate a reference control strategy.

8. The intelligent electroplating parameter adaptive control system based on big data analysis according to claim 7 is characterized in that: The multi-source knowledge fusion and strategy generation include: Recommended adjustment range generation: Based on the historical sample data in the candidate feature set, the upper and lower limits of key control parameters are calculated to construct the recommended adjustment range under the current operating status; Coupling relationship analysis and decoupling suggestions: Based on the edge feature set, a parameter partial derivative matrix is ​​constructed to obtain a set of control decoupling suggestions based on coupling relationship analysis; Electroplating cycle optimization suggestions: Call the cycle feedback information from historical cases and combine it with the process model to generate cycle optimization suggestions through reasoning; Generation of control reference strategy: Based on the generated recommended adjustment interval, control decoupling suggestion set and beat optimization suggestion results, a control reference strategy is constructed.

9. The intelligent electroplating parameter adaptive control system based on big data analysis according to claim 8, characterized in that: The adaptive control execution module includes: Control instruction generation and control execution: Combine the control reference strategy with the current edge feature set to generate control instructions in real time based on parameter deviations, adjust the gain to generate control instructions, and execute them; Control feedback and data update: After the control execution is completed, the measured response value and execution status returned by each execution unit are collected and fused with the original edge data to update it into a new round of edge multi-parameter data stream.

10. The intelligent electroplating parameter adaptive control system based on big data analysis according to claim 9, characterized in that: The control instruction generation and control execution include: Decision fusion and control instruction generation: Receive the control reference strategy and the current edge feature set, and make adaptive control decisions based on the strategy interval and current state; Control execution and parameter adjustment: The system distributes each parameter control instruction to the corresponding physical control unit.

Images