A method for generating anodizing current density adjustment schemes based on mutual information entropy

Abstract

This application relates to the field of process optimization technology, and particularly to a method for generating anodizing current density adjustment schemes based on mutual information entropy. The method includes: acquiring process parameters in real time and performing adaptive filtering to ensure the accurate and effective extraction of key feature parameters, providing a high-quality data foundation for optimization; determining key feature parameters based on the mutual information entropy values ​​of process parameters and preset optimization targets; filtering current density data to be optimized according to the influence weight of key parameters; extracting candidate current density curves matching preset oxidation quality standards from the candidate data, calculating the deviation between the current value and the curve as candidate optimization parameters, and transforming abstract standards into concrete criteria; and generating an adaptive adjustment scheme based on the candidate parameters to dynamically adapt to process changes, avoid fluctuations caused by static settings, achieve high consistency between current density control accuracy and oxide film quality stability, shorten the optimization cycle, and improve production efficiency.

Description

Technical Field

[0001] This application belongs to the field of process optimization technology, and in particular relates to a method for generating anodizing current density adjustment scheme based on mutual information entropy. Background Technology

[0002] In the field of aluminum profile processing, anodizing is a core step in improving the surface properties of aluminum profiles (such as corrosion resistance and decorative properties), and the precise control of its process parameters plays a decisive role in the quality of the final oxide film. With the continuous improvement of the surface quality requirements for aluminum profiles in modern manufacturing, parameter optimization technology for anodizing has become crucial for ensuring product competitiveness.

[0003] Currently, the industry commonly uses a "fixed parameter preset + post-production inspection and correction" model to optimize the current density of aluminum profile anodizing. Before production, a baseline current density value is set based on past experience. During production, parameter stability is monitored only through periodic manual inspections (e.g., recording current values ​​hourly). After each batch is completed, samples are taken to test the oxide film quality. If uneven thickness or insufficient hardness is found, the current density parameters for the next batch are adjusted based on experience. This approach has significant drawbacks: Firstly, current density is strongly coupled with parameters such as electrolyte temperature and concentration. Experience-based fixed parameters cannot dynamically adapt to real-time conditions such as changes in electrolyte composition and variations in aluminum profile loading during production, easily leading to fluctuations in oxide film quality at different locations within the same batch. Secondly, post-production inspection has a significant time lag. Completing one anodizing cycle takes 40-90 minutes; by the time quality problems are discovered, a large number of defective products have already been produced, resulting in waste of aluminum profile raw materials, electrolyte, and energy.

[0004] A more significant problem is that, in actual production, variations in the depth of the aluminum profile immersed in the anodizing tank and differences in local electrolyte flow rates can cause slight fluctuations in the real-time current density. Manual inspections cannot capture the impact of these instantaneous fluctuations on oxide film growth; adjustments can only be made by inferring the direction from the final test results, leading to a highly arbitrary optimization path. Especially when producing high-precision aluminum profiles, it is necessary to maintain parameter stability within an extremely narrow process window. The errors of traditional experience-based adjustment methods far exceed the allowable range, making it difficult to meet quality requirements. Summary of the Invention

[0005] This application provides a method for generating anodizing current density adjustment schemes based on mutual information entropy. This method can solve the problem that in the process of optimizing the current density of aluminum profile anodizing, the adjustment error far exceeds the allowable range due to reliance on traditional experience-based adjustment methods, making it difficult to meet quality requirements.

[0006] In a first aspect, embodiments of this application provide a method for generating anodizing current density adjustment schemes based on mutual information entropy, including:

[0007] The process parameter data of aluminum profiles at each moment during the anodizing process is obtained, and the key feature parameters affecting the current density optimization are extracted by an adaptive filtering algorithm based on the process parameter data at each moment; wherein the key feature parameters are determined based on the mutual information entropy value between the process parameter data and the preset current density optimization target. Based on the key feature parameters, the real-time current density data corresponding to parameters whose influence weight exceeds the preset influence weight are identified as current density data to be optimized. Extract candidate current density curves from the current density data to be optimized that match the preset oxidation quality standard within a preset matching threshold, and identify the deviation values ​​of the current density value corresponding to the current processing position of the aluminum profile from each candidate current density curve as candidate optimization parameters. An adaptive current density adjustment scheme is generated based on the proposed optimization parameters.

[0008] The technical solutions described in this application embodiment have at least the following technical effects: The anodizing current density adjustment scheme generation method based on mutual information entropy provided in this application obtains process parameter data at each moment during the anodizing process of aluminum profiles and processes it through an adaptive filtering algorithm, ensuring the accuracy and effectiveness of key feature parameter extraction and providing a high-quality data foundation for current density optimization. Key feature parameters are determined based on the mutual information entropy value between the process parameter data and the preset current density optimization target, ensuring a strong correlation between the extracted parameters and the current density optimization requirements, and avoiding irrelevant parameters interfering with the optimization direction. The current density data to be optimized is screened according to the influence weight of the key feature parameters, focusing the optimization on the core influencing dimensions, ensuring targeted optimization while reducing redundant calculations and improving optimization efficiency. Candidate current density curves matching the preset oxidation quality standard are extracted from the current density data to be optimized, and the deviation between the current current density value and the curve is calculated as the candidate optimization parameter, transforming the abstract quality standard into a concrete optimization basis and providing clear guidance for the formulation of the adjustment scheme. By combining the candidate optimization parameters to generate an adaptive current density adjustment scheme, it can dynamically adapt to parameter changes during the anodizing process. It effectively combines theoretical optimization logic with actual production scenarios, avoids process fluctuations caused by static parameter settings, achieves high consistency in current density control accuracy and oxide film quality stability, shortens the optimization cycle, and improves production efficiency.

[0009] Secondly, embodiments of this application provide a system for generating anodizing current density adjustment schemes based on mutual information entropy, the system comprising: The acquisition unit is used to acquire process parameter data of aluminum profiles at each moment during the anodizing process, and extract key feature parameters affecting current density optimization based on the process parameter data at each moment through an adaptive filtering algorithm; wherein, the key feature parameters are determined based on the mutual information entropy value between the process parameter data and the preset current density optimization target. The data unit is used to filter out the real-time current density data corresponding to parameters whose influence weight exceeds the preset influence weight based on the key feature parameters and confirm them as current density data to be optimized. The extraction unit is used to extract candidate current density curves from the current density data to be optimized that match the preset oxidation quality standard within a preset matching threshold, and to identify the deviation values ​​of the current density value corresponding to the current processing position of the aluminum profile from each candidate current density curve as candidate optimization parameters. The result unit is used to generate an adaptive current density adjustment scheme based on the candidate optimization parameters.

[0010] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any of the foregoing aspects.

[0011] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer instructions that, when executed on a computer, cause the computer to perform the method described in any of the preceding aspects.

[0012] Fifthly, embodiments of this application provide a computer program product that, when run on an electronic device, causes the electronic device to perform the method described in any of the preceding aspects.

[0013] It is understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant descriptions in the above aspects, and will not be repeated here. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart illustrating a method for generating anodizing current density adjustment schemes based on mutual information entropy, provided in an embodiment of this application. Figure 2 This is a schematic diagram of the operation of the anodizing current density adjustment scheme generation method based on mutual information entropy provided in an embodiment of this application; Figure 3 This is a schematic diagram of signal adjustment for a method for generating anodizing current density adjustment scheme based on mutual information entropy, provided in an embodiment of this application. Figure 4 This is a schematic diagram of the deviation and adjustment amount of the anodizing current density adjustment scheme generation method based on mutual information entropy provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an anodizing current density adjustment scheme generation system based on mutual information entropy provided in an embodiment of this application; Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0016] 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.

[0017] 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.

[0018] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0019] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determination" or "if the described condition or event is detected" may be interpreted, depending on the context, as "once determination," "in response to determination," "once the described condition or event is detected," or "in response to the detection of the described condition or event."

[0020] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] 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.

[0022] In the field of aluminum profile processing, anodizing is a core step in improving the surface properties of aluminum profiles (such as corrosion resistance and decorative properties), and the precise control of its process parameters plays a decisive role in the quality of the final oxide film. With the continuous improvement of the surface quality requirements for aluminum profiles in modern manufacturing, parameter optimization technology for anodizing has become crucial for ensuring product competitiveness.

[0023] Currently, the industry commonly uses a "fixed parameter preset + post-production inspection and correction" model to optimize the current density of aluminum profile anodizing. Before production, a baseline current density value is set based on past experience. During production, parameter stability is monitored only through periodic manual inspections (e.g., recording current values ​​hourly). After each batch is completed, samples are taken to test the oxide film quality. If uneven thickness or insufficient hardness is found, the current density parameters for the next batch are adjusted based on experience. This approach has significant drawbacks: Firstly, current density is strongly coupled with parameters such as electrolyte temperature and concentration. Experience-based fixed parameters cannot dynamically adapt to real-time conditions such as changes in electrolyte composition and variations in aluminum profile loading during production, easily leading to fluctuations in oxide film quality at different locations within the same batch. Secondly, post-production inspection has a significant time lag. Completing one anodizing cycle takes 40-90 minutes; by the time quality problems are discovered, a large number of defective products have already been produced, resulting in waste of aluminum profile raw materials, electrolyte, and energy.

[0024] A more significant problem is that, in actual production, variations in the depth of the aluminum profile immersed in the anodizing tank and differences in local electrolyte flow rates can cause slight fluctuations in the real-time current density. Manual inspections cannot capture the impact of these instantaneous fluctuations on oxide film growth; adjustments can only be made by inferring the direction from the final test results, leading to a highly unreliable optimization path. Especially when producing high-precision aluminum profiles, parameters need to be kept stable within an extremely narrow process window (e.g., current density stable at 1.8±0.1A / dm²). The errors of traditional experience-based adjustment methods far exceed the allowable range, making it difficult to meet quality requirements.

[0025] To address the aforementioned issues, this application provides a method for generating anodizing current density adjustment schemes based on mutual information entropy. This method acquires process parameter data at each moment during the anodizing process of aluminum profiles and processes it using an adaptive filtering algorithm, ensuring the accuracy and effectiveness of key feature parameter extraction and providing a high-quality data foundation for current density optimization. Key feature parameters are determined based on the mutual information entropy value between the process parameter data and the preset current density optimization target, ensuring a strong correlation between the extracted parameters and the current density optimization requirements, and avoiding irrelevant parameters interfering with the optimization direction. The current density data to be optimized is screened according to the influence weight of the key feature parameters, focusing the optimization on the core influencing dimensions, ensuring targeted optimization while reducing redundant calculations and improving optimization efficiency. Candidate current density curves matching the preset oxidation quality standard are extracted from the current density data to be optimized, and the deviation between the current current density value and the curve is calculated as the candidate optimization parameter, transforming the abstract quality standard into a concrete optimization basis and providing clear guidance for formulating the adjustment scheme. By combining the candidate optimization parameters to generate an adaptive current density adjustment scheme, it can dynamically adapt to parameter changes during the anodizing process. It effectively combines theoretical optimization logic with actual production scenarios, avoids process fluctuations caused by static parameter settings, achieves high consistency in current density control accuracy and oxide film quality stability, shortens the optimization cycle, and improves production efficiency.

[0026] The method for generating anodizing current density adjustment scheme based on mutual information entropy provided in this application embodiment can be applied to electronic devices. In this case, the electronic device is the execution subject of the method for generating anodizing current density adjustment scheme based on mutual information entropy provided in this application embodiment. This application embodiment does not impose any restrictions on the specific type of electronic device.

[0027] It is understandable that electronic devices can be various intelligent devices. For example, electronic devices can be terminal devices such as laptops, ultra-mobile personal computers (UMPCs), netbooks, personal digital assistants (PDAs), and desktop computers.

[0028] To better understand the method for generating anodizing current density adjustment scheme based on mutual information entropy provided in the embodiments of this application, the specific implementation process of the method for generating anodizing current density adjustment scheme based on mutual information entropy provided in the embodiments of this application will be described by way of example below.

[0029] Figure 1 A schematic flowchart of the anodizing current density adjustment scheme generation method based on mutual information entropy provided in this application is shown. Figure 2This paper illustrates a flowchart of the method for generating anodizing current density adjustment schemes based on mutual information entropy, as provided in an embodiment of this application. The method includes: S100: Obtain process parameter data of aluminum profiles at each moment during the anodizing process, and extract key feature parameters affecting current density optimization through an adaptive filtering algorithm based on the process parameter data at each moment; wherein, the key feature parameters are determined based on the mutual information entropy value between the process parameter data and the preset current density optimization target.

[0030] It can be understood that the process parameter data of the aluminum profile anodizing process refers to various physical and chemical parameters generated in real time during the anodizing process, covering electrolyte temperature, electrolyte concentration, tank voltage, processing time, aluminum profile surface temperature, etc. An adaptive filtering algorithm is an algorithm that can dynamically adjust the filtering coefficient according to the input signal to achieve signal optimization processing. Key feature parameters refer to process parameter characteristics that have a significant impact on current density optimization. A preset current density optimization target refers to a pre-set, desired current density standard, such as maintaining the current density stable within a certain range to ensure oxide film uniformity. Mutual information entropy is an indicator that quantifies the degree of correlation between two variables and can be used to measure the nonlinear relationship between features and optimization targets. By acquiring the process parameter data at each moment during the anodizing process, and then processing this data using an adaptive filtering algorithm, key feature parameters can be extracted. During the extraction process, the mutual information entropy value between the process parameter data and the preset current density optimization target is calculated to determine which feature parameters are key. Process parameter data can be collected in real time by sensors and stored as a time series dataset; the dataset is input into an adaptive filtering algorithm, which iteratively adjusts the parameters to separate the effective signal; the mutual information entropy value between each feature and the optimization target is calculated for the separated signal, and key feature parameters are selected based on the entropy value.

[0031] In one possible implementation, S100, process parameter data of the aluminum profile at each moment during the anodizing process is acquired, and based on the process parameter data at each moment, key feature parameters affecting current density optimization are extracted using an adaptive filtering algorithm, including: S110: Obtain the process parameter data of the aluminum profile at each moment during the anodizing process, and preprocess the process parameter data to obtain the first process parameter sequence.

[0032] It is understood that process parameter data specifically includes real-time monitoring data such as electrolyte temperature (unit: °C), electrolyte concentration (unit: g / L), applied voltage (unit: V), processing time (unit: min), and aluminum profile immersion depth (unit: cm). Preprocessing refers to the operation of processing the raw data to eliminate interference and unify the format. The first process parameter sequence refers to the set of process parameters with a unified format and temporal characteristics formed after preprocessing. The first process parameter sequence can be formed by acquiring the raw process parameter data at each moment in the anodizing process, performing preprocessing operations on it, and finally forming the first process parameter sequence. Raw data can be collected by sensors distributed at key locations in the oxidation tank; preprocessing may include removing outliers (such as data exceeding the reasonable range caused by sensor failure), filling missing values ​​(such as using interpolation of data from adjacent moments), and data normalization (such as converting parameters of different dimensions to the [0,1] interval); the processed data is arranged in chronological order to form the first process parameter sequence.

[0033] S120, the first process parameter sequence is separated by an adaptive filtering algorithm to determine the second process parameter sequence.

[0034] Adaptive filtering algorithms are primarily used to distinguish between effective and noise components in a signal. Their core principle is to dynamically adjust the filter coefficients to make the output signal closer to the desired signal. The first process parameter sequence is the pre-processed raw process parameter time series data, which may contain noise interference. Signal separation refers to the process of separating the effective signal related to current density optimization from the irrelevant noise signal in the sequence. The second process parameter sequence refers to the effective process parameter time series data after signal separation, removing noise interference. An adaptive filtering algorithm can be used to process the first process parameter sequence, separating the effective signal from the noise signal, and determining the effective signal portion as the second process parameter sequence. The first process parameter sequence is input into the adaptive filtering algorithm; processing begins with initially set filter coefficients, iteratively updating the coefficients to minimize the error between the output and the desired signal; when the error reaches a preset range, iteration stops, and the output signal at this point is the second process parameter sequence.

[0035] Optionally, in step S120, the first process parameter sequence is separated using an adaptive filtering algorithm to determine the second process parameter sequence, including: S121, Obtain the parameter sequence of aluminum profile anodizing under historical stable process conditions, and use the parameter sequence as the expected signal sequence; the expected signal sequence contains the process parameter fluctuation characteristics corresponding to qualified oxidation quality.

[0036] It can be understood that historically stable process conditions refer to the production conditions under which the anodizing quality of aluminum profiles was qualified and the process was stable in the past, such as specific electrolyte ratios, temperature ranges, and voltage curves. A parameter sequence refers to the set of various process parameter data recorded in chronological order under historically stable process conditions. A desired signal sequence refers to the signal sequence that serves as a reference standard for adaptive filtering algorithms, guiding the algorithm to separate effective signals. The process parameter fluctuation characteristics corresponding to qualified anodizing quality refer to the regularity and range of process parameter changes over time when the anodizing quality meets the standards, such as temperature fluctuations within ±2℃ and voltage stabilization around a certain value. Please refer to [link / reference]. Figure 4 It can obtain the anodizing parameter sequence under historically stable process conditions and set it as the desired signal sequence. The desired signal sequence contains the process parameter fluctuation characteristics corresponding to qualified oxidation quality.

[0037] For example, process parameter records of batches that have passed oxidation quality testing are selected from the production database; these parameters are then organized in chronological order to form a continuous parameter sequence; this parameter sequence is determined as the desired signal sequence for reference in subsequent adaptive filtering algorithms. S122, the first process parameter sequence is used as the input signal sequence of the adaptive filtering algorithm, and the filtering coefficient vector is initialized; the dimension of the filtering coefficient vector matches the number of parameter types in the first process parameter sequence.

[0038] The input signal sequence refers to the original signal processed by the adaptive filtering algorithm, which in this case is the first process parameter sequence. The filter coefficient vector is a set of parameters in the adaptive filtering algorithm, used to weight the input signal to generate the output signal. The number of parameter types refers to the number of different types of process parameters contained in the first process parameter sequence, such as temperature, concentration, and voltage, which are counted as three types. The first process parameter sequence is used as the input to the adaptive filtering algorithm to initially set the filter coefficient vector in the algorithm, and the dimension of the vector is consistent with the number of parameter types in the first process parameter sequence.

[0039] For example, the first process parameter sequence can be decomposed into multiple subsequences according to parameter type, and the number of parameter types is determined to be n; an n-dimensional filter coefficient vector is initialized, and the initial value can be set to a uniform value (such as all of 1 / n) or an initial weight set according to experience; this vector is used as the initial parameter of the algorithm.

[0040] S123, set the convergence factor according to the characteristics of the input signal sequence, calculate the filtered output signal and error signal at the current time based on the minimum mean square error criterion, and update the filter coefficient vector through iteration until the error signal is less than the preset error threshold or the number of iterations reaches the preset upper limit of the number of iterations; the error signal is the deviation between the desired signal sequence and the filtered output signal.

[0041] The convergence factor is a parameter in the adaptive filtering algorithm, used to control the update speed of the filter coefficient vector. Its value is related to the fluctuation characteristics of the input signal; a smaller value can be set to ensure stability when the fluctuation is large. The minimum mean square error criterion refers to the principle of adjusting the filter coefficients by minimizing the sum of squared errors between the desired signal and the output signal. The filtered output signal is the signal obtained after the input signal has been weighted by the filter coefficients. The error signal is the difference between the desired signal sequence and the filtered output signal, reflecting the degree of deviation between the output signal and the desired signal. The preset error threshold is a pre-set, acceptable upper limit value for error. The preset upper limit for the number of iterations is the maximum number of iterations set to avoid infinite iteration of the algorithm. The convergence factor can be determined based on the characteristics of the input signal sequence. According to the minimum mean square error criterion, the current filtered output signal and error signal are calculated. The filter coefficient vector is updated through multiple iterations until the error signal is less than the preset threshold or the number of iterations reaches the upper limit.

[0042] For example, the fluctuation amplitude and frequency of the input signal sequence can be analyzed, and an appropriate convergence factor (such as 0.01) can be set. In each iteration, the input signal is weighted using the current filter coefficient vector to obtain the output signal. The error signal between the output signal and the desired signal is calculated. The filter coefficient vector is updated according to the error signal and the convergence factor. The iteration process is repeated until the stopping condition is met.

[0043] S124, the filtered output signal sequence after iterative convergence is confirmed as the second process parameter sequence; the second process parameter sequence is the effective process parameter change sequence after removing noise interference, which contains signal components that are strongly correlated with current density optimization.

[0044] Iterative convergence can be understood as the state where the filter coefficient vector is updated to no longer change significantly, and the error signal stabilizes within a small range. The filtered output signal sequence refers to the signal sequence output by the adaptive filtering algorithm after iterative convergence. The second process parameter sequence is the effective process parameter time-series data after denoising. The second process parameter sequence is the effective process parameter change sequence after removing noise interference, containing signal components strongly correlated with current density optimization. Noise interference refers to interference components in the original signal that are unrelated to the essence of the process, such as sensor random errors and power grid fluctuations. The effective process parameter change sequence refers to the parameter sequence that can reflect the actual changes in the process, and the signal components strongly correlated with current density optimization refer to parameter change information that has a significant impact on current density adjustment. When iterative convergence is achieved, the filtered output signal sequence at this point is determined as the second process parameter sequence, which is the effective sequence after removing noise and contains signals strongly correlated with current density optimization.

[0045] S130, a multi-dimensional feature pool is constructed based on the second process parameter sequence to obtain the key feature parameters that affect the optimization of current density.

[0046] As can be understood, a multi-dimensional feature pool refers to a collection containing features of various types and dimensions, extracted from the second process parameter sequence. Key feature parameters refer to those selected from the feature pool that significantly impact current density optimization. Based on the second process parameter sequence, features of multiple dimensions can be extracted to form a feature pool, from which key feature parameters affecting current density optimization can then be selected.

[0047] For example, statistical features and time-series features are extracted for each parameter in the second process parameter sequence; these features of all parameters are summarized to form a multi-dimensional feature pool; and key feature parameters are selected by calculating the correlation between the features and the current density optimization target.

[0048] Optionally, in S130, a multi-dimensional feature pool is constructed based on the second process parameter sequence to obtain key feature parameters affecting current density optimization, including: S131, for each parameter in the second process parameter sequence, extract multi-dimensional features; the multi-dimensional features include statistical features and time series features, wherein the statistical features include the central tendency, dispersion and extreme value features of the parameter within a preset time window, and the time series features include the change trend, fluctuation frequency and correlation features of adjacent time points of the parameter.

[0049] It can be understood that each parameter in the second process parameter sequence refers to the time-series data of a single process parameter, such as temperature or concentration, contained in the sequence. Multi-dimensional features refer to a set of features describing the characteristics of a parameter from different perspectives. Statistical features are features extracted based on statistical methods, reflecting the distribution characteristics of a parameter within a certain time range. A preset time window refers to a pre-defined time interval (e.g., 5 minutes) used to calculate statistical features. Central tendency refers to the degree of concentration of parameter values ​​within the time window, such as the mean and median. Dispersion refers to the dispersion of parameter values, such as variance and standard deviation. Extreme value features refer to the maximum, minimum, and outlier values ​​within the time window. Time-series features reflect the pattern of parameter changes over time. The trend of change refers to the direction and rate of increase or decrease of parameter values ​​over time, such as slope and trend term. Fluctuation frequency refers to the frequency of periodic changes in parameter values. Correlation features between adjacent time points refer to the relationship between parameter values ​​at adjacent time points, such as the autocorrelation coefficient. For each parameter in the second process parameter sequence, multi-dimensional features, including statistical features and time-series features, are extracted. The statistical features cover the central tendency, dispersion and extreme value features within the preset time window, while the time-series features cover the change trend, fluctuation frequency and correlation features between adjacent time points.

[0050] For example, for the time series data of each parameter, multiple preset time windows (such as 1 min, 5 min) are pre-set; within each window, the mean, median (central tendency), variance, standard deviation (dispersion), maximum value, minimum value (extreme value feature) are calculated; the slope (change trend) is calculated through linear fitting, the fluctuation frequency is obtained through Fourier transform, and the difference or correlation coefficient of adjacent time data (adjacent association feature) is calculated; these features are summarized as the multi-dimensional features of the parameter.

[0051] S132, based on the multi-dimensional features of all parameters, forms a multi-dimensional feature pool containing features of multiple types and scales.

[0052] It can be understood that the multi-dimensional features of all parameters refer to the sum of statistical features, temporal features, etc., extracted from each parameter in the second process parameter sequence. Multi-type features refer to the diversity of feature types, including both statistical and temporal features. Multi-scale features refer to features calculated based on different time windows (e.g., 1 min, 5 min, 10 min), reflecting the parameter characteristics at different time scales. The multi-dimensional feature pool is a feature set formed by integrating the multi-type and multi-scale features of all parameters.

[0053] For example, the multi-dimensional features of all parameters in the second process parameter sequence can be integrated to form a multi-dimensional feature pool containing features of various types and multiple time scales. Statistical and temporal features of each parameter under different time windows can be extracted, and these features can be classified and labeled (e.g., labeled "temperature-5min-mean" and "voltage-10min-slope"). All labeled features can then be summarized to form a multi-dimensional feature pool.

[0054] S133, based on the multi-dimensional feature pool, each parameter in the second process parameter sequence is screened to obtain the key feature parameters that affect the current density optimization.

[0055] Screening, as we understand it, refers to the process of selecting parameters with significant impact from all parameters based on their correlation with the current density optimization objective. Key feature parameters refer to the process parameters that, after screening, have a significant impact on current density optimization. The feature information in the multi-dimensional feature pool can be used to evaluate and screen each parameter in the second process parameter sequence, ultimately obtaining the key feature parameters affecting current density optimization.

[0056] For example, the correlation between each feature in the multidimensional feature pool and the current density optimization target can be analyzed; each parameter can be comprehensively scored based on the correlation; and the parameters with higher scores can be selected as key feature parameters.

[0057] For example, in step S133, each parameter in the second process parameter sequence is filtered based on a multi-dimensional feature pool to obtain key feature parameters affecting current density optimization, including: S1331, calculate the mutual information entropy between each feature in the multidimensional feature pool and the preset current density optimization target; the mutual information entropy is used to quantify the degree of nonlinear correlation between the feature and the optimization target, and the higher the entropy value, the more significant the influence of the feature on the optimization target.

[0058] It can be understood that the preset current density optimization objective is a pre-defined ideal state of current density (e.g., stable at 2A / dm²). Mutual information entropy is a concept in information theory used to measure the dependency between two random variables, with a value ranging from [0,1]. The degree of nonlinear correlation refers to the tightness of the nonlinear relationship between features and the optimization objective. The degree of nonlinear correlation between the two can be quantified by calculating the mutual information entropy between each feature in the multidimensional feature pool and the preset current density optimization objective; a higher entropy value indicates a more significant impact of the feature on the optimization objective.

[0059] For example, the value of each feature and the value of the preset current density optimization target (or the target state) can be used as two variables; the mutual information entropy value of the two can be calculated using a mutual information entropy calculation function (such as a histogram-based estimation method); the entropy value corresponding to each feature can be recorded for subsequent screening.

[0060] S1332, based on mutual information entropy, filters each feature in the multi-dimensional feature pool to obtain key feature parameters that affect current density optimization.

[0061] Understandably, features in a multi-dimensional feature pool can be filtered based on the mutual information entropy value of each feature, retaining those with high correlation. The corresponding process parameters are then the key feature parameters affecting current density optimization. For example, a mutual information entropy filtering criterion can be pre-defined (e.g., retaining the top 30% of features by entropy value); features meeting the criteria can be selected based on the criterion; and the process parameters corresponding to these features can be determined as the candidate set of key feature parameters.

[0062] For example, in S1332, each feature in the multi-dimensional feature pool is filtered based on mutual information entropy to obtain key feature parameters affecting current density optimization, including: S13321, Select features whose mutual information entropy value is not less than the preset mutual information entropy threshold, and confirm the parameters corresponding to the selected features as initial candidate feature parameters.

[0063] Mutual information entropy is understood to be a quantified value of the degree of correlation between a feature and the optimization objective. A preset mutual information entropy threshold is a pre-defined critical value (e.g., 0.6) used to determine whether a feature has sufficient correlation. Selected features are those whose mutual information entropy values ​​reach or exceed the threshold. Initial candidate feature parameters are the process parameters corresponding to the selected features; they are preliminary candidates for key feature parameters. Features with mutual information entropy values ​​not lower than the preset threshold can be selected, and their corresponding process parameters can be determined as initial candidate feature parameters. For example, based on the preset mutual information entropy threshold (e.g., determined by historical data statistics), all features in the multidimensional feature pool are traversed, and their mutual information entropy values ​​are compared with the threshold; features with qualifying entropy values ​​are retained, and their corresponding process parameters are recorded; these process parameters are then summarized to form the initial candidate feature parameter set.

[0064] S13322, Redundancy check is performed on the initial candidate feature parameters. The mutual information entropy between any two initial candidate feature parameters is calculated. If the mutual information entropy between the two exceeds the preset redundancy threshold, feature redundancy is determined to exist, and the feature parameter with higher mutual information entropy than the optimization target is retained.

[0065] It is understandable that initial candidate feature parameters are process parameters that have been preliminarily screened and may affect current density optimization. Redundancy testing refers to the process of checking whether there is information duplication or high correlation between candidate parameters. The mutual information entropy between any two initial candidate feature parameters is an indicator of the degree of correlation between these two parameters; the higher the value, the stronger the correlation. The preset redundancy threshold is a critical value (e.g., 0.8) for judging whether two parameters have redundancy. Feature redundancy refers to the phenomenon that the information carried by two parameters is highly overlapping, and retaining one of them can reflect the relevant information. Redundancy testing can be performed on the initial candidate feature parameters by calculating the mutual information entropy between any two parameters. If the entropy value exceeds the preset redundancy threshold, redundancy is determined to exist, and the parameter with the higher mutual information entropy to the optimization target is retained.

[0066] For example, the initial candidate feature parameters can be combined in pairs; the mutual information entropy between each pair of parameters can be calculated; the calculation result can be compared with a preset redundancy threshold. If the threshold is exceeded, the mutual information entropy values ​​of the two parameters with the optimization target can be compared; the parameters with higher entropy values ​​can be retained, and the parameters with lower entropy values ​​can be removed.

[0067] S13323 identifies the characteristic parameters retained after redundancy testing as key characteristic parameters affecting current density optimization.

[0068] It is understandable that the feature parameters that are retained after the redundancy test can be identified as the key feature parameters affecting the current density optimization. The feature parameters that are not eliminated after the redundancy test are finally confirmed to ensure that their mutual information entropy values ​​meet the standards and there is no obvious redundancy. These parameters are then formally identified as key feature parameters.

[0069] S200: Based on key feature parameters, real-time current density data corresponding to parameters whose influence weight exceeds the preset influence weight are identified as current density data to be optimized.

[0070] It can be understood that influence weight is an indicator that measures the degree of influence of key feature parameters on current density optimization, and is determined based on the correlation between the parameter and the optimization objective (such as mutual information entropy). The preset influence weight is a pre-set critical value used to determine whether the parameter's influence is sufficiently large. Real-time current density data refers to the current density value measured in real time during the anodizing process. Current density data to be optimized refers to the real-time current density data that needs adjustment and optimization. Based on the influence weight of the key feature parameters, key feature parameters with influence weights greater than the preset value can be selected, and the real-time current density data corresponding to these key feature parameters can be identified as the current density data to be optimized.

[0071] For example, each key feature parameter can be assigned an influence weight based on the mutual information entropy value (such as the value after entropy normalization); a preset influence weight threshold (such as 0.5) can be set; key feature parameters whose influence weight exceeds the threshold can be selected; and the real-time current density data corresponding to these parameters can be extracted and identified as the current density data to be optimized.

[0072] In one possible implementation, S200, real-time current density data corresponding to parameters whose influence weight exceeds a preset influence weight are selected based on key feature parameters and confirmed as current density data to be optimized, including: S210, determine the influence weight of each parameter in the key feature parameters based on the mutual information entropy value with the preset current density optimization target, screen out the strong influence feature parameters whose influence weight is not less than the preset influence weight threshold, and extract the real-time current density data corresponding to the strong influence feature parameters to form the current density data confirmed as to be optimized.

[0073] It can be understood that the influence weight of each key feature parameter refers to the degree of influence of a single key feature parameter on the current density optimization target, which is obtained by converting the mutual information entropy value between the parameter and the preset optimization target (e.g., using the entropy value after normalization as the weight). The preset influence weight threshold is a critical value used to distinguish between parameters with strong and weak influence. A key feature parameter with strong influence refers to a key feature parameter whose influence weight reaches or exceeds the preset threshold. The real-time current density data corresponding to a key feature parameter with strong influence refers to the real-time current density measurement value associated with the key feature parameter with strong influence. The current density data to be optimized refers to the set of current density data that needs to be optimized and adjusted. First, the influence weight of each key feature parameter based on the mutual information entropy value is determined. Then, key feature parameters with influence weights not lower than the preset threshold are selected. Finally, the real-time current density data corresponding to these key feature parameters are extracted and identified as the current density data to be optimized.

[0074] For example, the mutual information entropy value between each key feature parameter and the optimization target is normalized (e.g., divided by the maximum entropy value) to obtain the influence weight; a preset influence weight threshold (e.g., 0.6) is set; parameters with influence weight ≥ the threshold are selected as strong influence feature parameters; current density data corresponding to these strong influence parameters are extracted from real-time monitoring data and summarized to form the current density data to be optimized.

[0075] S300 extracts candidate current density curves from the current density data to be optimized that match the preset oxidation quality standard within a preset matching threshold, and identifies the deviation values ​​of the current density value corresponding to the current processing position of the aluminum profile from each candidate current density curve as candidate optimization parameters.

[0076] It can be understood that the current density data to be optimized is the real-time current density data that needs to be optimized. The preset oxidation quality standard is a pre-defined range and variation pattern of current density for which oxidation quality is acceptable (e.g., the current density curve for achieving the required oxide film thickness). The matching degree measures the degree of conformity between the current density data to be optimized and the preset standard; a higher value indicates a better conformity. The preset matching threshold is the critical value for determining whether the data meets the standard (e.g., 0.8). The candidate current density curve refers to the current density variation curve within the threshold range of the matching degree with the preset standard. The current processing position of the aluminum profile refers to the current position of the aluminum profile in the anodizing bath (e.g., immersion depth, processing stage). The deviation value of the current density value from the curve refers to the difference between the current measured current density value and the theoretical value of the candidate curve at the corresponding moment. The candidate optimization parameters refer to the deviation values ​​used for subsequent optimization and adjustment. Candidate current density curves within the preset threshold range of the matching degree with the preset oxidation quality standard can be extracted from the current density data to be optimized. The deviation values ​​between the current density value at the current processing position of the aluminum profile and each candidate curve can be calculated, and these deviation values ​​can be determined as candidate optimization parameters.

[0077] For example, multiple candidate curves can be obtained by curve fitting of the current density data to be optimized; the matching degree between each candidate curve and the preset oxidation quality standard curve can be calculated (e.g., by curve similarity algorithm); curves with matching degree ≥ preset threshold can be selected as candidate current density curves; the time corresponding to the current processing position of the aluminum profile can be obtained, and the real-time current density value at that time can be read; the deviation (e.g., absolute difference, relative difference) between this value and the theoretical value of each candidate curve at the same time can be calculated; these deviation values ​​can be used as candidate optimization parameters.

[0078] In one possible implementation, S300, candidate current density curves with a matching degree within a preset matching threshold to the current density data to be optimized are extracted, and the deviation values ​​of the current density value corresponding to the current processing position of the aluminum profile from each candidate current density curve are identified as candidate optimization parameters, including: S310: The matching degree between the current density data to be optimized and the preset oxidation quality standard is calculated using a feature comparison algorithm. A feature comparison algorithm is used to compare the similarity between two sets of data or features (such as dynamic time warping algorithms, cosine similarity algorithms, etc.). The current density data to be optimized is the time-series current density data that needs to be optimized. The preset oxidation quality standard is a standard that includes current density features corresponding to qualified oxidation quality (such as feature points and trends of a standard current density curve). The matching degree is a quantitative value (usually ranging from [0,1]) that measures the similarity between the data to be optimized and the preset standard. The feature comparison algorithm can be used to compare the current density data to be optimized and the preset oxidation quality standard to calculate their matching degree.

[0079] For example, features (such as feature points, slope changes, and fluctuation ranges) of the current density data to be optimized can be extracted; corresponding features of the preset oxidation quality standard can be extracted; the two sets of features can be input into a feature comparison algorithm (such as a dynamic time warping algorithm to calculate the distance and convert it into a matching degree); and the matching degree value between the data to be optimized and the preset standard can be obtained.

[0080] S320 filters out current density data that meet or exceed a preset matching threshold and generates candidate current density curves through data fitting.

[0081] It can be understood that current density data to be optimized that reaches or exceeds a preset matching threshold refers to data that highly conforms to preset oxidation quality standards. The preset matching threshold is the critical value for determining whether data meets the standard. Data fitting refers to the process of fitting discrete data into a continuous curve using mathematical methods (such as least squares or polynomial fitting). The candidate current density curve refers to the current density change curve obtained based on the fitted data after screening, which may serve as an optimization reference. Current density data to be optimized that reaches or exceeds the preset threshold can be selected, and data fitting can be performed on these data to generate candidate current density curves.

[0082] For example, a matching threshold (e.g., 0.75) can be preset; data with a matching degree greater than or equal to the threshold can be selected from the current density data to be optimized; the selected data can be arranged in chronological order, and a suitable fitting method (e.g., cubic spline fitting) can be used to perform curve fitting; multiple candidate current density curves can be generated (e.g., fitting based on data from different time periods).

[0083] S330: Obtain the real-time current density measurement value at the current processing position of the aluminum profile, calculate the deviation between the real-time current density measurement value and the theoretical value of each candidate current density curve at the corresponding time, and confirm the deviation value as the candidate optimization parameter. It can be understood that the real-time current density measurement value at the current processing position of the aluminum profile refers to the current density value measured in real time at the current processing position of the aluminum profile (such as a certain depth or a certain processing stage). The theoretical value of each candidate current density curve at the corresponding time refers to the theoretical current density value of the candidate curve at the time point corresponding to the current processing position. The deviation value is the difference between the real-time measurement value and the theoretical value (e.g., ΔI = measured value - theoretical value). The candidate optimization parameter refers to the quantified deviation value used for subsequent adjustment of the current density.

[0084] For example, the current processing position of the aluminum profile can be determined by a position sensor and associated with the corresponding processing time; the real-time current density measurement value at that time can be read; the theoretical current density value at that time can be extracted from each candidate current density curve; the deviation between the measured value and each theoretical value (such as absolute deviation and relative deviation) can be calculated; and these deviation values ​​can be recorded as candidate optimization parameters.

[0085] S400 generates an adaptive current density adjustment scheme based on the selected optimization parameters.

[0086] It can be understood that the candidate optimization parameters are quantified values ​​reflecting the deviation between the current current density and the candidate curve. An adaptive current density adjustment scheme refers to a strategy that can automatically adjust the current density based on real-time parameter changes, exhibiting dynamic adaptability. This step can be summarized as: based on the candidate optimization parameters, developing an adaptive scheme that can dynamically adjust the current density.

[0087] For example, the adjustment direction and magnitude of the current density can be determined by analyzing the magnitude and direction of the candidate optimization parameters (such as positive deviation and negative deviation); combining process experience and constraints; formulating rules for dynamic adjustment as the deviation changes (such as the larger the deviation, the larger the adjustment magnitude); and integrating these rules into an adaptive current density adjustment scheme.

[0088] In one possible implementation, S400 generates an adaptive current density adjustment scheme based on the selected optimization parameters, including: S410, input the candidate optimization parameters into the preset adjustment mapping model to obtain the current density adjustment amount of the aluminum profile during the anodizing process; wherein, the adjustment mapping model is a machine learning model that has been pre-trained.

[0089] It can be understood that the candidate optimization parameters are the deviation values ​​used as model inputs. The preset adjustment mapping model refers to a pre-built and trained model used to map the input parameters to current density adjustment amounts. The current density adjustment amount refers to the value that needs to be adjusted for the current density (e.g., +0.2A / dm², -0.1A / dm²). The pre-trained machine learning model refers to an algorithmic model (e.g., neural network, random forest, etc.) trained on historical data, capable of learning the mapping relationship between input and output. This step can be summarized as follows: the candidate optimization parameters are input into the pre-trained adjustment mapping model, and the model outputs the current density adjustment amount during the aluminum profile anodizing process. In specific implementation, the candidate optimization parameters are preprocessed according to the format required by the model (e.g., normalization); the preprocessed parameters are input into the adjustment mapping model (e.g., a trained neural network); the model outputs the corresponding current density adjustment amount through internal calculation; and this adjustment amount is recorded for subsequent steps.

[0090] S420 verifies the rationality of the current density adjustment based on the preset process constraints of aluminum profile anodizing, formulates dynamic adjustment rules according to the process sequence, and determines the dynamic adjustment rules as an adaptive current density adjustment scheme.

[0091] It is understandable that the preset process constraints for aluminum profile anodizing refer to the limitations that must be followed during the anodizing process, such as the upper limit of current density (to avoid ablation), the lower limit (to ensure oxide film formation), and the adjustment rate limit (to avoid drastic changes). Rationality verification refers to the process of checking whether the current density adjustment meets the process constraints. Process sequence refers to the time sequence and stage division of the anodizing process (such as the heating stage, the isothermal stage, and the cooling stage). Dynamic adjustment rules refer to rules that allow for dynamic adjustment of current density based on different process stages and real-time parameter changes. The rationality of the current density adjustment can be checked based on the preset process constraints, and dynamic adjustment rules can be formulated based on the process sequence, defining these rules as an adaptive current density adjustment scheme.

[0092] For example, the current density adjustment amount can be compared with process constraints (such as current density range of 0.5-3A / dm²), and the adjustment amount can be corrected if it exceeds the range; adjustment rules can be formulated according to the process sequence (such as different processing stages) (such as allowing a larger adjustment range in the heating stage and a smaller adjustment range in the isothermal stage); the verified adjustment amount and the rules of each stage can be integrated to form an adaptive current density adjustment scheme.

[0093] In one possible implementation, before inputting the candidate optimization parameters into a preset adjustment mapping model, the method further includes: S510: Obtain the sample optimization parameters and corresponding current density adjustment and oxidation quality results data from historical processes.

[0094] It can be understood that "historical process" refers to past aluminum profile anodizing production processes. "Sample candidate optimization parameters" refers to candidate optimization parameters (such as historical deviation values) extracted from historical processes and used as training samples for the model. "Corresponding current density adjustment" refers to the actual current density adjustment used for these sample parameters historically. "Oxidation quality result data" refers to the oxidation quality test results of the aluminum profiles after applying the corresponding adjustment (such as oxide film thickness, hardness, uniformity, etc.). This can be achieved through the sample candidate optimization parameters from the historical process, along with the corresponding current density adjustment and oxidation quality result data.

[0095] For example, please refer to Figure 4 This can be achieved by filtering complete historical process records from the production database; extracting candidate optimization parameters (such as historical deviation values) for each process batch from the records; extracting the actual current density adjustment amount used in the corresponding batch; extracting the oxidation quality test data of the aluminum profiles in that batch (such as various indicators in the test report); and associating these three types of data by batch to form a training sample set.

[0096] S520, based on the oxidation quality result data, determines the optimal current density adjustment amount corresponding to this candidate optimization parameter from the corresponding current density adjustment amount.

[0097] It can be understood that oxidation quality result data reflects the detection indicators (such as pass / fail, specific numerical values) of aluminum profile oxidation quality. The corresponding current density adjustment amount refers to multiple possible adjustment amounts historically used for a specific sample parameter. The optimal current density adjustment amount corresponding to the sample candidate optimization parameter refers to the current density adjustment amount that produces the best oxidation quality result in historical data. This step can be summarized as follows: based on the oxidation quality result data, select the adjustment amount that produces the best oxidation quality from multiple current density adjustment amounts corresponding to the sample candidate optimization parameter, and use it as the optimal adjustment amount for that sample parameter. In practice, for each sample candidate optimization parameter, collect all corresponding current density adjustment amounts and the corresponding oxidation quality results; score the oxidation quality results (e.g., 100 points for pass, 0 points for fail, or score according to the indicator value); find the current density adjustment amount corresponding to the oxidation quality result with the highest score; and determine this adjustment amount as the optimal current density adjustment amount for that sample candidate optimization parameter.

[0098] The S530 uses the sample candidate optimization parameters as input and the corresponding optimal current density adjustment amount as the expected output to train the initial machine model, thus obtaining the trained adjustment mapping model.

[0099] It can be understood that the sample candidate optimization parameters are the training sample data used as input to the model. The corresponding optimal current density adjustment is the target value expected to be output during model training. The initial machine model refers to an untrained machine learning model (such as a neural network with initial weights) used to learn the mapping relationship between input and output. The adjusted mapping model (post-training) refers to a model that has been trained to output a suitable current density adjustment based on the input candidate optimization parameters. This step can be summarized as follows: using the sample candidate optimization parameters as input and the corresponding optimal current density adjustment as the expected output, the initial machine model is trained using these data to obtain the trained adjusted mapping model. In specific implementation, the sample candidate optimization parameters and the corresponding optimal adjustment are divided into a training set and a validation set; the machine learning model is initialized (e.g., setting the number of layers and nodes in the neural network); the model is trained using the training set data, and the model parameters are adjusted using algorithms such as backpropagation to make the model output as close as possible to the expected optimal adjustment; the model performance is evaluated using the validation set, and if it does not meet the target, the model structure or parameters are adjusted and retrained; finally, a post-trained adjusted mapping model with satisfactory performance is obtained.

[0100] Corresponding to the above embodiment of the anodizing current density adjustment scheme generation method based on mutual information entropy, this application embodiment also provides an anodizing current density adjustment scheme generation system based on mutual information entropy. Each unit of the system can implement each step of the anodizing current density adjustment scheme generation method based on mutual information entropy. Figure 5The diagram shows a structural block diagram of the anodizing current density adjustment scheme generation system based on mutual information entropy provided in the embodiments of this application. For ease of explanation, only the parts related to the embodiments of this application are shown.

[0101] Reference Figure 5 The anodizing current density adjustment scheme generation system based on mutual information entropy includes: The acquisition unit is used to acquire process parameter data of aluminum profiles at each moment during the anodizing process, and extract key feature parameters affecting current density optimization based on the process parameter data at each moment through an adaptive filtering algorithm; wherein, the key feature parameters are determined based on the mutual information entropy value between the process parameter data and the preset current density optimization target. The data unit is used to filter out the real-time current density data corresponding to parameters whose influence weight exceeds the preset influence weight based on the key feature parameters and confirm them as current density data to be optimized. The extraction unit is used to extract candidate current density curves from the current density data to be optimized that match the preset oxidation quality standard within a preset matching threshold, and to identify the deviation values ​​of the current density value corresponding to the current processing position of the aluminum profile from each candidate current density curve as candidate optimization parameters. The result unit is used to generate an adaptive current density adjustment scheme based on the candidate optimization parameters.

[0102] It should be noted that the information interaction and execution process between the above systems / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0103] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit module can exist physically separately, or two or more unit modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0104] This application also provides an electronic device. Figure 6This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 6 As shown, the electronic device 6 of this embodiment includes: at least one processor 60 ( Figure 6 Only one is shown in the image), at least one memory 61 ( Figure 6 (Only one is shown in the image) and a computer program 62 stored in the at least one memory 61 and executable on the at least one processor 60, wherein when the processor 60 executes the computer program 62, it causes the electronic device 6 to implement the steps in any of the above embodiments of the method for generating anodizing current density adjustment scheme based on mutual information entropy, or causes the electronic device 6 to implement the functions of each unit in the above system embodiments.

[0105] For example, the computer program 62 may be divided into one or more units, which are stored in the memory 61 and executed by the processor 60 to complete this application. The one or more units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program 62 in the electronic device 6.

[0106] Electronic device 6 can be a computing device or terminal device such as a mobile phone, tablet computer, desktop computer, laptop, handheld computer, and cloud server. This electronic device may include, but is not limited to, a processor 60 and a memory 61. Those skilled in the art will understand that... Figure 6 This is merely an example of electronic device 6 and does not constitute a limitation on electronic device 6. It may include more or fewer components than shown, or combine certain components, or different components, such as input / output devices, network access devices, buses, etc.

[0107] The processor 60 can be a Central Processing Unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0108] In some embodiments, the memory 61 may be an internal storage unit of the electronic device 6, such as a hard disk or memory of the electronic device 6. In other embodiments, the memory 61 may be an external storage device of the electronic device 6, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 6. Furthermore, the memory 61 may include both internal and external storage units of the electronic device 6. The memory 61 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 61 can also be used to temporarily store data that has been output or will be output.

[0109] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0110] This application provides a computer program product that, when run on an electronic device, causes the electronic device to perform the steps in any of the above method embodiments.

[0111] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to an electronic device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0112] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0113] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0114] In the embodiments provided in this application, it should be understood that the disclosed anodizing current density adjustment scheme generation system / electronic device and method based on mutual information entropy can be implemented in other ways. For example, the embodiments of the anodizing current density adjustment scheme generation system / electronic device based on mutual information entropy described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0115] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0116] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for generating anodizing current density adjustment schemes based on mutual information entropy, characterized in that, The method includes: The process parameter data of aluminum profiles at each moment during the anodizing process is obtained, and based on the process parameter data at each moment, the key feature parameters affecting the current density optimization are extracted by an adaptive filtering algorithm; wherein, the key feature parameters are determined based on the mutual information entropy value between the process parameter data and the preset current density optimization target. Based on the key feature parameters, the real-time current density data corresponding to parameters whose influence weight exceeds the preset influence weight are identified as current density data to be optimized. Extract candidate current density curves from the current density data to be optimized that match the preset oxidation quality standard within a preset matching threshold, and identify the deviation values ​​of the current density value corresponding to the current processing position of the aluminum profile from each candidate current density curve as candidate optimization parameters. Based on the proposed optimization parameters, an adaptive current density adjustment scheme is generated; The process of acquiring process parameter data of the aluminum profile at each moment during the anodizing process, and extracting key feature parameters affecting current density optimization based on the process parameter data at each moment using an adaptive filtering algorithm, includes: The process parameter data of aluminum profiles at each moment during the anodizing process are obtained, and the process parameter data is preprocessed to obtain the first process parameter sequence; The first process parameter sequence is separated by an adaptive filtering algorithm to determine the second process parameter sequence. A multi-dimensional feature pool is constructed based on the second process parameter sequence to obtain the key feature parameters that affect current density optimization. The construction of a multi-dimensional feature pool based on the second process parameter sequence yields key feature parameters affecting current density optimization, including: For each parameter in the second process parameter sequence, multi-dimensional features are extracted; the multi-dimensional features include statistical features and time series features, wherein the statistical features include the central tendency, dispersion and extreme value features of the parameter within a preset time window, and the time series features include the change trend, fluctuation frequency and correlation features of adjacent time points of the parameter. Based on the multi-dimensional features of all parameters, a multi-dimensional feature pool containing multi-type and multi-scale features is formed; Based on the multidimensional feature pool, each parameter in the second process parameter sequence is filtered to obtain the key feature parameters that affect current density optimization; The step of filtering each parameter in the second process parameter sequence based on the multidimensional feature pool to obtain key feature parameters affecting current density optimization includes: Calculate the mutual information entropy between each feature in the multidimensional feature pool and the preset current density optimization target; the mutual information entropy is used to quantify the degree of nonlinear correlation between the feature and the optimization target, and the higher the entropy value, the more significant the influence of the feature on the optimization target; Based on the mutual information entropy, each feature in the multidimensional feature pool is filtered to obtain the key feature parameters that affect the current density optimization; The step of filtering each feature in the multi-dimensional feature pool based on the mutual information entropy to obtain key feature parameters affecting current density optimization includes: Features whose mutual information entropy value is not less than a preset mutual information entropy threshold are selected, and the parameters corresponding to the selected features are confirmed as initial candidate feature parameters. Redundancy checks are performed on the initial candidate feature parameters. The mutual information entropy between any two initial candidate feature parameters is calculated. If the mutual information entropy between the two exceeds a preset redundancy threshold, feature redundancy is determined to exist, and the feature parameter with a higher mutual information entropy than the optimization target is retained. The feature parameters retained after redundancy testing were identified as key feature parameters affecting current density optimization.

2. The method as described in claim 1, characterized in that, The step of performing signal separation on the first process parameter sequence using an adaptive filtering algorithm to determine the second process parameter sequence includes: Obtain the parameter sequence of aluminum profile anodizing under historical stable process conditions, and use the parameter sequence as the expected signal sequence; the expected signal sequence contains the process parameter fluctuation characteristics corresponding to qualified anodizing quality; The first process parameter sequence is used as the input signal sequence of the adaptive filtering algorithm to initialize the filter coefficient vector; the dimension of the filter coefficient vector matches the number of parameter types in the first process parameter sequence. A convergence factor is set based on the characteristics of the input signal sequence. Based on the minimum mean square error criterion, the filtered output signal and error signal at the current time are calculated. The filter coefficient vector is updated iteratively until the error signal is less than a preset error threshold or the number of iterations reaches a preset upper limit. The error signal is the deviation between the desired signal sequence and the filtered output signal. The filtered output signal sequence after iterative convergence is identified as the second process parameter sequence; the second process parameter sequence is the effective process parameter change sequence after removing noise interference, which contains signal components strongly correlated with current density optimization.

3. The method as described in claim 1, characterized in that, The step of identifying real-time current density data corresponding to parameters whose influence weight exceeds a preset influence weight based on the key feature parameters and confirming them as current density data to be optimized includes: The influence weight of each parameter in the key feature parameters is determined based on the mutual information entropy value with the preset current density optimization target. Strongly influential feature parameters with influence weights exceeding the preset influence weight threshold are selected, and the real-time current density data corresponding to the strongly influential feature parameters are extracted to form the current density data to be optimized.

4. The method as described in claim 1, characterized in that, The step of extracting candidate current density curves from the current density data to be optimized, whose matching degree with the preset oxidation quality standard is within a preset matching threshold, and confirming the deviation values ​​of the current density value corresponding to the current processing position of the aluminum profile from each candidate current density curve as candidate optimization parameters includes: The matching degree between the current density data to be optimized and the preset oxidation quality standard is calculated using a feature comparison algorithm. Select the current density data to be optimized that meet or exceed the preset matching threshold, and generate the candidate current density curve through data fitting; Obtain the real-time current density measurement value at the current processing position of the aluminum profile, calculate the deviation value between the real-time current density measurement value and the theoretical value of each candidate current density curve at the corresponding time, and confirm the deviation value as the candidate optimization parameter.

5. The method as described in claim 1, characterized in that, The step of generating an adaptive current density adjustment scheme based on the candidate optimization parameters includes: The candidate optimization parameters are input into a preset adjustment mapping model to obtain the current density adjustment amount of the aluminum profile during the anodizing process; wherein, the adjustment mapping model is a pre-trained machine learning model; The rationality of the current density adjustment is verified based on the preset process constraints of aluminum profile anodizing, and dynamic adjustment rules are formulated according to the process sequence. The dynamic adjustment rules are then determined as an adaptive current density adjustment scheme.

6. The method as described in claim 5, characterized in that, Before inputting the candidate optimization parameters into a preset adjustment mapping model, the method further includes: Obtain the sample optimization parameters and corresponding sample current density adjustment and sample oxidation quality results from historical processes; Based on the sample oxidation quality results data, the optimal current density adjustment amount corresponding to the sample candidate optimization parameters is determined from the corresponding current density adjustment amount; Using the sample candidate optimization parameters as input and the corresponding optimal current density adjustment amount as the expected output, the initial machine learning model is trained to obtain the trained adjustment mapping model.

Images