Information judgment method in metal object oxidation process in electroplating process

Abstract

The invention belongs to the field of metal object oxidation data judgment methods in the electroplating process, and particularly relates to an information judgment method in the metal object oxidation process in the electroplating process. The method comprises the steps of firstly collecting oxidation process information and identifying an oxidation abnormity type; then selecting a key reference index, and obtaining feature data; and arranging the data by using the performance index matrix, and carrying out normalization processing through a minimum-maximum standardization method. And further setting a standard performance index matrix, and calculating the cosine similarity between the target matrix and the standard matrix to identify the oxidation anomaly. Through systematic data analysis, automatic monitoring of the oxidation process is achieved, and the production efficiency and the reliability of metal objects in the electroplating process are improved.

Description

Technical Field

[0001] The invention belongs to the field of oxidation information judgment methods, and specifically relates to an information judgment method during the oxidation process of a metal object in an electroplating process. Background Art

[0002] In the electroplating process, the oxidation process of metal objects is a key link, and its quality directly affects the performance and service life of electroplated products. However, in the actual production process, metal objects are prone to various abnormalities during oxidation.

[0003] Uneven oxide layer thickness is one of the more common problems, which may lead to inconsistent local performance of the product, affecting the overall protection effect and appearance quality. For example, in the electroplating of some electronic components with high flatness requirements, uneven oxide layer thickness may cause unstable circuit conductivity.

[0004] Too fast an oxidation rate can also bring many adverse effects. On the one hand, it may make the oxide layer structure loose and reduce its density, thus affecting the protective effect on the metal matrix; on the other hand, too fast an oxidation rate may make it difficult to control the oxidation process, increasing the uncertainty in the production process.

[0005] Abnormal oxidation products should not be ignored either. When their chemical composition or crystal structure does not meet expectations, the adhesion of the electroplating layer may decrease, or the expected corrosion resistance, wear resistance and other performance requirements may not be achieved.

[0006] At present, the monitoring and abnormal judgment of the oxidation process of metal objects in the electroplating process mainly rely on manual experience and simple detection methods, and lack a systematic, accurate and efficient information judgment method. Manual detection is not only time-consuming and laborious, but also difficult to guarantee accuracy, and is prone to misjudgment or missed judgment. Simple detection methods can often only obtain limited information and cannot comprehensively and deeply analyze the various complex situations in the oxidation process. This urgently requires a method that can accurately monitor and analyze the oxidation process and timely and accurately judge the type of abnormality, so as to improve the quality and production efficiency of electroplating products, reduce production costs, and enhance the competitiveness of enterprises in the market. Summary of the invention

[0007] In view of the shortcomings of the prior art, the present invention proposes a method for determining information during the oxidation process of metal objects in the electroplating process. Through systematic data collection, analysis and processing, the method can accurately identify the abnormal types of metal objects in the oxidation process in the electroplating process and make timely adjustments and treatments. This not only improves the oxidation quality and prolongs the life of metal objects in the electroplating process, but also greatly improves production efficiency and reliability through automated and intelligent data analysis.

[0008] The technical solution of the present invention is as follows:

[0009] A method for determining information during oxidation of a metal object in an electroplating process, comprising:

[0010] Obtain characteristic data of metal objects in the electroplating process during oxidation, analyze the characteristic data during oxidation through IsolationForest, determine the type of oxidation anomaly, and determine the reference index for analysis for each type of oxidation anomaly;

[0011] The analysis results of the characteristic data in each oxidation process under different oxidation anomaly types are used as elements of the performance indicator matrix to construct the performance indicator matrix V;

[0012] The performance indicator matrix V is standardized so that the value of each element is in the range of [0,1] to obtain the target performance indicator matrix;

[0013] Set up a standard performance indicator matrix corresponding to each oxidation anomaly type;

[0014] Determine the coordinates of the indicators in the target performance indicator matrix as the outlier values, screen the target performance indicator matrix according to the outlier value coordinates to obtain the final target performance indicator matrix; screen the standard matrix corresponding to the current oxidation anomaly type according to the outlier value coordinates to obtain the final standard performance indicator matrix; calculate the cosine similarity of the final target performance indicator matrix and the final standard performance indicator matrix as the similarity between the target performance indicator matrix and the current oxidation anomaly type;

[0015] The similarity of each oxidation anomaly type is compared, and the oxidation anomaly type with the greatest similarity is determined as the final oxidation anomaly type of the metal object in the target electroplating process.

[0016] Furthermore, the process of determining the type of oxidation anomaly is as follows:

[0017] Analyze the feature data through IsolationForest to obtain the anomaly score;

[0018] and determining whether the characteristic data point is an abnormal point according to the abnormal score, and classifying the abnormal point to determine the type of oxidation abnormality;

[0019] The calculation formula of the anomaly score is:

[0020] s(x)=2-E(h(x)) / c(n);

[0021] Where s(x) is the anomaly score of feature data point x; h(x) is the path length of feature data point x in the book; E(h(x)) is the expectation of h(x), which is obtained by averaging the path lengths of a large number of normal feature data points; c(n) is a constant related to the number of data points n, which is used to normalize the anomaly score;

[0022] The calculation formula of c(n) is:

[0023] c(n)=2H(n-1)-2(n-1) / n;

[0024] Among them, H(n-1) is the harmonic number, which is approximately ln(i)+0.5772156649.

[0025] Furthermore, the abnormal points are classified to determine the types of oxidation abnormalities, including: uneven thickness of the oxidation layer, excessively fast oxidation rate, and abnormal oxidation products;

[0026] Among them, the reference index of the uneven thickness of the oxide layer is: the thickness measurement value of the oxide layer at different positions;

[0027] The reference index for too fast oxidation speed is: the increase in the thickness of the oxide layer per unit time;

[0028] Reference indicators for abnormal oxidation products are: chemical composition analysis and crystal structure analysis of oxidation products.

[0029] Further, the analysis result of the uneven thickness of the oxide layer includes the thickness measurement values ​​of the oxide layer at different positions and the degree of difference between these measurement values;

[0030] The analysis result of the excessive oxidation rate includes the increase in the thickness of the oxide layer per unit time and the comparison with the normal oxidation rate;

[0031] The abnormality analysis results of the oxidation products include the chemical composition analysis results, crystal structure analysis results and the differences from the expected oxidation products.

[0032] Furthermore, the performance indicator matrix V is constructed as follows:

[0033] Number of rows = number of oxidation anomaly types, number of columns = number of oxidation-related data;

[0034] The analysis result of each oxidation-related data under a specific oxidation anomaly type is filled into the intersection of the corresponding row and column in the matrix.

[0035] Furthermore, the performance indicator matrix V is standardized by using minimum-maximum standardization to calculate each element in the performance indicator matrix V to obtain the standardized element value, that is, for the element Vij in the matrix V, the standardization formula is:

[0036]

[0037] Among them, Vmin is the minimum value of all elements in the matrix V, Vmax is the maximum value of all elements in the matrix V, and Vstdij is the standardized element value.

[0038] Furthermore, after the performance indicator matrix V is standardized, all standardized element values ​​are combined into a matrix according to the original position relationship, which is the target performance indicator matrix.

[0039] Furthermore, the standard performance indicator matrix is ​​constructed by determining the number of matrix rows according to the oxidation anomaly type, determining the number of matrix columns according to the relevant performance indicators, and filling the determined standard values ​​or ranges into corresponding positions of the standard performance indicator matrix.

[0040] Furthermore, the cosine similarity calculation formula is: Assume that the final target performance indicator matrix is The final standard performance indicator matrix is The cosine similarity calculation formula is:

[0041]

[0042] Among them, a11, a12, a21, and a22 represent the elements of the final target performance indicator matrix;

[0043] b11, b12, b21, and b22 represent the elements of the final standard performance indicator matrix.

[0044] The beneficial effects of the present invention are: 1) It can accurately determine the type of oxidation anomaly that may occur in a metal object during an electroplating process, and improve the accuracy of abnormality type judgment through in-depth analysis of the oxidation process and clarification of reference indicators.

[0045] 2) Selecting appropriate reference indicators for different types of oxidation anomalies will help to more accurately analyze oxidation anomalies and detect problems in a timely manner.

[0046] 3) By acquiring and analyzing the characteristic data of the metal objects in the electroplating process during the oxidation process, the oxidation state of the metal objects in the electroplating process can be fully understood, providing a reliable basis for subsequent processing.

[0047] 4) Constructing a performance indicator matrix and performing standardization processing eliminates the dimensional differences between different indicators and facilitates unified analysis and comparison.

[0048] 5) A standard performance indicator matrix corresponding to each oxidation anomaly type is set to provide a standard reference for judging the oxidation anomaly of metal objects in the electroplating process.

[0049] 6) Calculating the cosine similarity between the target performance indicator matrix and the standard performance indicator matrix can accurately measure the similarity between the target performance indicator matrix and the current oxidation anomaly type, thereby more accurately determining the final oxidation anomaly type.

[0050] 7) The holistic approach helps to improve the reliability and safety of metal objects in the electroplating process, detect and solve oxidation problems in a timely manner, and extend the service life of metal objects in the electroplating process.

[0051] 8) This method is based on data information processing, is scientific and objective, and can effectively improve the efficiency and accuracy of oxidation anomaly judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is the overall flow chart of the present invention;

[0053] Figure 2 The specific steps of S100;

[0054] Figure 3 Specific steps for S200;

[0055] Figure 4 The step of normalizing the performance indicator matrix V in S300;

[0056] Figure 5 Steps to set up a standard performance indicator matrix for S400;

[0057] Figure 6 Steps for calculating cosine similarity for S500;

[0058] Figure 7 The step of determining the final oxidation abnormality type is S600. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0060] See also Figure 1 , which shows a method for determining information during oxidation of a metal object in an electroplating process provided by an embodiment of the present invention, comprising:

[0061] S100: Acquire characteristic data of a metal object in an oxidation process in an electroplating process, analyze the characteristic data in the oxidation process through IsolationForest, determine the oxidation anomaly type, and determine a reference indicator for analysis for each oxidation anomaly type.

[0062] S200: Taking the analysis results of the characteristic data in each oxidation process under different oxidation anomaly types as elements of the performance indicator matrix, a performance indicator matrix V is constructed.

[0063] S300: Standardize the performance indicator matrix V so that the value of each element is within the range of [0, 1], and obtain the target performance indicator matrix.

[0064] S400: setting a standard performance indicator matrix corresponding to each oxidation anomaly type;

[0065] S500: Determine the coordinates of the outliers as the values ​​of the indicators in the target performance indicator matrix, filter the target performance indicator matrix according to the coordinates of the outliers to obtain a final target performance indicator matrix; filter the standard matrix corresponding to the current oxidation anomaly type according to the coordinates of the outliers to obtain a final standard performance indicator matrix; calculate the cosine similarity between the final target performance indicator matrix and the final standard performance indicator matrix as the similarity between the target performance indicator matrix and the current oxidation anomaly type.

[0066] S600: Compare the similarity of each oxidation anomaly type, and determine the oxidation anomaly type with the greatest similarity as the final oxidation anomaly type of the metal object in the target electroplating process.

[0067] S100 involves in-depth research on the oxidation process of metal objects in the electroplating process, using appropriate detection technology and equipment to obtain relevant characteristic data of metal objects in the electroplating process during the oxidation process, so as to conduct accurate analysis, and by obtaining characteristic data of metal objects in the electroplating process during the oxidation process, the data in the oxidation process is analyzed through IsolationForest to determine the type of oxidation anomaly. For each type of oxidation anomaly, specific reference indicators that can effectively analyze the anomaly are identified.

[0068] In some possible embodiments, Figure 2 As shown, S100 specifically includes:

[0069] S110: Use professional testing equipment and technology to test metal objects in the electroplating process, obtain characteristic data related to oxidation, including temperature, humidity, oxidation time, material of metal objects in the electroplating process, etc., and pre-process the data;

[0070] S120: Analyze the feature data through IsolationForest to obtain an anomaly score.

[0071] Randomly select some data from the feature data as subsamples, and randomly select a feature. Randomly select a split point on the selected feature to split the data into two subsets, so that samples less than the split value are on the left and samples greater than the split value are on the right. Repeat the split process until the number of samples in the subset is less than or equal to a predetermined threshold, or the preset tree depth is reached. Calculate the path length of each feature data in the tree, that is, for each feature data, starting from the root node, the path length increases by 1 each time it is split until the leaf node is reached. By counting the path lengths of multiple trees, the average path length is calculated. Determine the anomaly score based on the average path length. The calculation formula for the anomaly score is:

[0072] s(x)=2-E(h(x)) / c(n) ；

[0073] Where s(x) is the anomaly score of feature data point x; h(x) is the path length of feature data point x in the book; E(h(x)) is the expectation of h(x), which is obtained by averaging the path lengths of a large number of normal feature data points; c(n) is a constant related to the number of data points n, which is used to normalize the anomaly score;

[0074] Specifically, the calculation formula of c(n) is:

[0075] c(n)=2H(n-1)-2(n-1) / n;

[0076] Among them, H(n-1) is the harmonic number, which is approximately ln(i)+0.5772156649.

[0077] S130: Determine whether the characteristic data point is an abnormal point according to the abnormal score, classify the abnormal point, determine the oxidation abnormality type, set the abnormal score threshold, and the data point above the threshold is considered to be an abnormal point. According to the characteristics of the abnormal point and related knowledge, determine the possible oxidation abnormality type.

[0078] Based on the previous analysis, possible types of oxidation anomalies include:

[0079] Uneven oxide layer thickness: The oxide layer has different distribution thickness on the surface of the metal object during the electroplating process.

[0080] Too fast oxidation rate: The oxidation process of the metal object in the electroplating process progresses abnormally quickly, exceeding the normal oxidation rate.

[0081] Abnormal oxidation products: The substances generated after oxidation are different from the expected oxidation products and may have abnormalities in chemical composition or structure.

[0082] S140: For each determined oxidation anomaly type, select reference indicators that can effectively reflect the anomaly type based on its characteristics and relevant scientific theories. These indicators are representative and sensitive and can accurately capture the characteristics of oxidation anomalies.

[0083] Specifically, the reference indicators of possible oxidation abnormalities are:

[0084] Non-uniform oxide layer thickness: The thickness of the oxide layer measured at different locations.

[0085] By measuring the oxide layer thickness at multiple locations on the surface of a metal object in the electroplating process, the distribution of the thickness can be obtained. If there is a large difference between these measured values, which exceeds the allowable tolerance range, it indicates that there is an abnormality of uneven oxide layer thickness. For example, select several measurement points on a metal object in an electroplating process, measure the oxide layer thickness at each point, and then calculate the standard deviation of these thickness values. If the standard deviation is large, it means that the thickness distribution is uneven.

[0086] Oxidation rate is too fast: the increase in oxide layer thickness per unit time.

[0087] Monitor the change in oxide layer thickness of metal objects in the electroplating process over a certain period of time, and calculate the increase in thickness per unit time. If this increase exceeds the normal oxidation rate range, it means that the oxidation rate is too fast. For example, measure the oxide layer thickness of metal objects in the electroplating process at regular intervals, calculate the increase in thickness between two adjacent measurements, and then compare it with the expected increase under normal oxidation rate.

[0088] Abnormal oxidation products: chemical composition analysis and crystal structure analysis of oxidation products.

[0089] The chemical composition of the oxidized product is analyzed to check whether the elemental composition is consistent with the expected oxidized product. At the same time, the crystal structure analysis is performed to check whether there is an abnormal structure. For example, the content of each element in the oxidized product is determined by spectral analysis and compared with the standard oxidized product composition; or the crystal structure of the oxidized product is analyzed by X-ray diffraction and other techniques to see whether it is consistent with the normal situation.

[0090] S200 involves analyzing characteristic data obtained during each oxidation process, determining analysis results of these data under different oxidation anomaly types, and using these analysis results as elements for constructing a performance indicator matrix V, wherein each element in the performance indicator matrix V corresponds to an analysis result of an oxidation-related data under a specific oxidation anomaly type.

[0091] In some possible embodiments, Figure 3 As shown, S200 specifically includes:

[0092] S210: Arrange and classify the characteristic data analysis results for different oxidation anomaly types in each oxidation process to ensure the orderliness and accuracy of the analysis results, so as to facilitate the subsequent construction of the matrix.

[0093] S220: Determine the number of rows and columns of the performance indicator matrix V according to the number of oxidation anomaly types and the number of oxidation-related data.

[0094] The number of rows = the number of oxidation anomaly types, and the number of columns = the number of oxidation-related data.

[0095] For example, if there are three types of oxidation anomalies and five oxidation-related data, the number of rows and columns of the matrix V is 3 and 5, respectively.

[0096] S230: Fill the sorted analysis results into each element of the performance indicator matrix V accordingly. After determining the number of rows and columns of the performance indicator matrix V, fill the sorted analysis results into each element of the matrix accordingly according to certain rules. Specifically, the rows of the matrix represent the oxidation anomaly type, and the columns represent the oxidation-related data. For each analysis result of oxidation-related data under a specific oxidation anomaly type, fill it into the intersection of the corresponding row and column in the matrix. By reasonably filling the matrix elements, the relationship between each oxidation-related data and different oxidation anomaly types can be clearly displayed, so that the performance indicator matrix can comprehensively and accurately reflect the various indicators of the oxidation process. This helps to analyze and process the matrix later, so as to more effectively judge the oxidation state and anomaly type of metal objects in the electroplating process.

[0097] The performance index matrix V has 3 rows, representing three types of oxidation anomalies: uneven oxide thickness, excessive oxidation rate, and abnormal oxidation products; and 5 columns, representing five oxidation-related data: oxide thickness, oxidation time, oxygen concentration, temperature, and humidity. For the analysis results of oxide thickness under the abnormal type of uneven oxide thickness, fill it into the intersection of the first row of the matrix and the corresponding oxide thickness column; for the analysis results of oxidation time under the abnormal type of excessive oxidation rate, fill it into the intersection of the second row of the matrix and the corresponding oxidation time column, and so on.

[0098] S300 involves normalizing the performance indicator matrix V. The normalization process is to normalize the element values ​​in the performance indicator matrix V so that they fall within the range of [0, 1]. This can eliminate the dimensional differences between different indicators and facilitate subsequent analysis and comparison.

[0099] In some possible embodiments, Figure 4 As shown, S300 specifically includes:

[0100] S310: Use minimum-maximum standardization to calculate each element in the performance indicator matrix V to obtain a standardized element value. Ensure that each element in the matrix is ​​standardized to the range of [0,1] to make the data comparable.

[0101] For the element Vij in the matrix V, the normalization formula is:

[0102]

[0103] Among them, Vmin is the minimum value of all elements in the matrix V, Vmax is the maximum value of all elements in the matrix V, and Vstdij is the standardized element value.

[0104] S320: All standardized element values ​​are combined into a matrix according to their original positional relationship, which is the target performance indicator matrix. The obtained target performance indicator matrix has a unified scale, which is convenient for subsequent analysis and processing.

[0105] S400 involves setting a standard performance indicator matrix according to the previously determined oxidation anomaly type. The elements in the standard performance indicator matrix represent standard values ​​or ranges of various performance indicators corresponding to the oxidation anomaly type under normal circumstances.

[0106] Standard values ​​can be derived from historical data, experimental results, expert experience or industry standards, etc.

[0107] In some possible embodiments, Figure 5 As shown, S400 specifically includes:

[0108] S410: Determine the number of matrix rows according to the oxidation anomaly type. The number of oxidation anomaly types determines the number of rows of the standard performance indicator matrix, and each oxidation anomaly type corresponds to a row. The number of matrix rows corresponds to the oxidation anomaly type one by one, which is convenient for subsequent analysis and comparison.

[0109] S420: Determine the number of matrix columns according to the relevant performance indicators. The number of relevant performance indicators determines the number of columns of the standard performance indicator matrix, and each performance indicator corresponds to one column. The number of matrix columns corresponds to the performance indicator one by one, which can fully reflect the standard situation of each oxidation anomaly type in each performance indicator.

[0110] S430: Fill the determined standard value or range into the corresponding position of the standard performance indicator matrix, so that the standard performance indicator matrix fully reflects the standard performance indicator of each oxidation abnormality type.

[0111] S500 first finds the coordinates of the elements in the target performance indicator matrix whose values ​​are outliers. Then, based on the coordinates of these outliers, the target performance indicator matrix is ​​screened, and the rows and columns containing outliers are removed to obtain the final target performance indicator matrix. Similarly, the standard matrix corresponding to the current oxidation anomaly type is screened based on the coordinates of the outliers to obtain the final standard performance indicator matrix. Finally, the cosine similarity between the final target performance indicator matrix and the final standard performance indicator matrix is ​​calculated to measure the similarity between the target performance indicator matrix and the current oxidation anomaly type.

[0112] In some possible embodiments, Figure 6 As shown, S500 includes:

[0113] S510: traverse the target performance indicator matrix, find out the element whose value is an abnormal value, and record the row and column coordinates of the element.

[0114] The target performance indicator matrix is Among them, NAN represents an outlier; the determined outlier coordinates are (2,2).

[0115] S520: According to the outlier coordinates, remove the rows and columns containing the outliers from the target performance indicator matrix to obtain a final target performance indicator matrix.

[0116] For the above matrix M, after removing the second row and second column, the final target performance indicator matrix is ​​obtained as follows:

[0117]

[0118] S530: In the same manner as S520, the standard matrix corresponding to the current oxidation anomaly type is screened according to the outlier value coordinates to obtain a final standard performance indicator matrix.

[0119] The standard matrix corresponding to the current oxidation anomaly type is After screening according to the same outlier coordinates (2,2), the final standard performance indicator matrix is ​​obtained as follows:

[0120] S540: Calculate the similarity between the final target performance indicator matrix and the final standard performance indicator matrix using the cosine similarity formula.

[0121] For the final target performance indicator matrix M′ and the final standard performance indicator matrix S′

[0122] The cosine similarity calculation formula is:

[0123]

[0124] Among them, a11, a12, a21, and a22 represent the elements of the final target performance indicator matrix; b11, b12, b21, and b22 represent the elements of the final standard performance indicator matrix.

[0125] Substituting the specific value into the formula, the value of cosine similarity can be obtained.

[0126] For example: The final target performance indicator matrix is The final standard performance indicator matrix is Substitute into the cosine similarity calculation formula to obtain Through the above calculation, the cosine similarity between the final target performance indicator matrix and the final standard performance indicator matrix is ​​about 0.857.

[0127] In the previous step, the similarities between the target performance indicator matrix and each oxidation anomaly type have been calculated. In S600, these similarities need to be compared. The oxidation anomaly type with the greatest similarity is found and determined as the final oxidation anomaly type of the metal object in the target electroplating process.

[0128] In some possible embodiments, Figure 7 As shown, S600 includes:

[0129] S610: Summarize the similarity data of the target performance indicator matrix and each oxidation anomaly type calculated previously to ensure that similarity information of any oxidation anomaly type is not omitted.

[0130] S620: Compare the collected similarity data to find the largest value.

[0131] S630: Determine the oxidation anomaly type with the greatest similarity as the final oxidation anomaly type of the metal object in the target electroplating process.

[0132] Example 2

[0133] Use professional testing equipment to obtain characteristic data of metal objects in the electroplating process, such as:

[0134] Temperature: During the oxidation process, the temperature data obtained by multiple measurements were [25℃, 26℃, 27℃, 28℃, 29℃].

[0135] Humidity: The corresponding humidity data is [40%, 42%, 41%, 43%, 42%].

[0136] Oxidation time: The recorded oxidation time is [10h, 11h, 12h, 13h, 14h].

[0137] Material of metal objects in electroplating process: The material of metal objects in electroplating process is aluminum alloy.

[0138] The feature data is analyzed by IsolationForest to obtain the anomaly score. Some data are randomly selected as subsamples, and features are randomly selected for segmentation to build multiple trees. The path length of each feature data in the tree is calculated. For example, the path length of a data point is 5. The average path length is calculated by counting the path lengths of multiple trees. The average path length is 8. The anomaly score is calculated according to the anomaly score calculation formula, c(n) = 1.5, E(h(x)) = 8, then the anomaly score of the data point is s(x) = 2^(-8 / 1.5) = 0.282.

[0139] Determine whether the feature data point is an outlier based on the anomaly score, classify the anomaly, and determine the type of oxidation anomaly. Set the anomaly score threshold to 0.5, and data points above the threshold are considered outliers. Determine the type of oxidation anomaly based on the characteristics of the anomaly and related knowledge. It was found that the oxide layer thickness of some metal objects in the electroplating process is uneven.

[0140] For the abnormal type of uneven oxide layer thickness, the thickness measurement values ​​of the oxide layer at different positions are selected as reference indicators.

[0141] The oxide layer thickness was measured at multiple locations on the surface of the metal object during the electroplating process, and the thickness measurement values ​​were [5μm, 6μm, 8μm, 4μm, 7μm]. The standard deviation of these thickness values ​​was calculated, and the standard deviation was 1.5μm, which exceeded the allowable tolerance range of 0.5μm, indicating that there was an abnormality of uneven oxide layer thickness.

[0142] Arrange and classify the characteristic data analysis results for different oxidation anomaly types in each oxidation process. Determine the number of rows and columns of the performance indicator matrix V. There are three types of oxidation anomalies: uneven oxide layer thickness, excessive oxidation rate, and abnormal oxidation products, as well as five oxidation-related data: oxide layer thickness, oxidation time, oxygen concentration, temperature, and humidity. The number of rows and columns of matrix V is 3 and 5, respectively.

[0143] Fill the sorted analysis results into the performance indicator matrix V accordingly. For example, the analysis result of the oxide layer thickness under the oxide layer thickness unevenness abnormal type is [5μm, 6μm, 8μm, 4μm, 7μm], which is filled into the intersection of the first row of the matrix and the corresponding oxide layer thickness column; the analysis result of the oxidation time under the oxidation speed too fast abnormal type is [10h, 11h, 12h, 13h, 14h], which is filled into the intersection of the second row of the matrix and the corresponding oxidation time column, and so on.

[0144] Use minimum-maximum normalization to calculate each element in the performance indicator matrix V. The minimum value of all elements in the matrix V is 0 and the maximum value is 10. For element 8, the standardized element value is (8-0) / (10-0)=0.8. All standardized element values ​​are combined into the target performance indicator matrix according to the original position relationship.

[0145] According to the oxidation anomaly type, the number of rows of the standard performance indicator matrix is ​​determined to be 3. According to the relevant performance indicators, the number of matrix columns is determined to be 5. The determined standard value or range is filled into the corresponding position of the standard performance indicator matrix. For example, for the oxide layer thickness uneven anomaly type, the standard range of oxide layer thickness is [4μm, 6μm], which is filled into the intersection of the first row of the matrix and the corresponding oxide layer thickness column.

[0146] Traverse the target performance indicator matrix, find the elements with outliers, and record their row and column coordinates. According to the outlier coordinates, remove the rows and columns containing outliers from the target performance indicator matrix to obtain the final target performance indicator matrix. In the same way as S520, filter the standard matrix corresponding to the current oxidation anomaly type according to the outlier coordinates to obtain the final standard performance indicator matrix. Use the cosine similarity formula to calculate the similarity between the final target performance indicator matrix and the final standard performance indicator matrix. The final target performance indicator matrix is The final standard performance indicator matrix is Then the cosine similarity is: 0.94.

[0147] Summarize the similarity data of the target performance indicator matrix calculated previously and each oxidation anomaly type. Compare the collected similarity data and find the largest value. Determine the oxidation anomaly type with the largest similarity as the final oxidation anomaly type of the metal object in the target electroplating process.

[0148] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0149] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0150] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0151] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features of the present invention.

[0152] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for determining information during oxidation of a metal object in an electroplating process, characterized in that: include: Obtain characteristic data of metal objects in the electroplating process during oxidation, analyze the characteristic data during oxidation through IsolationForest, determine the type of oxidation anomaly, and determine the reference index for analysis for each type of oxidation anomaly; The analysis results of the characteristic data in each oxidation process under different oxidation anomaly types are used as elements of the performance indicator matrix to construct the performance indicator matrix V; The performance indicator matrix V is standardized so that the value of each element is in the range of [0,1] to obtain the target performance indicator matrix; Set up a standard performance indicator matrix corresponding to each oxidation anomaly type; Determine the coordinates of the indicators in the target performance indicator matrix as the values ​​of the outliers, screen the target performance indicator matrix according to the coordinates of the outliers, and obtain the final target performance indicator matrix; screen the standard matrix corresponding to the current oxidation anomaly type according to the coordinates of the outliers, and obtain the final standard performance indicator matrix; Calculate the cosine similarity between the final target performance indicator matrix and the final standard performance indicator matrix as the similarity between the target performance indicator matrix and the current oxidation anomaly type; The similarity of each oxidation anomaly type is compared, and the oxidation anomaly type with the greatest similarity is determined as the final oxidation anomaly type of the metal object in the target electroplating process.

2. The method for determining information during oxidation of a metal object in an electroplating process according to claim 1, characterized in that: The process of determining the type of oxidation anomaly is as follows: Analyze the feature data through IsolationForest to obtain the anomaly score; and determining whether the characteristic data point is an abnormal point according to the abnormal score, and classifying the abnormal point to determine the type of oxidation abnormality; The calculation formula of the anomaly score is: s(x)=2-E(h(x)) / c(n); Where s(x) is the anomaly score of feature data point x; h(x) is the path length of feature data point x in the book; E(h(x)) is the expectation of h(x), which is obtained by averaging the path lengths of a large number of normal feature data points; c(n) is a constant related to the number of data points n, which is used to normalize the anomaly score; The calculation formula of c(n) is: c(n)=2H(n-1)-2(n-1) / n; Among them, H(n-1) is the harmonic number, which is approximately ln(i)+0.5772156649.

3. The method for determining information during oxidation of a metal object in an electroplating process according to claim 2, characterized in that: The abnormal points are classified to determine the types of oxidation abnormalities, including: uneven thickness of the oxidation layer, too fast oxidation speed, and abnormal oxidation products; Among them, the reference index of the uneven thickness of the oxide layer is: the thickness measurement value of the oxide layer at different positions; The reference index for too fast oxidation speed is: the increase in the thickness of the oxide layer per unit time; Reference indicators for abnormal oxidation products are: chemical composition analysis and crystal structure analysis of oxidation products.

4. The method for determining information during oxidation of a metal object in an electroplating process according to claim 3, characterized in that: The analysis result of the uneven thickness of the oxide layer includes the thickness measurement values ​​of the oxide layer at different positions and the degree of difference between these measurement values; The analysis result of the excessive oxidation rate includes the increase in the thickness of the oxide layer per unit time and the comparison with the normal oxidation rate; The abnormal analysis results of the oxidation products include chemical composition analysis results, crystal structure analysis results and differences from expected oxidation products.

5. The method for determining information during oxidation of a metal object in an electroplating process according to claim 1, characterized in that: The performance indicator matrix V is constructed as follows: Number of rows = number of oxidation anomaly types, number of columns = number of oxidation-related data; The analysis result of each oxidation-related data under a specific oxidation anomaly type is filled into the intersection of the corresponding row and column in the matrix.

6. The method for determining information during oxidation of a metal object in an electroplating process according to claim 1, characterized in that: The method of standardizing the performance indicator matrix V is: using minimum-maximum standardization, calculating each element in the performance indicator matrix V to obtain the standardized element value, that is, for the element Vij in the matrix V, the standardization formula is: Among them, Vmin is the minimum value of all elements in the matrix V, Vmax is the maximum value of all elements in the matrix V, and Vstdij is the standardized element value.

7. The method for determining information during oxidation of a metal object in an electroplating process according to claim 6, characterized in that: After the performance indicator matrix V is standardized, all standardized element values ​​are combined into a matrix according to the original position relationship, which is the target performance indicator matrix.

8. The method for determining information during oxidation of a metal object in an electroplating process according to claim 1, characterized in that: The standard performance indicator matrix is ​​constructed in the following manner: the number of matrix rows is determined according to the oxidation anomaly type, the number of matrix columns is determined according to the relevant performance indicators, and the determined standard values ​​or ranges are filled into the corresponding positions of the standard performance indicator matrix.

9. The method for determining information during oxidation of a metal object in an electroplating process according to claim 1, characterized in that: The cosine similarity calculation formula is: Assume that the final target performance indicator matrix is The final standard performance indicator matrix is The cosine similarity calculation formula is: Among them, a11, a12, a21, and a22 represent the elements of the final target performance indicator matrix; b11, b12, b21, and b22 represent the elements of the final standard performance indicator matrix.

Images