Intelligent control system for gold bonding wire based on big data analysis

Abstract

The embodiment of the present invention provides an intelligent control system for bonding gold wire based on big data analysis, which belongs to the field of data processing technology. The system includes a data collection module for collecting initial parameter data and initial production results corresponding to the production process of bonding gold wire; a data segmentation module for segmenting the initial parameter data according to fuzzy rules to obtain target segmentation data; an abnormality analysis module for analyzing the abnormality degree of the target segmentation data to obtain the abnormality degree; a data screening module for determining the target parameter data according to the abnormality degree, and obtaining the target production result from the initial production result; a curve determination module for determining any two parameter information from the target parameter data, and determining the target change curve corresponding to the parameter information according to the any two parameter information and the target production result; a strategy determination module for determining the target production state corresponding to the production process according to the target change curve, and determining the intelligent control strategy according to the target production state.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an intelligent gold bonding wire control system based on big data analysis. Background Art

[0002] Gold bonding wire is a thin metal wire used in the microelectronic packaging process, mainly used to connect the electrical path between the semiconductor chip and the external circuit. They are usually made of metals with good electrical conductivity such as gold (Au), aluminum (Al) or copper (Cu). Gold bonding wire is an important material used in the manufacture of electronic products. Controlling the stability and accuracy of its production process is crucial to product quality. However, the modeling analysis technology used in related technologies cannot accurately identify abnormal conditions in the production process of gold bonding wire, which affects product quality and reduces production efficiency. Summary of the invention

[0003] The main purpose of the embodiments of the present invention is to provide an intelligent control system for bonding wires based on big data analysis, aiming to solve the problem that the technology using modeling analysis in related technologies cannot accurately identify abnormal conditions in the production process of bonding wires, thereby affecting product quality and reducing production efficiency.

[0004] In a first aspect, an embodiment of the present invention provides an intelligent gold bonding wire control system based on big data analysis, the system comprising:

[0005] A data collection module, used to collect initial parameter data corresponding to the bonding wire production process and initial production results corresponding to the initial parameter data;

[0006] A data segmentation module, used for segmenting the initial parameter data according to fuzzy rules to obtain target segmentation data corresponding to the initial parameter data;

[0007] An abnormality analysis module, used to analyze the abnormality degree of the target segmentation data to obtain the abnormality degree corresponding to the target segmentation data;

[0008] A data screening module, used to determine target parameter data according to the abnormality degree, and obtain a target production result corresponding to the target parameter data from the initial production result;

[0009] A curve determination module, used to determine any two parameter information from the target parameter data, and determine a target change curve corresponding to the parameter information according to the any two parameter information and the target production result;

[0010] A strategy determination module is used to determine the corresponding target production state in the bonding wire production process according to the target change curve, and determine the intelligent control strategy according to the target production state.

[0011] The embodiment of the present invention provides an intelligent control system for bonding wire based on big data analysis, the system includes a data collection module, a data segmentation module, an abnormal analysis module, a data screening module, a curve determination module, and a strategy determination module. In the present application, the initial parameter data and initial production results in the production process of bonding wire are collected according to the data collection module, so that a reliable data foundation can be established, the possibility of incomplete or distorted data can be reduced, and support can be provided for the subsequent formulation of more effective decisions and strategies to improve production efficiency and quality. According to the data segmentation module, data segmentation processing is performed through fuzzy rules, and the relationship between different parameters can be clearly displayed, which helps to understand the role and influence between parameters, so that the influence and trend of parameters can be analyzed more accurately, and a better basis can be provided for subsequent abnormal analysis and decision-making. According to the abnormal analysis module, through the abnormal degree analysis, abnormal situations in the production process of bonding wire can be discovered in time, which helps to quickly identify potential problems and take solutions to avoid accidents in the production process. Further using the data screening module to determine the target parameter data according to the abnormal degree helps to screen out data that may be abnormal in the production process of bonding wire, thereby reducing unnecessary analysis and processing and improving work efficiency. The curve determination module is then used to determine the target change curve corresponding to any two parameter information, which can intuitively display the correlation between the parameters and help understand the change rules between the parameters. Finally, the strategy determination module is used to determine the corresponding target production state in the bonding wire production process according to the target change curve, which helps to formulate intelligent control strategies and improve the automation and production efficiency of the production process. It also solves the problem that the modeling analysis technology used in related technologies cannot accurately identify abnormal conditions in the bonding wire production process, thereby affecting product quality and reducing production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0013] Figure 1 A schematic diagram of the module structure of a gold bonding wire intelligent control system based on big data analysis provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0014] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0015] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0016] It should be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0017] The embodiment of the present invention provides an intelligent control system for gold bonding wires based on big data analysis. The intelligent control system for gold bonding wires based on big data analysis can be applied to a terminal device, which can be an electronic device such as a tablet computer, a laptop computer, a desktop computer, a personal digital assistant, and a wearable device. The terminal device can be a server or a server cluster.

[0018] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0019] Please refer to Figure 1 , Figure 1 A schematic diagram of the module structure of a gold bonding wire intelligent control system based on big data analysis provided in an embodiment of the present invention.

[0020] like Figure 1As shown, the intelligent control system 200 for bonding wire based on big data analysis includes a data collection module 201, a data segmentation module 202, an abnormality analysis module 203, a data screening module 204, a curve determination module 205, and a strategy determination module 206, wherein the data collection module 201 is used to collect the initial parameter data corresponding to the bonding wire production process and the initial production results corresponding to the initial parameter data; the data segmentation module 202 is used to segment the initial parameter data according to fuzzy rules to obtain the target segmentation data corresponding to the initial parameter data; the abnormality analysis module 203 is used to segment the target segmentation data An abnormality degree analysis is performed to obtain the abnormality degree corresponding to the target segmentation data; a data screening module 204 is used to determine the target parameter data according to the abnormality degree, and obtain the target production result corresponding to the target parameter data from the initial production result; a curve determination module 205 is used to determine any two parameter information from the target parameter data, and determine the target change curve corresponding to the parameter information according to any two parameter information and the target production result; a strategy determination module 206 is used to determine the corresponding target production state in the bonding wire production process according to the target change curve, and determine the intelligent control strategy according to the target production state.

[0021] Exemplarily, the data collection module 201 is used to determine the initial parameter data in the gold bonding wire production process that needs to be collected, and the initial parameter data includes parameter information such as temperature, pressure, and speed. And the initial production result is obtained by testing the corresponding gold bonding wire product after producing it according to the initial parameter information. The initial production result can be any one of the test passing or the test failing, and the initial production result can also be conductive or non-conductive, and can also be the size of conductive resistance, etc.

[0022] Exemplarily, a fuzzy logic system is established using the data segmentation module 202, parameter data is converted into fuzzy sets, and the relationship and operation rules between fuzzy sets are defined to perform fuzzy reasoning and segmentation processing. Then, a suitable fuzzy tool or software platform is selected, such as a fuzzy logic controller or a fuzzy reasoning system, to realize the operation and analysis of fuzzy rules. Thus, based on the selected fuzzy tool, a fuzzy reasoning system is established, including modules such as input parameters, fuzzification processing, fuzzy rule base, reasoning engine and output results. Thus, according to production requirements and fuzzy rules, a corresponding parameter segmentation strategy is formulated to determine how to segment the initial parameter data to obtain the target segmentation data. Finally, the initial parameter data is input into the fuzzy reasoning system, and fuzzy reasoning and segmentation processing are performed according to the pre-set fuzzy rules and segmentation strategy to obtain the target segmentation data.

[0023] Exemplarily, the degree of abnormality is used to evaluate the abnormal value of the target segmentation data, which can be the degree of data deviation from the mean, the degree of data variation, or the deviation from the expected situation. The target segmentation data is preprocessed by the abnormality analysis module 203, such as removing noise, smoothing data, etc., to ensure the accuracy and reliability of the data, and then according to the characteristics of the abnormality analysis index and the distribution of the data, a suitable abnormality analysis method is selected, such as a statistical method, a machine learning algorithm or professional knowledge. Based on the selected abnormality analysis method, an abnormality evaluation model is established, and the target segmentation data is input into the model for analysis and calculation to obtain an abnormality evaluation result. According to actual needs and business background, the threshold value of the abnormality or the range of the abnormality is determined to judge the abnormality of the target segmentation data. The target segmentation data is input into the abnormality evaluation model, the abnormality analysis is performed, the abnormality of each data point or data set is calculated, and the abnormality evaluation result is obtained, so as to determine the abnormality corresponding to the target segmentation data according to the abnormality evaluation result.

[0024] Exemplarily, the data screening module 204 is used to screen out abnormal data from the target segmentation data according to the calibration of the abnormality degree, as target parameter data, for subsequent analysis and processing, and then the target parameter data is associated with the initial production result to determine the initial production result corresponding to the target parameter data, and the association operation can be performed by database query, data matching, etc. Thus, the initial production result corresponding to the target parameter data obtained by association is determined as the target production result corresponding to the target parameter data, that is, the target production result is the initial production result corresponding to the abnormal data.

[0025] Exemplarily, the curve determination module 205 is used to select any two parameter information from the target parameter data, and these parameter information can be any indicators related to the bonding wire production process, such as temperature, pressure, speed, etc. Then, the relevant data of the selected parameters are extracted from the target parameter data, including time series data or other relevant information. Based on the data of the selected parameter information and the target production results, an association model between the parameter information and the production results is established, and the model can be established by regression analysis, neural network and other methods. Then, the established association model is used to fit the data of the selected parameters and the target production results to obtain the target change curve corresponding to the parameter information.

[0026] Exemplarily, the bonding wire production process is divided into different production states, such as a stable state, an adjustment state, an abnormal state, etc., and a corresponding intelligent control strategy is formulated for each production state, such as maintaining parameter stability in a stable state, implementing parameter adjustment in an adjustment state, triggering exception processing in an abnormal state, etc., and then using the strategy determination module 206 to determine the production state corresponding to the target change curve according to the trend and characteristics of the target change curve, thereby obtaining the intelligent control strategy corresponding to the production state, and finally implementing the designed intelligent control strategy into the production process, and realizing real-time monitoring and control of the production process through an automated control system or a human-computer interaction interface.

[0027] In addition, based on actual production conditions and feedback information, the intelligent control strategy is continuously optimized to enable it to better adapt to changes in the production environment and improve production efficiency and product quality. The performance of the production process and the effect of the intelligent control strategy are continuously monitored, and the control algorithm and strategy are continuously optimized to achieve continuous improvement and optimization of the production process.

[0028] Specifically, the present application determines the target change curve based on the initial parameter information and the initial production results, and then obtains the corresponding target production state in the production process of the bonding wire, which is helpful to formulate an intelligent control strategy, thereby improving the automation level and production efficiency of the production process. It also solves the problem that the technology using modeling analysis in the related art cannot accurately identify abnormal conditions in the production process of the bonding wire, thereby affecting product quality and reducing production efficiency.

[0029] In some embodiments, the data segmentation module includes: an initial segmentation submodule, used to obtain an initial segmentation window and an initial moving step, and segment the initial parameter data according to the initial segmentation window and the initial moving step to obtain the initial segmentation data; a rule determination submodule, used to perform serial processing based on the initial segmentation data and related segmentation data adjacent to the initial segmentation data to obtain the fuzzy rule corresponding to the initial parameter data; a window determination submodule, used to determine the target segmentation window and target moving step corresponding to the initial parameter data according to the fuzzy rule; a target segmentation submodule, used to re-segment the initial parameter data according to the fuzzy rule, the target segmentation window and the target moving step to obtain the target segmentation data.

[0030] Exemplarily, the data segmentation module includes an initial segmentation submodule, a rule determination submodule, a window determination submodule, and a target segmentation submodule.

[0031] Exemplarily, the initial segmentation submodule determines the size of the initial segmentation window and the value of the initial moving step corresponding to the initial segmentation window according to specific needs and analysis requirements. The segmentation window refers to the window size sampled in the parameter data, and the moving step refers to the distance the window moves each time. Thus, the initial parameter data is segmented according to the determined initial segmentation window and initial moving step to obtain the corresponding initial segmentation data. The segmentation process can adopt a sliding window method, starting from the starting position of the initial parameter data, sliding the window step by step according to the set moving step, and sampling in each window.

[0032] Exemplarily, the rule determination submodule is used to determine the corresponding adjacent related segmentation data according to the initial segmentation data, and then the initial segmentation data and the related segmentation data are processed in series, that is, they are combined to form a complete data sequence. This can be achieved by connecting adjacent data together or considering their association relationship. A prediction model is established through the initial segmentation data, and then the prediction model is used to infer the adjacent prediction data corresponding to the initial segmentation data, so as to compare the prediction data with the related segmentation data, and then when the error between the prediction data and the related segmentation data is small, the rule corresponding to the prediction model is determined as a fuzzy rule, and when the error between the prediction data and the related segmentation data is large, the prediction model is updated again until the fuzzy rule is obtained. Among them, the fuzzy rule describes the fuzzy relationship and law between the initial segmentation data and the related segmentation data, which can help understand the fuzziness and uncertainty of the initial segmentation data and the related segmentation data.

[0033] Exemplarily, the window determination submodule is used to parse and analyze the fuzzy rules corresponding to each initial segmentation data and related segmentation data, and then the data of the obtained multiple fuzzy rules are merged to obtain merged fuzzy rules, and then the target segmentation window and target moving step are determined according to the merged fuzzy rules.

[0034] Exemplarily, the target segmentation submodule is used to parse and analyze the initial parameter data according to the previously determined fuzzy rules to understand the fuzzy relationships and rules between the data. This will provide guidance to determine the target segmentation window and target moving step required for re-segmenting the data. Based on the parsing results of the fuzzy rules, the target segmentation window and target moving step suitable for describing the fuzzy relationship are determined. Then, the initial parameter data is re-segmented using the determined target segmentation window and target moving step. This means resampling the parameter data according to the new window size and step size to obtain the target segmentation data.

[0035] Specifically, by obtaining the initial segmentation window and the initial moving step, and segmenting the initial parameter data, the data can be integrated and the understanding of the data can be improved. By processing the initial segmentation data and the related segmentation data in series, fuzzy rules can be generated to help understand the fuzzy relationships and laws between the data. Determining the target segmentation window and the target moving step according to the fuzzy rules helps to infer the target parameters suitable for describing the data relationship. By re-segmenting the initial parameter data to obtain the target segmentation data, the data representation can be optimized and better applied to subsequent analysis and processing.

[0036] In some embodiments, the rule determination submodule includes: an aggregation processing submodule, which is used to perform aggregation processing on the initial segmentation data to obtain an aggregation result; a data fusion submodule, which is used to perform information fusion and data fitting based on the aggregation result to obtain a fitting result; a data adjustment submodule, which is used to obtain the associated segmentation data corresponding to the relevant segmentation data from the fitting result, and adjust the initial segmentation window and the initial moving step according to the associated segmentation data and the relevant segmentation data; a rule acquisition submodule, which is used to determine the fuzzy rule based on the adjusted initial segmentation window and the initial moving step.

[0037] Exemplarily, the rule determination submodule includes an aggregation processing submodule, a data fusion submodule, a data adjustment submodule, and a rule acquisition submodule.

[0038] Exemplarily, the aggregation processing submodule first determines an aggregation method suitable for the initial segmented data, and then processes the initial segmented data according to the selected aggregation method. This can be achieved by grouping, merging or calculating the data to obtain an aggregation result.

[0039] Exemplarily, the data fusion submodule fuses the aggregation results with other relevant information to enrich the content and characteristics of the data. This may involve integrating information from different data sources to obtain a more comprehensive view of the data. Based on the fused data, the data is fitted using appropriate data fitting methods, such as regression analysis, machine learning models, etc. This will help identify relationships and trends between the data and generate fitting results.

[0040] Exemplarily, the data adjustment submodule infers the associated segmentation data corresponding to the associated segmentation data based on the fitting results, and then analyzes the difference between the associated segmentation data and the associated segmentation data, and adjusts the initial segmentation window and the moving step size accordingly. This may involve operations such as modifying the size of the segmentation window and adjusting the moving step size. Then, the adjusted initial segmentation window and initial moving step size are verified to ensure that they meet the characteristics and relationships of the data. As needed, the adjustment process is continuously optimized and adjusted to obtain better results. Thus, the initial segmentation window and initial moving step size are obtained.

[0041] Exemplarily, the rule acquisition submodule is used to analyze the data characteristics and relationships reflected by the adjusted initial segmentation window and moving step size. This includes characteristics of data distribution, trend, periodicity, etc. Based on the analyzed data characteristics, suitable fuzzy rules are determined. This may include using fuzzy logic, fuzzy sets and other methods to describe the fuzzy relationships and laws between data.

[0042] Specifically, by performing aggregation processing on the initial segmentation data, the data can be integrated and the complexity of the data can be reduced, making the data easier to analyze and understand. At the same time, information fusion and data fitting based on the aggregation results can optimize the data, reveal the associations and rules between the data, and provide stronger support for subsequent processing and decision-making. Comparing and adjusting the fitting results with the associated segmentation data corresponding to the relevant segmentation data can help discover the associations and influencing factors between the data, so as to understand the data more accurately. Adjusting the initial segmentation window and moving step size can make data processing more accurate and efficient.

[0043] In some embodiments, the abnormality analysis module includes: a limit determination submodule, which is used to determine the target upper limit and the target lower limit corresponding to the target segmentation data according to the fuzzy rule; a range determination submodule, which is used to obtain the running maximum value and the running minimum value corresponding to each parameter in the target segmentation data; an abnormality calculation submodule, which is used to determine the abnormality degree corresponding to the target segmentation data according to the target upper limit, the target lower limit, the running maximum value and the running minimum value; and obtain the abnormality degree according to the following formula:

[0044] ;

[0045] value represents the abnormality degree, x1 represents the target lower limit, x2 represents the target upper limit, x3 represents the running maximum value, x4 represents the running minimum value, and n represents the number of target segmentation data.

[0046] Exemplarily, the anomaly analysis module includes a limit determination submodule, a range determination submodule, and an anomaly calculation submodule.

[0047] Exemplarily, the restriction determination submodule performs fuzzy reasoning on the target segmentation data according to the determined fuzzy rules. This step involves applying the fuzzy rules to the target segmentation data to determine the fuzzy relationship and conclusion between the data. According to the result of the fuzzy reasoning, the fuzzy output corresponding to the target segmentation data can be obtained. This output can be a set of fuzzy sets that describe the possible value range of the target segmentation data. According to the fuzzy output, the target upper limit and target lower limit of the target segmentation data can be determined by the operation of the fuzzy set.

[0048] Exemplarily, the range determination submodule performs numerical comparison on the target segmentation data to obtain a running maximum value and a running minimum value corresponding to each parameter in the target segmentation data.

[0049] Exemplarily, the abnormality calculation submodule obtains the abnormality degree according to the following formula:

[0050] ;

[0051] value indicates the degree of abnormality, x1 indicates the target lower limit, x2 indicates the target upper limit, x3 indicates the running maximum value, x4 indicates the running minimum value, and n indicates the number of target segmentation data.

[0052] Specifically, determining the target upper and lower limits of the target segmentation data, as well as the running maximum and running minimum values ​​corresponding to each parameter according to the fuzzy rules, helps to identify abnormal situations and abnormal behaviors in the data. By comparing the deviation between the actual data and the upper and lower limits, the degree of abnormality of the data can be determined, so that abnormal situations can be discovered and handled in a timely manner.

[0053] In some embodiments, the curve determination module includes: a model building submodule, used to determine any two parameter information from the target parameter data, and build a three-dimensional model based on the two parameter information and the target production results to obtain initial three-dimensional data; a data screening submodule, used to determine the ideal production result, and screen the initial three-dimensional data based on the ideal production result to obtain target three-dimensional data; a curve fitting submodule, used to perform curve fitting on the target three-dimensional data to obtain the target change curve corresponding to the parameter information.

[0054] Exemplarily, the curve determination module includes a model building submodule, a data screening submodule, and a curve fitting submodule.

[0055] Exemplarily, the model building submodule selects any two parameter information from the target parameter data. The two parameters should be representative and important and can have a significant impact on the target production result. Data preparation and cleaning are performed on the selected two parameter information and the target production result to ensure the integrity and accuracy of the data. This may involve steps such as data cleaning, removal of outliers, and missing value processing. A three-dimensional model is established based on the selected two parameter information and the target production result, and then the initial three-dimensional data is determined based on the selected two parameter information and the target production result.

[0056] Exemplarily, the ideal production result is the desired optimal production state or the best output result. This can be a set of specific values ​​or an expected output state, which is used to guide the standard of data screening. The data screening submodule screens the initial three-dimensional data according to the requirements of the ideal production result, screens out the initial three-dimensional data that does not meet the ideal production result, and then determines the three-dimensional data that does not meet the ideal production result as the target three-dimensional data.

[0057] Exemplarily, the curve fitting submodule selects a suitable curve fitting model according to the characteristics and change trends of the data. This may involve different types of function curve fitting models such as linear fitting, polynomial fitting, exponential fitting, logarithmic fitting, etc. The selected fitting model is used to perform curve fitting on the target three-dimensional data. This can be achieved by corresponding functions or libraries in mathematical software or programming languages, such as using the Scipy library in Python for curve fitting operations. The results of the fitted curve are evaluated to check the degree of fit and the quality of fit. The accuracy and reliability of the fit can be evaluated by calculating the fitting error, residual analysis, etc., so as to obtain the target change curve.

[0058] In some embodiments, the strategy determination module includes a correlation analysis submodule, which is used to obtain the degree of correlation between any two of the parameter information, and determine the credibility of the target change curve according to the degree of correlation; a state determination submodule, which is used to determine the initial production state corresponding to each of the target change curves; and a state fusion submodule, which is used to fuse the initial production state according to the credibility to obtain the target production state corresponding to the bonding wire production process.

[0059] Exemplarily, the strategy determination module includes a correlation analysis submodule, a state determination submodule, and a state fusion submodule.

[0060] Exemplarily, the correlation analysis submodule uses appropriate statistical methods (such as Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) to calculate the correlation between any two parameter information. This can help quantify the degree of linear or nonlinear relationship between parameters. According to the results of the correlation analysis, the degree of correlation between any two parameter information is evaluated. The higher the degree of correlation, the more consistent or correlated the change trends between the two parameters are. Based on the degree of correlation and other factors (such as data quality, sample size, etc.), the credibility corresponding to the target change curve is determined. When the degree of correlation is high and the data quality is good, the credibility is usually high.

[0061] Exemplarily, the state determination submodule analyzes the trend and characteristics of each target change curve, including the amplitude, speed, periodicity, etc. of the change. This can help understand the changes in the production state reflected by the curve. According to the characteristics of the target change curve and the expected initial production state, formulate standards and rules for judging the initial state. This may include methods such as setting thresholds and comparing curve characteristics. According to the established judgment criteria, the state of each target change curve is judged, and then the initial production state corresponding to the target change curve is determined according to the trend of the curve and the compliance with the rules.

[0062] Exemplarily, the state fusion submodule performs weighted processing on each initial production state according to its credibility. An initial state with high credibility has a larger weight, and an initial state with low credibility has a smaller weight. The weighted initial production states are fused to obtain the overall target production state. Weighted averaging or other suitable fusion methods can be used to ensure that the contribution of each initial state is proportional to its credibility.

[0063] Specifically, the initial production state is fused according to the credibility to obtain the corresponding target production state in the bonding wire production process. This helps to comprehensively consider the credibility of different parameter information and curves, so as to more accurately describe the state and changes of the production process.

[0064] In some embodiments, the strategy determination module also includes: a state determination submodule, which is used to identify the abnormality of the target parameter data when the target production state is an abnormal state to obtain the corresponding abnormal parameters in the target parameter data; a parameter prediction submodule, which is used to predict the data corresponding to the abnormal parameters based on the target parameter data to obtain target data; and a strategy generation submodule, which is used to determine the intelligent control strategy corresponding to the abnormal parameters based on the target data.

[0065] Exemplarily, the strategy determination module includes a state determination submodule, a parameter prediction submodule, and a strategy generation submodule.

[0066] Exemplarily, when the target production state is abnormal, the state determination submodule selects an appropriate anomaly detection method according to the abnormality definition. Common methods include statistical methods (such as mean, standard deviation, box plot, etc.), machine learning methods (such as clustering, classification, anomaly detection algorithm, etc.), and time series analysis methods. The selected anomaly detection method is used to identify anomalies in the target parameter data. This may involve techniques such as calculating the statistical characteristics of the parameter data, model fitting, and pattern recognition. After identifying the abnormal target parameter data, the corresponding abnormal parameters are determined.

[0067] Exemplarily, the parameter prediction submodule selects an appropriate prediction model according to the characteristics of the data and the prediction requirements. Commonly used models include linear regression, decision tree, support vector machine, neural network, etc. The selected prediction model is trained using historical data. This includes dividing the data into training sets and test sets, and performing parameter tuning and performance evaluation on the model. The trained model is used to predict the data corresponding to the abnormal parameters. This may involve inputting the target parameter data and the abnormal parameter data into the model and obtaining the target data. The target data is the normal data corresponding to the abnormal parameters.

[0068] Exemplarily, the strategy generation submodule selects an appropriate control algorithm to implement the intelligent control strategy. Depending on the specific situation, different algorithms such as PID control, fuzzy control, and neural network control can be selected. The selected control algorithm is parameter tuned to ensure that the control value corresponding to the abnormal parameter can be the target data in actual application, and the control strategy and parameters are adjusted in time according to the monitoring results to maintain the stability and reliability of the production process.

[0069] Specifically, the intelligent control strategy corresponding to the abnormal parameters is determined according to the target data. This helps to deal with abnormal situations in a timely manner during the production of bonding wires and ensure the stability, safety and efficiency of production.

[0070] In some embodiments, the parameter prediction submodule includes: a parameter determination submodule, used to obtain a first parameter and a second parameter other than the abnormal parameter from the target parameter data; a curve acquisition submodule, used to obtain a first association curve corresponding to the first parameter and the abnormal parameter and a second association curve corresponding to the second parameter and the abnormal parameter; a first prediction submodule, used to obtain the first data corresponding to the abnormal parameter from the first association curve according to the target parameter data; a second prediction submodule, used to obtain the second data corresponding to the abnormal parameter from the second association curve according to the target parameter data; and a fusion prediction submodule, used to determine the target data corresponding to the abnormal parameter based on the first data and the second data.

[0071] Exemplarily, the parameter prediction submodule includes a parameter determination submodule, a curve acquisition submodule, a first prediction submodule, a second prediction submodule, and a fusion prediction submodule.

[0072] Exemplarily, the parameter determination submodule screens the target parameter data, excludes data corresponding to abnormal parameters, and retains other parameter data, and determines the first parameter and the second parameter in the target parameter data.

[0073] Exemplarily, the curve acquisition submodule uses a statistical or machine learning method to perform correlation analysis between parameters. Correlation analysis, regression analysis or other association rule mining techniques can be used to find the correlation between parameters. A first correlation curve between the first parameter and the abnormal parameter, and a second correlation curve between the second parameter and the abnormal parameter are determined based on the result of the correlation analysis.

[0074] Exemplarily, the first prediction submodule analyzes the first correlation curve to ensure that the correlation between the abnormal parameter and the first parameter is understood so as to accurately obtain the corresponding data, and then on the first correlation curve, according to the numerical value corresponding to the abnormal parameter in the target parameter data, find its corresponding position on the curve. The first data corresponding to the abnormal parameter can be found by interpolation, fitting or other mathematical methods.

[0075] Exemplarily, the second prediction submodule analyzes the second correlation curve to ensure that the correlation between the abnormal parameter and the second parameter is understood so as to accurately obtain the corresponding data, and then on the second correlation curve, according to the value corresponding to the abnormal parameter in the target parameter data, find its corresponding position on the curve. The second data corresponding to the abnormal parameter can be found by interpolation, fitting or other mathematical methods.

[0076] Exemplarily, the fusion prediction submodule fuses the first data and the second data to obtain target data corresponding to the abnormal parameters.

[0077] Specifically, by combining the first data and the second data, the target data corresponding to the abnormal parameters can be determined more comprehensively, thereby providing good support for subsequently obtaining accurate intelligent control strategies.

[0078] In some embodiments, the strategy determination module also includes: an edge processing submodule, which is used to extract edges from the target change curve to obtain a target edge curve when the target production state is normal; and a curve updating submodule, which is used to update the curves corresponding to any two parameter information in the target parameter data according to the target edge curve to obtain a target association curve.

[0079] Exemplarily, the strategy determination module further includes an edge processing submodule and a curve updating submodule.

[0080] Exemplarily, when the target production status is normal, the edge processing submodule indicates that the production process is in normal operation. At this time, the target change curve can be edge extracted to obtain the target edge curve, and then the ideal curve when the production status is normal can be obtained according to the target edge curve.

[0081] Exemplarily, the curve updating submodule analyzes the target edge curve to understand the relationship and change rule between it and any two parameter information in the target parameter data. According to the characteristics of the target edge curve and the update requirements, a suitable curve updating method is selected. According to the selected update method, the curve corresponding to any two parameter information in the target parameter data is updated. Ensure that the updated curve can more accurately reflect the relationship between the parameter information. Compare and analyze the updated curve with other related curves to ensure that the target correlation curve is obtained. This correlation curve can reflect the new relationship and trend between the updated parameter information.

[0082] Specifically, the curves corresponding to any two parameter information in the target parameter data are updated according to the target edge curve, and finally the target association curve is obtained, so as to more accurately understand the association and influence between the parameters and achieve the purpose of real-time updating.

[0083] In some embodiments, the system also includes: a quality determination module, used to determine the product quality corresponding to the target production result based on the target production status; a product strategy module, used to obtain the production range corresponding to the abnormal product from the target production result based on the product quality, and determine the product processing strategy based on the production range.

[0084] Exemplarily, the intelligent gold bonding wire control system based on big data analysis also includes a quality determination module and a product strategy module.

[0085] Exemplarily, the quality determination module establishes a corresponding relationship between the target production state and the product quality according to the target production state. For example, when the target production state is abnormal, the product quality is abnormal, and when the target production state is normal, the product quality is normal. Then, the product quality corresponding to the target production result is determined according to the target production state and the corresponding relationship.

[0086] For example, the product strategy module determines the production range corresponding to the abnormal products obtained in the target production results according to the product quality. The production range can be a product number range or a production time range corresponding to the target production results. Then the product strategy module determines the product processing strategy according to the production range. The product processing strategy is to re-test or recycle the corresponding products in the production range.

[0087] Specifically, by determining the association between the target production status and product quality, the production process can be monitored and controlled more accurately, thereby improving the quality management efficiency of the product. Establishing an association model between quality and production range can timely identify the production range corresponding to abnormal products, which helps to timely warn and take measures during the production process, thereby reducing the generation of abnormal products.

[0088] An embodiment of the present invention also provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement any step of the intelligent control system for bonding gold wires based on big data analysis as provided in the description of the embodiment of the present invention.

[0089] The storage medium may be an internal storage unit of the terminal device in the aforementioned embodiment, such as a hard disk or memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal device.

[0090] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware embodiment, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0091] It should be understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0092] The serial numbers of the embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. An intelligent control system for gold bonding wire based on big data analysis, characterized in that: The system comprises: A data collection module, used to collect initial parameter data corresponding to the bonding wire production process and initial production results corresponding to the initial parameter data; A data segmentation module, used for segmenting the initial parameter data according to fuzzy rules to obtain target segmentation data corresponding to the initial parameter data; An abnormality analysis module, used to analyze the abnormality degree of the target segmentation data to obtain the abnormality degree corresponding to the target segmentation data; A data screening module, used to determine target parameter data according to the abnormality degree, and obtain a target production result corresponding to the target parameter data from the initial production result; A curve determination module, used to determine any two parameter information from the target parameter data, and determine a target change curve corresponding to the parameter information according to the any two parameter information and the target production result; A strategy determination module, used to determine a corresponding target production state in a bonding wire production process according to the target change curve, and determine an intelligent control strategy according to the target production state; Wherein, the data segmentation module includes: An initial segmentation submodule, used to obtain an initial segmentation window and an initial moving step length, and to segment the initial parameter data according to the initial segmentation window and the initial moving step length to obtain initial segmentation data; A rule determination submodule, configured to perform serial processing on the initial segmentation data and the related segmentation data adjacent to the initial segmentation data to obtain the fuzzy rule corresponding to the initial parameter data; A window determination submodule, used to determine the target segmentation window and target moving step length corresponding to the initial parameter data according to the fuzzy rule; The target segmentation submodule is used to re-segment the initial parameter data according to the fuzzy rule, the target segmentation window and the target moving step to obtain the target segmentation data.

2. The system according to claim 1, characterized in that The rule determination submodule includes: An aggregation processing submodule, used to perform aggregation processing on the initial segmented data to obtain an aggregation result; A data fusion submodule, used to perform information fusion and data fitting according to the aggregation result to obtain a fitting result; A data adjustment submodule, configured to obtain associated segmentation data corresponding to the associated segmentation data from the fitting result, and adjust the initial segmentation window and the initial moving step size according to the associated segmentation data and the associated segmentation data; The rule acquisition submodule is used to determine the fuzzy rule according to the adjusted initial segmentation window and the initial moving step length.

3. The system according to claim 1, characterized in that The abnormality analysis module comprises: A limit determination submodule, used for determining a target upper limit and a target lower limit corresponding to the target segmentation data according to the fuzzy rule; A range determination submodule, used to obtain a running maximum value and a running minimum value corresponding to each parameter in the target segmentation data; an abnormality calculation submodule, used to determine the abnormality degree corresponding to the target segmentation data according to the target upper limit, the target lower limit, the running maximum value and the running minimum value; The abnormality degree is obtained according to the following formula: value represents the abnormality degree, x1 represents the target lower limit, x2 represents the target upper limit, x3 represents the running maximum value, x4 represents the running minimum value, and n represents the number of target segmentation data.

4. The system according to claim 1, characterized in that The curve determination module comprises: A model building submodule, used to determine any two parameter information from the target parameter data, and to build a three-dimensional model according to the two parameter information and the target production result to obtain initial three-dimensional data; A data screening submodule, used to determine an ideal production result, and screen the initial three-dimensional data according to the ideal production result to obtain target three-dimensional data; The curve fitting submodule is used to perform curve fitting on the target three-dimensional data to obtain the target change curve corresponding to the parameter information.

5. The system according to claim 4, characterized in that The strategy determination module includes A correlation analysis submodule, used to obtain the correlation degree between any two pieces of parameter information, and determine the credibility of the target change curve according to the correlation degree; A state determination submodule, used to determine the initial production state corresponding to each target change curve; The state fusion submodule is used to fuse the initial production state according to the credibility to obtain the target production state corresponding to the bonding wire production process.

6. The system according to claim 1, characterized in that The strategy determination module further includes: A state determination submodule, used for, when the target production state is an abnormal state, performing abnormal identification on the target parameter data to obtain corresponding abnormal parameters in the target parameter data; A parameter prediction submodule, used to predict the data corresponding to the abnormal parameter according to the target parameter data to obtain target data; A strategy generation submodule is used to determine the intelligent control strategy corresponding to the abnormal parameter according to the target data.

7. The system according to claim 6, characterized in that The parameter prediction submodule comprises: A parameter determination submodule, used for obtaining a first parameter and a second parameter except the abnormal parameter from the target parameter data; A curve acquisition submodule, used to obtain a first correlation curve corresponding to the first parameter and the abnormal parameter and to obtain a second correlation curve corresponding to the second parameter and the abnormal parameter; A first prediction submodule, configured to obtain first data corresponding to the abnormal parameter from the first correlation curve according to the target parameter data; A second prediction submodule, configured to obtain second data corresponding to the abnormal parameter from the second association curve according to the target parameter data; A fusion prediction submodule is used to determine the target data corresponding to the abnormal parameter based on the first data and the second data.

8. The system according to claim 1, characterized in that The strategy determination module further includes: An edge processing submodule, for performing edge extraction on the target change curve to obtain a target edge curve when the target production state is a normal state; The curve updating submodule is used to update the curves corresponding to any two parameter information in the target parameter data according to the target edge curve to obtain a target association curve.

9. The system according to claim 1, characterized in that The system further comprises: A quality determination module, used to determine the product quality corresponding to the target production result according to the target production status; A product strategy module is used to obtain a production range corresponding to abnormal products from the target production results according to the product quality, and determine a product processing strategy according to the production range.

Images