Method, medium and system for capturing BOP data to generate EWI

Abstract

The invention provides a method, medium and system for generating EWI through BOP data capturing, and belongs to the technical field of production business data.The method includes the steps that firstly, a multi-dimensional BOP data load matrix is built, contribution values of all parameters are accurately calculated through a thermodynamic entropy equation, then a standard EWI basic matrix is built, and the standard EWI basic matrix is obtained; and utilizing a minimum spanning tree algorithm to optimize a BOP data capture strategy and calculating a data integrity score. In a data processing stage, an EWI mergeable index is accurately generated through a Lagrange multiplier method, a transverse error is calculated by adopting an Euclidean distance function so as to analyze a spatial relationship among parameters, a longitudinal error is calculated by applying Fourier transform so as to monitor a time change trend of the parameters, and the two kinds of errors are synthesized to form a mixed error evaluation result. And finally, combining the EWI mergeable index to output a comprehensive index for accurately reflecting the production quality state.

Description

Technical Field

[0001] The present invention belongs to the technical field of production business data, and in particular, relates to a method, medium and system for capturing BOP data and generating EWI. Background Art

[0002] In industrial production, Early Warning Indicator (EWI) systems are key tools for assessing and controlling production quality. Traditionally, they are constructed by collecting and analyzing data from key bill of process (BOP) parameters during the production process. Currently, mainstream EWI construction methods rely primarily on simple threshold monitoring and linear weighted summation, comparing acquired production parameters with preset standards and triggering an early warning mechanism when the parameters deviate from the predetermined threshold. In industries with extremely high production precision requirements, such as semiconductor manufacturing, precision machining, pharmaceutical production, and chemical synthesis, these systems are widely used to monitor production quality status in real time.

[0003] However, traditional EWI construction methods face the problem of insufficiently accurate EWI generation from BOP data capture. Existing technologies typically process BOP data using empirical formulas or fixed-weight models, which fail to accurately reflect the actual contribution of each parameter to quality status at different production stages and under different operating conditions. Especially when processing high-dimensional BOP data, traditional linear weighting methods struggle to capture the complex nonlinear relationships between parameters, resulting in significant deviations between the generated EWI indicators and the actual production quality status. Furthermore, existing methods lack systematic optimization strategies during data capture, making it difficult to ensure the most representative set of data points, which impacts the accuracy of subsequent EWI indicators.

[0004] In modern high-precision manufacturing environments, the lack of accuracy in generating EWI from BOP data has become a key bottleneck restricting the effectiveness of quality control systems. Current technologies struggle to simultaneously consider the changing characteristics of BOP parameters in both the spatial dimension (the relationship between parameters) and the temporal dimension (the trend of parameter evolution). They also lack a systematic error analysis method and are unable to accurately characterize the quality status of the production process. Particularly in complex scenarios where production conditions fluctuate and multiple parameters influence each other, the accuracy issues of traditional BOP data capture and EWI generation methods are even more prominent. There is an urgent need for a method that can significantly improve the accuracy of generating EWI from BOP data capture. In other words, the existing technology suffers from the technical problem of insufficient accuracy in generating EWI from BOP data capture. Summary of the Invention

[0005] In view of this, the present invention provides a method, medium and system for capturing BOP data to generate EWI, which can solve the technical problem in the prior art that generating EWI by capturing BOP data is not accurate enough.

[0006] The present invention is implemented as follows: In a first aspect, the present invention provides a method for capturing BOP data and generating EWI, including: constructing a multidimensional BOP data load matrix; calculating the BOP data contribution value, and using the thermodynamic entropy equation to analyze the influence of each BOP data dimension; establishing an EWI basic matrix; implementing BOP data capture rate calculation, and using the minimum spanning tree algorithm to obtain the optimized data capture path and BOP data integrity score; generating an EWI mergeability index, and weightedly fusing the matching degree of each dimension of BOP data with the EWI basic matrix through the Lagrange multiplier method; calculating the lateral error, and using the Euclidean distance function to analyze the degree of deviation between different BOP parameters at the same time point; calculating the longitudinal error, and using the Fourier transform equation to track the change trend of a single BOP parameter in the time series; generating a hybrid error evaluation; and generating a comprehensive indicator reflecting the production quality status as the final EWI indicator based on the hybrid error evaluation result and the EWI mergeability index and outputting it.

[0007] The BOP data load matrix is ​​a multidimensional data structure that constructs key parameters in the production process according to time and type, and is used to represent the load status of each parameter at different times and their mutual influence relationship.

[0008] Among them, the BOP data contribution value is the influence weight of each BOP data on the formation process of the final EWI index, which is calculated by the thermodynamic entropy equation and reflects the degree of influence of different parameters on production quality.

[0009] The EWI basic matrix is ​​a reference standard for quality control of the production process, which contains the standard values ​​and fluctuation thresholds of each parameter and is used as a benchmark for data analysis.

[0010] The BOP data capture rate is the percentage of valid BOP data successfully acquired during the production monitoring process to the total BOP data that should be acquired theoretically, reflecting the reliability of the data acquisition system.

[0011] Among them, the EWI combinable index is a process of integrating multi-dimensional parameter BOP data into a single indicator through the Lagrange multiplier method, which facilitates the intuitive evaluation of the overall production status.

[0012] Among them, the lateral error is the relative deviation between different BOP parameters at the same time point, which is used to identify abnormal correlations between BOP parameters; the longitudinal error is the change deviation of a single BOP parameter in the time series, which is used to monitor whether the change trend of the BOP parameter meets expectations; the mixed error is a comprehensive indicator obtained by fusing the lateral error and the longitudinal error according to specific weights, which comprehensively reflects the quality status of the production process.

[0013] Among them, the hybrid error assessment is to conduct a comprehensive analysis of the lateral error and the longitudinal error. Specifically, a comprehensive quality evaluation system is constructed based on the distance matrix between parameters, abnormal correlation identification, parameter energy distribution and abnormal frequency points to analyze and obtain the hybrid error assessment results.

[0014] A second aspect of the present invention provides a computer-readable storage medium having program instructions stored therein. When the program instructions are executed in a computer, the method for capturing BOP data and generating EWI is executed.

[0015] The third aspect of the present invention provides a system for capturing BOP data and generating EWI, which includes the above-mentioned computer-readable storage medium. The system is any one of a computer, a server, and a single-chip microcomputer. The computer-readable storage medium is set in the system, and the system is provided with a microprocessor that executes the program instructions stored in the computer-readable storage medium.

[0016] The present invention constructs a multidimensional BOP data load matrix, uses the thermodynamic entropy equation to scientifically calculate the contribution value of each parameter, combines the minimum spanning tree algorithm to optimize the data capture path, and uses the Lagrange multiplier method to accurately generate the comprehensive EWI index, thereby achieving high-precision monitoring of the production quality status.

[0017] This method addresses the lack of accuracy inherent in conventional techniques, often caused by arbitrary and subjective parameter weighting, through a weight allocation mechanism based on information entropy theory. Specifically, it introduces a dual-dimensional analysis system for lateral and longitudinal errors, accurately assessing the spatial relationship and temporal evolution of parameters using the Euclidean distance function and Fourier transform, respectively. This overcomes the limitations of conventional methods, which struggle to simultaneously capture complex relationships between parameters and temporal trends, significantly improving the accuracy of the EWI indicator.

[0018] Through this multi-dimensional, multi-algorithm integrated processing method, the present invention effectively solves the technical problem of inaccurate EWI generation from BOP data capture in the existing technology, so that the generated EWI indicators can accurately reflect the actual quality status of the production process, providing a reliable basis for precise quality control and timely intervention in the production process, and greatly improving the accuracy and effectiveness of quality management. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of the method of the present invention.

[0020] Figure 2 The graph shows the change of the normalized values ​​of the five key parameters in Example 2 over time.

[0021] Figure 3This is the EWI index change curve calculated based on the five key parameters in Example 2.

[0022] Figure 4 This is a spectrum analysis result diagram of the P4 parameter (RF bias) in Example 2.

[0023] Figure 5 This is the evaluation diagram of the lateral error, longitudinal error and their mixed error in Example 2. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0025] like Figure 1 FIG. 1 is a flowchart of a method for capturing BOP data and generating EWI according to the first aspect of the present invention. The method includes the following steps:

[0026] S01. Construct a multi-dimensional BOP data load matrix, classify and organize the BOP data collected during the production process according to time series and parameter types, and form a standardized multi-dimensional matrix structure;

[0027] S02. Calculate the BOP data contribution value. Use the thermodynamic entropy equation to analyze the influence of each BOP data dimension. Assign a weight coefficient to each BOP data dimension based on parameter fluctuation amplitude, fluctuation frequency, system response time, system self-organizing ability, and environmental interference factors. Normalize the data to ensure that the sum of the BOP data contribution values ​​of each dimension is 1.

[0028] S03. Establish an EWI basic matrix, set a benchmark threshold based on historical production data, and construct an EWI basic matrix containing standard parameters and allowable fluctuation ranges;

[0029] S04. Calculate the BOP data capture rate by using a minimum spanning tree algorithm to monitor the ratio of the actual amount of valid BOP data captured during the production process to the theoretical amount of BOP data that should be captured based on data point distribution density, data point weights, inter-point connection costs, network topology, and system resource constraints, and obtain the optimal data capture path and BOP data integrity score.

[0030] S05. Generate the EWI mergibility index by weighting and fusing the matching degree of each dimension of BOP data with the EWI basic matrix through the Lagrange multiplier method based on the upper and lower weight constraints, the correlation matrix between parameters, the historical accuracy evaluation, the objective function type, and the convergence threshold to form a single measurement indicator;

[0031] S06. Calculate the lateral error and use the Euclidean distance function based on the parameter vector group, dimension weight coefficient, reference standard vector, allowable deviation range, and correlation threshold to analyze the degree of deviation between different BOP parameters at the same time point, and obtain the distance matrix between the parameters and the abnormal correlation mark;

[0032] S07. Calculate the longitudinal error by using the Fourier transform equation to track the changing trend of a single BOP parameter in the time series based on the time series parameter value, sampling interval, frequency resolution, window function type, and spectrum smoothing coefficient, and obtain the energy distribution of the parameter at different frequencies and abnormal frequency points;

[0033] S08. Generate a mixed error assessment, conduct a comprehensive analysis of the lateral error and the longitudinal error, specifically, build a comprehensive quality evaluation system based on the distance matrix between parameters, abnormal correlation identification, parameter energy distribution, and abnormal frequency points, and analyze and obtain the mixed error assessment result;

[0034] S09. Output the final EWI index, that is, generate a comprehensive index reflecting the production quality status based on the mixed error evaluation result and the EWI combinable index.

[0035] Among them, the BOP data load matrix refers to a multidimensional data structure that constructs key parameters in the production process according to time and type, and is used to represent the load status of each parameter at different times and their mutual influence relationship.

[0036] Among them, the BOP data contribution value refers to the influence weight of each BOP data on the formation of the final EWI index, which is calculated through the thermodynamic entropy equation and reflects the degree of influence of different parameters on production quality.

[0037] Among them, the EWI basic matrix is ​​the reference standard for production process quality control, which contains the standard values ​​and fluctuation thresholds of each parameter and is used as the benchmark point for data analysis.

[0038] Among them, the BOP data capture rate refers to the percentage of valid BOP data successfully obtained during the production monitoring process to the total BOP data that should be obtained theoretically, reflecting the reliability of the data acquisition system.

[0039] Among them, the EWI combinable index is a process of integrating multidimensional parameter BOP data into a single index through the Lagrange multiplier method, which facilitates the intuitive evaluation of the overall production status.

[0040] Among them, the lateral error refers to the relative deviation between different BOP parameters at the same time point, which is used to identify abnormal correlations between BOP parameters; the longitudinal error refers to the change deviation of a single BOP parameter in the time series, which is used to monitor whether the change trend of the BOP parameter is in line with expectations; the mixed error is a comprehensive indicator obtained by fusing the lateral error and the longitudinal error according to specific weights, which comprehensively reflects the quality status of the production process.

[0041] Among them, the inter-parameter distance matrix is ​​a matrix representation of the distance between BOP parameters calculated using the Euclidean distance function, which is used for mixed error assessment; the abnormal correlation identifier is an indicator that marks abnormal correlation between BOP parameters, which is used to identify quality anomalies; the parameter energy distribution is the energy distribution of BOP parameters in the frequency domain obtained by Fourier transform, which is used to identify abnormal patterns; the abnormal frequency point is the frequency position in the BOP parameter spectrum that is significantly different from the normal pattern, indicating potential anomalies.

[0042] The thermodynamic entropy equation is used to analyze the contribution of different BOP parameters in a production system to system stability. The inputs include parameter fluctuation amplitude, fluctuation frequency, system response time, system self-organizing ability, and environmental interference factors, and the output is the entropy contribution coefficient of each BOP parameter. Among them, the parameter fluctuation amplitude refers to the difference between the maximum and minimum values ​​of the BOP data within a certain time window, which is used to measure the stability of the BOP data; the fluctuation frequency refers to the number of times the BOP parameter changes in unit time, reflecting the sensitivity of the system response; the system response time refers to the time interval required from the input change to the corresponding change in the system output, which measures the reaction speed of the system; the system self-organizing ability refers to the ability indicator of the production system to maintain a stable state under external disturbances, reflecting the robustness of the system; the environmental interference factor refers to the quantitative representation of external environmental factors that affect the accuracy of the BOP data, including temperature, humidity, vibration, etc.

[0043] The Fourier transform equation is used to convert BOP parameter changes in the time domain into frequency domain analysis, identifying the periodic characteristics and abnormal patterns of BOP parameter fluctuations. The input includes time series parameter values, sampling intervals, frequency resolutions, window function types, and spectrum smoothing coefficients. The output is the energy distribution and abnormal frequency points of the BOP parameters at different frequencies. Among them, the time series parameter values ​​are a single BOP parameter numerical sequence arranged in chronological order and are the input data of the Fourier transform. Among them, the sampling interval is the time interval for obtaining BOP data, which affects the frequency range of the Fourier transform. The frequency resolution is the minimum frequency difference that can be distinguished in the Fourier transform, which affects the degree of analysis precision. The window function type is a function form for weighting the time series data before the Fourier transform to reduce spectrum leakage. The spectrum smoothing coefficient is a parameter for smoothing the Fourier transform results to reduce the impact of noise.

[0044] The minimum spanning tree algorithm is used to optimize the BOP data capture strategy to ensure that the most representative set of BOP data points is obtained under limited resources. The input includes data point distribution density, data point weight, connection cost between points, network topology, and system resource limitations. The output is the optimized data capture path and BOP data integrity score. Among them, data point distribution density refers to the distribution of the number of BOP data points per unit space or time, which is used for weight calculation of the minimum spanning tree algorithm; data point weight is a weighting coefficient assigned according to the importance of BOP data, which is used to optimize the data capture strategy; connection cost between points refers to the resource consumption required to connect two data points in the data capture network, which is used to construct the minimum spanning tree; network topology refers to the description of the connection relationship between nodes in the data acquisition network, which provides the basis for the minimum spanning tree algorithm; system resource limitations refer to the constraints of the data acquisition system on processing power, storage space and communication bandwidth, which affect the data capture strategy.

[0045] The Euclidean distance function is used to calculate the similarity and degree of deviation between different dimensions in the BOP parameter space. The input includes a parameter vector group, a dimension weight coefficient, a reference standard vector, an allowable deviation range, and a correlation threshold. The output is a distance matrix between parameters and an abnormal association identifier. Among them, the parameter vector group is a vector set consisting of multiple BOP parameters at the same time point, which is used for Euclidean distance function calculation; the dimension weight coefficient is the importance coefficient assigned to different BOP parameter dimensions in the Euclidean distance function calculation, which affects the distance calculation result; the reference standard vector is the BOP parameter vector under ideal production conditions, which serves as the benchmark for calculating the degree of deviation; the allowable deviation range is the maximum limit to which the BOP parameter is allowed to deviate from the reference standard vector, which is used to judge the abnormal state; the correlation threshold is the critical value for judging whether the association between BOP parameters is abnormal, which is used to screen effective associations.

[0046] The Lagrange multiplier method is used to solve the optimal fusion weight of the EWI index under multiple constraints. The input includes upper and lower weight constraints, parameter correlation matrix, historical accuracy evaluation, objective function type, and convergence threshold. The output is the optimal weight distribution scheme that meets all constraints. Among them, the upper and lower weight constraints refer to the limitation of the value range of each BOP parameter weight in the Lagrange multiplier method to ensure reasonable weight distribution. The parameter correlation matrix is ​​a numerical representation that quantifies the relationship between different BOP parameters and is used to calculate the EWI combinable index. The historical accuracy evaluation is the evaluation result of the prediction accuracy of each BOP parameter based on historical data, which is used to adjust the weight distribution. The objective function type refers to the mathematical function form used to optimize the weight in the Lagrange multiplier method, which affects the convergence of the optimization result. The convergence threshold is the judgment standard for judging whether the iterative calculation of the Lagrange multiplier method has reached the convergence state, which affects the calculation efficiency and accuracy.

[0047] The specific implementation of the above steps is described in detail below.

[0048] The specific implementation of step S01 involves constructing a multidimensional BOP data load matrix. First, key parameter data from the production process is collected and arranged into a time series format according to timestamps. Parameter types are then classified and organized, with similar parameters grouped into the same dimension to form a preliminary parameter classification structure. Next, data format standardization is performed, converting parameter data of different dimensions into a dimensionless form. Maximum and minimum normalization is employed to ensure that all data values ​​fall within the interval [0, 1]. A matrix index structure is then established, with time as the first dimension index and parameter type as the second, thus forming a two-dimensional or multidimensional data structure. Finally, a data integrity check is performed to address missing values. Linear interpolation can be used to fill missing data with short time intervals, or forward filling can be used to handle consecutive missing values. This series of processes ultimately creates a standardized multidimensional BOP data load matrix, providing a structured data foundation for subsequent analysis.

[0049] The specific implementation of step S02 is to calculate the contribution value of the BOP data and analyze the impact of each dimension of data on the system based on the thermodynamic entropy equation. First, the parameter fluctuation amplitude is calculated by calculating the standard deviation of each parameter within a fixed time window (e.g., 10 minutes) as the fluctuation amplitude indicator. The fluctuation frequency is then measured by analyzing the main frequency components of the parameter per unit time using a Fourier transform to obtain the frequency of significant changes. Next, the system response time is evaluated by calculating the time delay from parameter change to system output response using a cross-correlation function. Generally, a faster response time (e.g., less than 5 seconds) indicates a greater parameter impact. The system's self-organizing ability is also measured by calculating the system's ability to recover to a stable state after a disturbance using the Lyapunov exponent. The exponent value typically ranges from [-1 to 1], with closer to -1 indicating stronger system self-organizing ability. Environmental interference factors are then quantified based on the influence of environmental parameters such as temperature (standard operating temperature deviation does not exceed ±5°C), humidity (relative humidity is controlled within the range of 45% to 65%), and vibration (amplitude does not exceed 0.1mm). The entropy weight method is used to calculate the entropy value of each dimensional parameter. According to the principle of information entropy, parameters with large fluctuations and poor regularity contain more information. Finally, normalization is performed to ensure that the sum of the contribution values ​​of the BOP data of all dimensions is 1. This forms a weight distribution scheme and provides a basis for the subsequent EWI indicator calculation.

[0050] The specific implementation of step S03 involves establishing an EWI foundational matrix, forming a benchmark framework for quality assessment based on historical production data. First, historical high-quality production batch data is screened, and the best-performing production cycle data (typically batches with quality inspection scores in the top 10%) are selected based on product quality inspection results. Baseline values ​​for each parameter are then calculated, and the parameter means corresponding to these high-quality batches are used as standard reference points. Next, parameter fluctuation thresholds are determined. By calculating the standard deviation of the parameters for high-quality batches, the allowable fluctuation range is set to ±3 times the standard deviation. Fluctuations within this range are considered normal. A parameter correlation model is then established, using the Pearson correlation coefficient to analyze the relationships between different parameters. The correlation coefficient threshold is typically set at 0.7, with parameter pairs above this value considered strongly correlated. A threshold matrix is ​​then constructed, organizing the baseline values ​​and allowable fluctuation ranges for each parameter into a matrix format, forming a standardized basis for judgment. Finally, verification and adjustment are performed, using additional historical data to verify the applicability of the thresholds and, if necessary, fine-tuning to ensure that the EWI foundational matrix accurately reflects the quality status of the production process.

[0051] The specific implementation of step S04 involves calculating the BOP data capture rate and optimizing the data collection strategy using a minimum spanning tree algorithm. First, the data point distribution density is analyzed. The kernel density estimation method is used to calculate the data distribution density function in the production parameter space and identify high-density areas. Next, data point weights are determined. Each data point is assigned a weight based on its parameter contribution, typically in the range of [0.1, 1.0], reflecting its importance. Next, the cost of connecting points is calculated, taking into account the system resource consumption required to collect different data points, including computing resources, storage resources, and communication bandwidth, to form a cost matrix. Subsequently, the network topology is analyzed to establish a connection diagram between data collection nodes and determine feasible data transmission paths. Furthermore, considering system resource limitations, resource constraints for data collection are set based on the actual hardware configuration, such as processor utilization not exceeding 80%, storage space utilization not exceeding 75%, and communication bandwidth utilization not exceeding 90%. A minimum spanning tree algorithm is then applied, using either the Kruskal algorithm or the Prim algorithm, to construct a minimum-cost tree connecting the data points, ensuring maximum data collection efficiency within resource constraints. Finally, the data capture rate indicator is calculated. The ratio of the number of valid data points actually collected to the theoretical number of data points that should be collected is used as the BOP data capture rate to obtain a score reflecting the data integrity. The ideal capture rate should usually be maintained above 95%.

[0052] The specific implementation of step S05 involves generating an EWI combinability index and employing the Lagrange multiplier method to achieve a comprehensive assessment of multidimensional parameters. First, upper and lower weight constraints are set, determining the weight range based on the importance of each parameter. Typically, the lower limit for core parameter weights is no less than 0.1, and the upper limit is no more than 0.5, ensuring a reasonable weight distribution. A parameter correlation matrix is ​​then constructed, and the correlation coefficients between different BOP parameters are calculated, forming a matrix structure that characterizes the mutual influence between parameters. Next, historical accuracy is assessed. Backtesting is performed to analyze the accuracy of each parameter's prediction of the historical EWI indicator. Accuracy is typically scored using the root mean square error (RMSE), ideally less than 0.05. An objective function type is then selected, and an optimization objective function is set based on the principle of minimizing the overall prediction error. Mean square error or cross-entropy loss functions are typically used. A convergence threshold is then determined, and the iterative calculation termination criteria are set. For example, convergence is considered when the weight change between two consecutive iterations is less than 0.001. Next, the Lagrange multiplier method is applied to solve the optimal weight distribution solution while satisfying all constraints. By introducing Lagrange multipliers, the constrained optimization problem is transformed into an unconstrained one. Finally, the EWI mergibility index is calculated, and the obtained optimal weight is weighted and fused with the matching degree of each dimension BOP data and the EWI basic matrix to form a single index in the range of [0, 1]. Among them, above 0.8 indicates good production status, 0.6-0.8 indicates acceptable production status, and below 0.6 indicates abnormal production status.

[0053] The specific implementation of step S06 involves calculating lateral errors and analyzing the relationships between different parameters at the same time point using the Euclidean distance function. First, a parameter vector group is organized, combining multiple BOP parameter values ​​at the same time point into a vector form to construct a parameter space. Dimension weight coefficients are then determined, based on the importance of each parameter, to reflect the importance of each dimension in the distance calculation. A reference standard vector is then selected, using the parameter vector under high-quality production conditions as a benchmark. Typically, data from the best historical production batch is used as the reference. An allowable deviation range is then set. The allowable deviation from the standard vector value is determined based on production process requirements, typically within ±5% of the standard vector value. A correlation threshold is then determined, setting a critical value for determining parameter correlation anomalies. A value below this threshold is typically set at 0.75; values ​​below this threshold are considered abnormal. Next, a weighted Euclidean distance is calculated, using a formula to calculate the weighted Euclidean distance between the actual parameter vector and the reference standard vector, reflecting the overall degree of deviation. A distance matrix is ​​then constructed, calculating the distances between different parameter pairs to form a complete distance matrix. Finally, an abnormal correlation marker is generated. By comparing the calculated distances with a set threshold, parameter pairs with abnormal correlations are identified, providing lateral analysis results for mixed error assessment.

[0054] The specific implementation of step S07 is to calculate the longitudinal error and use the Fourier transform equation to track the time variation characteristics of the parameters. First, prepare the time series parameter values ​​and arrange the single BOP parameters in chronological order to form a discrete time series. Then determine the sampling interval and set the sampling time interval according to the data acquisition frequency, usually in milliseconds (such as 100ms) or seconds (such as 1s). Then select the frequency resolution and set the fineness of the Fourier transform. Usually, a resolution of not less than 0.01Hz is selected to ensure that the key frequency components can be captured. Then select the window function type and apply window functions such as Hanning window or Hamming window to weight the original signal to reduce the spectrum leakage effect. Then set the spectrum smoothing coefficient. Usually, a smoothing coefficient of 0.05 to 0.2 is selected to reduce the impact of random noise on spectrum analysis. Next, perform a fast Fourier transform to convert the time domain signal into a frequency domain representation and obtain the frequency components and their amplitudes. Then analyze the energy distribution characteristics, calculate the energy proportion of different frequency intervals, and determine the main frequency characteristics of the parameter changes. Finally, the abnormal frequency points are identified, the obtained spectrum is compared with the standard spectrum under normal working mode, and the frequency points where the energy is abnormally concentrated or significantly deviated are marked, providing a longitudinal analysis basis for the mixed error assessment.

[0055] The specific implementation of step S08 involves generating a hybrid error assessment, integrating lateral and longitudinal errors to form a complete quality assessment system. First, the error data is integrated, combining the inter-parameter distance matrix and anomaly association markers obtained from the lateral analysis with the parameter energy distribution and anomaly frequency points obtained from the longitudinal analysis into a unified data structure. An error fusion model is then constructed, using an adaptive weighting method to determine the fusion weights for the lateral and longitudinal errors. Typically, the initial weights are 0.5 each, which are then dynamically adjusted based on the error characteristics. A comprehensive anomaly score is then calculated. Both the lateral and longitudinal anomaly scores are calculated for each parameter, and then combined according to the fusion weights to form a comprehensive anomaly score. Anomaly pattern classification is then performed. Based on the distribution characteristics of the comprehensive anomaly score, detected anomalies are classified into three levels: mild anomalies (scores between 0.1 and 0.3), moderate anomalies (scores between 0.3 and 0.6), and severe anomalies (scores greater than 0.6). Anomaly source tracing is then performed. Based on the correlation analysis of the anomaly parameters, the possible root causes of the anomaly are traced and a causal relationship diagram is constructed. The impact of the anomaly is then assessed, analyzing the extent to which the anomaly parameter affects other related parameters and predicting potential chain reactions. Finally, a hybrid error assessment report is generated, which includes anomaly level, anomaly description, impact range and recommended measures, providing a comprehensive quality assessment basis for the final EWI index calculation.

[0056] The specific implementation of step S09 is to output the final EWI index and generate a comprehensive quality assessment index based on the above analysis results. First, the mixed error weight is determined. The weight ratio of the mixed error in the final index is set according to the characteristics of different production processes, usually set to 0.6-0.7. Then the weight of the merging index is determined as a supplement to the mixed error. The weight is usually set to 0.3-0.4. Then a weighted calculation is performed, and the mixed error evaluation result and the EWI merging index are weighted averaged according to the set weights to obtain a preliminary EWI index value. Subsequently, nonlinear correction is applied to correct the preliminary index through an S-type function (such as a Sigmoid function) so that the final index reflects the change in quality status more sensitively. Then the index threshold is set and the judgment standard is determined according to the product quality requirements. Generally, an EWI index above 0.9 is considered to be a high-quality state, 0.7-0.9 is a normal state, 0.5-0.7 is a warning state, and below 0.5 is an abnormal state. Next, a trend analysis is performed to calculate the change rate and acceleration of the EWI index to predict the development trend of the quality status in the short term. The system then generates indicator visualizations, converting EWI values ​​into intuitive dashboards or trend charts, enabling production managers to quickly identify quality status. Finally, it generates a comprehensive quality report containing EWI values, quality status assessments, key parameter anomalies, trend forecasts, and improvement suggestions, providing a basis for decision-making in production adjustments and quality control.

[0057] A second aspect of the present invention provides a computer-readable storage medium having program instructions stored therein. When the program instructions are executed in a computer, the method for capturing BOP data and generating EWI is executed.

[0058] The third aspect of the present invention provides a system for capturing BOP data and generating EWI, which includes the above-mentioned computer-readable storage medium. The system is any one of a computer, a server, and a single-chip microcomputer. The computer-readable storage medium is set in the system, and the system is provided with a microprocessor that executes the program instructions stored in the computer-readable storage medium.

[0059] The mathematical model or calculation process involved in the present invention is described in detail below.

[0060] The process of constructing a multi-dimensional BOP data load matrix in step S01 involves data standardization, which is specifically expressed as follows:

[0061]

[0062] Where, X norm is the normalized parameter value; X is the original parameter value; X min is the minimum value of the parameter; X max is the maximum value of the parameter.

[0063] This equation uses the maximum and minimum normalization method to uniformly convert parameters of different dimensions to the range of [0, 1] to ensure data comparability. This method is chosen because it preserves the relative relationship between parameters without changing the data distribution. It is suitable for scenarios where the relative differences of the original data need to be maintained. The parameter acquisition method is: X is directly obtained from the production equipment sensor; X min and X max Obtained through statistical analysis of historical data, usually determined based on data from at least 30 production cycles.

[0064] The multi-dimensional BOP data load matrix structure is expressed as:

[0065]

[0066] Where, M is the BOP data load matrix; m ij represents the normalized value of the jth parameter at the i-th time point; T is the total number of time points; and P is the total number of parameters.

[0067] The matrix construction principle is based on the expression of spatiotemporal relationships. The rows represent the time dimension and the columns represent the parameter dimension. This realizes the two-dimensional mapping between time series and parameter types, which facilitates subsequent analysis and processing.

[0068] The calculation of the BOP data contribution value in step S02 involves the entropy weight method, which is specifically expressed as follows:

[0069]

[0070] Where H j is the information entropy of the jth parameter; n is the number of samples; p ij is the proportion of the jth parameter in the i-th sample, calculated as

[0071] The greater the information entropy, the greater the parameter volatility and the more information it contains. The entropy weight method is based on the principle of information theory and measures the importance of parameters by calculating their uncertainty. It is suitable for processing production systems containing multiple uncertainties. The parameters are obtained by calculation, where m ij The data load matrix constructed from step S01.

[0072] The parameter weight coefficient calculation formula is:

[0073]

[0074] Where w j is the weight coefficient of the jth parameter; H j is the information entropy of the jth parameter; P is the total number of parameters.

[0075] The core concept of this formula is that parameters with lower information entropy have stronger discriminatory power and should be assigned higher weights. This weighting method eliminates subjective influences and objectively reflects data characteristics. Parameters are obtained through calculation, without the need for additional experimental measurement.

[0076] In the thermodynamic entropy equation, a variety of influencing factors are considered, and the comprehensive formula is expressed as:

[0077]

[0078] Where S j is the entropy contribution coefficient of the jth parameter; A j is the parameter fluctuation amplitude, the value range is [0, 1]; F j is the fluctuation frequency, ranging from [0, 100] Hz; T j is the system response time, in seconds, with a value range of [0.1, 10]s; j is the system self-organizing ability, with a value range of [-1, 1]; E j is the environmental interference factor, with a value range of [0, 1]; α, β, γ, δ, ∈ are weight coefficients, and the sum is 1.

[0079] This equation is constructed based on the principle of entropy increase and comprehensively considers various factors that affect system stability. The formula uses a weighted sum form and adopts an inverse relationship for the response time, reflecting the characteristic that the faster the response, the greater the impact. Parameter acquisition method: A j Obtained by calculating the standard deviation of the parameter within a 10-minute window; F j The main frequency is obtained by Fourier transform analysis; T j Calculate the input and output delay time through the cross-correlation function; j Obtained by calculating the Lyapunov exponent; E j The environmental impact index is calculated through environmental sensor data.

[0080] The establishment of the EWI basic matrix in step S03 involves the calculation of standard parameter values ​​and fluctuation ranges:

[0081]

[0082] Where, EWI base is the EWI basic matrix; e i1 Indicates the reference value of the i-th parameter; e i2 represents the allowable fluctuation range of the i-th parameter; P is the total number of parameters.

[0083] Benchmark value calculation formula:

[0084]

[0085] Where K is the number of selected high-quality batches; v ik is the average value of the i-th parameter in the k-th high-quality batch.

[0086] Fluctuation range calculation formula:

[0087] e i2 =3·σ i ;

[0088] Where, σ i is the standard deviation of the i-th parameter in the high-quality batch, and the calculation formula is

[0089] This matrix is ​​constructed based on statistical principles, using the mean of high-quality batches as the benchmark and three times the standard deviation as the fluctuation range. This conforms to the 99.7% confidence interval under a normal distribution, ensuring the scientific validity of the judgment criteria. Parameters are obtained through historical data analysis and calculation, and batches with quality inspection scores in the top 10% are selected as high-quality batches.

[0090] The core calculation process of the minimum spanning tree algorithm in step S04 involves the following formula:

[0091] Inter-point connection cost matrix:

[0092]

[0093] Where C is the cost matrix; c ij represents the connection cost from point i to point j; N is the total number of data points.

[0094] Connection cost calculation formula:

[0095]

[0096] Where, d ij is the Euclidean distance between point i and point j; w i and w j are the weights of point i and point j respectively; r ij is the resource consumption factor; λ1, λ2, λ3 are weight coefficients, and λ1+λ2+λ3=1.

[0097] This formula takes into account three factors: spatial distance, point importance, and resource consumption, and adopts a weighted sum form. The point weight adopts an inverse relationship, which reflects the characteristic that the connection between important points has a higher priority. Parameter acquisition method: d ij Obtained by calculating the Euclidean distance between two points in the parameter space; w i and w j The parameter weights calculated in step S02; ijObtained through system resource monitoring, including processor usage, storage space, and communication bandwidth consumption.

[0098] BOP data capture rate calculation formula:

[0099]

[0100] Where R capture is the data capture rate; N actual is the number of valid data points actually captured; N theoretical The theoretical number of data points that should be captured.

[0101] This formula uses a ratio relationship to intuitively reflect the integrity of data collection. The ideal capture rate should be no less than 95%. Parameter acquisition method: N actual Obtained through data acquisition system recording; N theoretical Calculated based on the sampling frequency and time period.

[0102] The Lagrange multiplier method in step S05 solves the optimization problem of the optimal weight as follows:

[0103] Objective function:

[0104]

[0105] Where J(W) is the objective function; W = [w1, w2, ..., w P ] is the weight vector; y i is the actual EWI value of the i-th sample; x ij is the matching degree between the jth parameter in the i-th sample and the basic matrix; n is the number of samples; P is the total number of parameters.

[0106] Constraints:

[0107]

[0108] w min ≤w j ≤w max , j = 1, 2, ..., P;

[0109] Where w min is the lower limit of weight, usually 0.1; w max It is the upper limit of weight, usually 0.5.

[0110] Lagrangian function:

[0111]

[0112] Where λ and μ 1j 、μ 2j is the Lagrange multiplier.

[0113] This optimization problem is based on the least squares principle, and the goal is to minimize the prediction error while satisfying the weight constraints. The Lagrange multiplier method transforms the constrained optimization problem into an unconstrained problem, and obtains the optimal solution by solving the system of equations with zero partial derivatives. Parameter acquisition method: y i Obtained through the EWI index value of historical high-quality batches; x ij It is obtained by calculating the matching degree between the parameter value and the basic matrix. The calculation formula is: where v ij is the actual value of the jth parameter in the i-th sample.

[0114] EWI Consolidation Index calculation formula:

[0115]

[0116] Where, EWI combined is the EWI mergibility index, ranging from [0, 1]; w j is the optimal weight of the jth parameter; x j is the matching degree between the jth parameter and the basic matrix.

[0117] The formula adopts the weighted sum form to achieve a comprehensive evaluation of multi-dimensional parameters. The closer the index value is to 1, the better the production status. The parameters are obtained by calculation, where w j Solved by Lagrange multiplier method; x j Obtained by calculating the matching degree between the parameters and the basic matrix.

[0118] The weighted Euclidean distance calculation formula in step S06 is:

[0119]

[0120] Where, D(V, V ref ) is the weighted Euclidean distance; V = [v1, v2, ..., v P ] is the actual parameter vector; V ref =[v ref，1 , v ref，2 ,...,v ref，P ] is the reference standard vector; ω j is the weight coefficient of the j-th dimension, and

[0121] This formula is based on the concept of Euclidean space distance, and introduces weight coefficients to enhance the influence of important parameters. The calculation form of square root and square root allows the differences in each dimension to be taken into account. Parameter acquisition method: V is obtained through real-time parameter collection during the production process; V ref Obtain the reference value from the EWI basic matrix in step S03; ω jObtained through parameter weight calculation in step S02.

[0122] Distance matrix calculation formula:

[0123]

[0124] Where DM is the distance matrix; d ij Represents the weighted Euclidean distance between parameter i and parameter j, and the calculation formula is

[0125] This matrix represents the relationships between parameters, with diagonal elements set to zero and off-diagonal elements reflecting the distances between pairs of parameters. The purpose of constructing this matrix is ​​to identify unusual correlations between parameters and provide a basis for evaluating mixing errors. Parameters are obtained through calculations based on the differences between actual parameter values ​​and reference standard values.

[0126] Abnormal correlation identifier calculation formula:

[0127]

[0128] Where A ij is the abnormal association identifier between parameter i and parameter j; d ij is the distance between parameters; τ cor is the correlation threshold, usually 0.75.

[0129] This formula uses a binary judgment method to intuitively mark abnormal associations and simplify subsequent analysis and processing. Parameter acquisition method: d ij From the distance matrix calculation result; τ cor Critical values ​​determined through historical data analysis.

[0130] The core calculation formula of Fourier transform in step S07 is:

[0131]

[0132] Where, X k is the complex value at the kth frequency point; x n is the value of the nth point in the time series; N is the sequence length; i is the imaginary unit; k is the frequency index, k = 0, 1, ..., N-1.

[0133] This formula is based on the principle of Fourier transform, which converts the time domain signal into the frequency domain and realizes the analysis of the signal frequency characteristics. The reason for choosing Fourier transform is that it can effectively identify periodic patterns and abnormal changes. Parameter acquisition method: x n Obtained through parameter time series data during the production process; N is the number of sampling points, usually 2 mpoints to improve computing efficiency, such as 4096 points.

[0134] Window function application formula:

[0135] x′ n =x n w n ;

[0136] Where x′ n is the sequence value after applying the window function; x n is the original sequence value; w n is the window function value.

[0137] For the Hanning window, w n The calculation formula is:

[0138]

[0139] The application of window function reduces the spectrum leakage effect and improves the accuracy of spectrum analysis. Parameter acquisition method: x n From time series data; w n Calculated using window functions.

[0140] Energy distribution calculation formula:

[0141] E k =|X k | 2 ;

[0142] Where, E k is the energy value of the kth frequency point; X k is the kth value of the Fourier transform result; |X k | represents the plural X k Model.

[0143] This formula calculates the energy distribution at each frequency point, reflecting the signal strength at different frequencies. The parameters are obtained by calculating the square of the modulus of the Fourier transform result.

[0144] Abnormal frequency point identification formula:

[0145]

[0146] Where, F abnormal is the set of abnormal frequency points; E k is the actual energy distribution; E k，ref is the reference energy distribution; τ f is the frequency anomaly threshold, usually set to 0.3.

[0147] This formula determines the abnormal frequency by relative deviation and marks the frequency point that is significantly different from the normal mode. Parameter acquisition method: Ek Calculated by Fourier transform; E k，ref Obtained by analyzing the spectra of historical high-quality batches; τ f Determined through historical data analysis.

[0148] The key calculation formula for the mixed error evaluation in step S08 is:

[0149] Horizontal anomaly score calculation:

[0150]

[0151] Where S horizontal is the horizontal anomaly score; A ij is the abnormal association identifier between parameter i and parameter j; P is the total number of parameters.

[0152] This formula calculates the density of abnormal correlation, reflecting the degree of abnormal relationship between parameters. Parameter acquisition method: A ij The calculation result of the abnormal association identifier in step S06.

[0153] Longitudinal anomaly score calculation:

[0154]

[0155] Where S vertical is the vertical anomaly score; |F abnormal | is the number of abnormal frequency points; N / 2 is the total number of valid frequency points (taking into account conjugate symmetry).

[0156] This formula calculates the proportion of abnormal frequency, reflecting the abnormal degree of parameter time change. Parameter acquisition method: |F abnormal |The abnormal frequency point identification result from step S07.

[0157] Comprehensive anomaly score calculation:

[0158] S combined =α·S horizontal +(1-α)·S vertical ;

[0159] Where S combined is the comprehensive anomaly score; α is the lateral anomaly weight, the initial value is usually 0.5, and it can be dynamically adjusted later.

[0160] The formula adopts the weighted sum form, which takes into account the horizontal and vertical abnormalities in a balanced manner and realizes a comprehensive quality assessment. horizontal and S vertical They are derived from the calculation results of the horizontal anomaly score and the vertical anomaly score respectively; α is initially set to 0.5 and is subsequently dynamically adjusted according to the anomaly characteristics.

[0161] The final EWI indicator calculation formula in step S09 is:

[0162] Preliminary EWI indicator calculation:

[0163] EWI initial =β·(1-S combined )+(1-β)·EWI combined ;

[0164] Where, EWI initial is the preliminary EWI indicator; S combined is the comprehensive anomaly score; EWI combined is the EWI mergibility index; β is the mixing error weight, which is usually between 0.6 and 0.7.

[0165] This formula comprehensively considers the abnormality assessment and the mergeability index, and realizes the quality assessment from multiple angles. Parameter acquisition method: S combined The result of calculating the comprehensive anomaly score in step S08; EWI combined The EWI from step S05 can be combined with the index calculation result.

[0166] Nonlinear correction formula:

[0167]

[0168] Where, EWI final is the final EWI index; γ is the sensitivity coefficient, usually taken as 10; θ is the midpoint parameter, usually taken as 0.5.

[0169] This formula uses the Sigmoid function for nonlinear correction, making the indicator more sensitive to changes in the middle area and improving the indicator's resolution. Parameter acquisition method: EWI initial Derived from preliminary EWI indicator calculation results; γ and θ are determined by historical data verification to obtain the best indicator response characteristics.

[0170] Optionally, the specific implementation formula of the minimum spanning tree algorithm is as follows:

[0171] MST={(i, j)|(i, j)∈E, satisfies the minimum cost tree condition};

[0172] Where MST is the minimum spanning tree; E is the edge set, which contains all possible data point connections.

[0173] Optionally, a mathematical expression of the steps of Kruskal's algorithm:

[0174]

[0175] For the edges in E, calculate the cost cij Sort from smallest to largest;

[0176] Examine each edge (i, j) in turn. If adding (i, j) does not form a loop, then T = T∪{(i, j)};

[0177] Until |T|=N-1;

[0178] Where T is the currently constructed tree; |T| is the number of edges in the tree; and N is the total number of data points.

[0179] The algorithm is based on a greedy strategy and gradually selects the edge with the lowest cost to construct a minimum spanning tree, ensuring optimal data collection under resource constraints. Parameter acquisition method: The cost of the edge c ij The connection cost calculation result from step S04.

[0180] Optional, calculation formula for parameter correlation matrix:

[0181]

[0182] Where R is the parameter correlation matrix; r ij It represents the Pearson correlation coefficient between parameter i and parameter j, and the calculation formula is:

[0183]

[0184] Where, v ki represents the i-th parameter value of the k-th sample; represents the mean value of the i-th parameter; n is the number of samples.

[0185] This formula, based on statistical principles, calculates the degree of linear correlation between parameters. Its range is [-1, 1]. Larger absolute values ​​indicate stronger correlation. Parameter Acquisition Method: Parameter values ​​are obtained from historical data during the production process; the average value is calculated by taking the arithmetic mean of this historical data.

[0186] Specifically, the principle of the present invention is: based on information entropy theory, spectrum analysis, graph optimization and multi-objective constrained optimization principles, the present invention constructs a systematic BOP data capture and EWI generation framework. Its core principle is to achieve precise control of the BOP data capture and EWI generation process through the synergy of multiple algorithms.

[0187] First, the contribution of each BOP parameter is calculated based on the thermodynamic entropy equation. This process follows the theory of information entropy, which views the production system as an energy-information conversion system. The parameter fluctuation characteristics, system response properties, and environmental impacts are quantified as entropy increase indicators, scientifically measuring the impact of each parameter on system stability. Compared with traditional empirical weighting, this information entropy-based weighting method objectively reflects the actual importance of parameters, eliminates the accuracy bias caused by subjective judgment, and provides a scientific and accurate data foundation for subsequent analysis.

[0188] Secondly, the BOP data capture strategy was optimized using a minimum spanning tree algorithm. Based on the construction principle of minimum spanning trees in graph theory, this algorithm comprehensively considers the connection cost between data points and the importance of the data to determine the optimal data collection path under limited resource conditions. This approach ensures the capture of the most representative set of BOP data points, improving the quality and representativeness of the raw data and laying a solid foundation for generating accurate EWI indicators.

[0189] Furthermore, a key innovation of this invention lies in the simultaneous introduction of lateral and longitudinal error analysis mechanisms, establishing a two-dimensional, precise analysis framework. Lateral errors precisely calculate the degree of deviation between different BOP parameters at the same time point using the Euclidean distance function, capturing the complex nonlinear relationships between these parameters. Longitudinal errors, on the other hand, precisely detect abnormal variations in a single BOP parameter over a time series using a Fourier transform, effectively identifying the parameter's temporal evolution. This dual-dimensional analysis approach overcomes the limitations of traditional single-dimensional analysis, enabling comprehensive and accurate assessment of BOP data quality and generation of precise EWI indicators.

[0190] Finally, the EWI fusion weights are accurately solved under multi-constraint conditions through the Lagrange multiplier method. This method is based on the multi-objective constrained optimization theory and can find the optimal weight configuration while satisfying all constraints. The generated EWI index can not only accurately reflect the actual contribution of each BOP parameter, but also meet the system's stability requirements.

[0191] The reason why this multi-algorithm fusion technical solution can effectively solve the core problem of inaccurate EWI generation from BOP data capture is that it fundamentally changes the single, static, and linear analysis model of traditional technology, introduces a more scientific weight calculation method, a more optimized data capture strategy, a more comprehensive error analysis system, and a more accurate data fusion algorithm, constructs a rigorous and efficient EWI generation framework, and provides an accurate and reliable BOP data capture and EWI generation method.

[0192] A specific embodiment 1 of the present invention is provided below. The specific implementation of each step in this embodiment 1 is described in detail as follows.

[0193] The specific implementation of step S01 involves constructing a multidimensional BOP data load matrix. Key parameter data from the production process is first collected and arranged into a time series format according to timestamps. Parameter types are then categorized and organized, with similar parameters grouped into the same dimension to form a preliminary parameter classification structure. Data format standardization is then performed, converting parameter data of different dimensions into a dimensionless form. Maximum and minimum normalization is employed to ensure that all data values ​​fall within the [0, 1] interval. The calculation formula is:

[0194]

[0195] Where, X norm is the normalized parameter value; X is the original parameter value; X min is the minimum value of the parameter; X max is the maximum value of the parameter. X is directly obtained from the production equipment sensor; X min and X max It is obtained through statistical analysis of historical data, usually based on data from at least 30 production cycles. A matrix index structure is then established, with time as the first dimension index and parameter type as the second dimension index, thus forming a multi-dimensional data structure, which can be expressed as:

[0196]

[0197] Where, M is the BOP data load matrix; m ij represents the standardized value of the jth parameter at the i-th time point; T is the total number of time points; and P is the total number of parameters. Finally, a data integrity check is performed to address missing values. Linear interpolation can be used to fill missing data for short time intervals, or forward filling can be used to handle consecutive missing values. This series of processes ultimately creates a standardized multidimensional BOP data loading matrix, providing a structured data foundation for subsequent analysis.

[0198] The specific implementation of step S02 is to calculate the contribution value of the BOP data and analyze the impact of each dimension of data on the system based on the thermodynamic entropy equation. First, the parameter fluctuation amplitude is calculated by calculating the standard deviation of each parameter within a fixed time window (e.g., 10 minutes) as the fluctuation amplitude indicator. Then, the fluctuation frequency is measured. The main frequency components of the parameter per unit time are analyzed through Fourier transform to obtain the frequency of significant changes. Next, the system response time is evaluated. The time delay from parameter change to system output response is calculated using the cross-correlation function. Generally, a faster response time (e.g., less than 5 seconds) indicates a greater parameter impact. The system's self-organizing ability is also measured. The Lyapunov exponent is used to calculate the system's ability to recover a stable state after a disturbance. The exponent value typically ranges from [-1 to 1], with the closer to -1, the stronger the system's self-organizing ability. The environmental interference factor is then quantified based on the influence of environmental parameters such as temperature (standard operating temperature deviation does not exceed ±5°C), humidity (relative humidity is controlled within the range of 45% to 65%), and vibration (amplitude does not exceed 0.1mm). The entropy value of each dimension parameter is calculated using the entropy weight method:

[0199]

[0200] Where H j is the information entropy of the jth parameter; n is the number of samples; p ij is the proportion of the jth parameter in the i-th sample, calculated as Then calculate the parameter weight coefficient:

[0201]

[0202] Where w j is the weight coefficient of the jth parameter; H j is the information entropy of the jth parameter; P is the total number of parameters. Finally, considering various influencing factors, the comprehensive formula is expressed as:

[0203]

[0204] Where S j is the entropy contribution coefficient of the jth parameter; A j is the parameter fluctuation amplitude, the value range is [0, 1]; F j is the fluctuation frequency, ranging from [0, 100] Hz; T j is the system response time, in seconds, with a value range of [0.1, 10]s; j is the system self-organizing ability, with a value range of [-1, 1]; E jis the environmental interference factor, with a value range of [0, 1]; α, β, γ, δ, ∈ are weight coefficients, and their sum is 1. Through normalization processing, the sum of the contribution values ​​of BOP data in all dimensions is ensured to be 1, thus forming a weight distribution scheme, which provides a basis for the subsequent EWI indicator calculation.

[0205] The specific implementation of step S03 is to establish an EWI foundation matrix, forming a benchmark framework for quality assessment based on historical production data. First, select historical high-quality production batch data and select the best-performing production cycle data based on product quality inspection results (usually batches with quality inspection scores in the top 10%). Then, calculate the baseline value of each parameter, and use the parameter mean corresponding to these high-quality batches as the standard reference point:

[0206]

[0207] Where K is the number of selected high-quality batches; v ik is the average value of the i-th parameter in the k-th high-quality batch. Then determine the parameter fluctuation threshold. By calculating the standard deviation of the high-quality batch parameters, the allowable fluctuation range is set to ±3 times the standard deviation:

[0208] e i2 =3·σ i ;

[0209] Where σ i is the standard deviation of the i-th parameter in the high-quality batch, and the calculation formula is Then, the correlation model between parameters is established, and the Pearson correlation coefficient is used to analyze the relationship between different parameters. The correlation coefficient threshold is usually set to 0.7. Parameters above this value are considered to be strongly correlated. Then, the EWI basic matrix is ​​constructed:

[0210]

[0211] Where, EWI base is the EWI basic matrix; e i1 Indicates the reference value of the i-th parameter; e i2 = represents the allowable fluctuation range of the i-th parameter; P is the total number of parameters. Finally, verification and adjustment are performed. Another portion of historical data is used to verify the applicability of the threshold value. If necessary, fine-tuning is performed to ensure that the EWI basic matrix can accurately reflect the quality status of the production process.

[0212] The specific implementation of step S04 involves calculating the BOP data capture rate and optimizing the data collection strategy using a minimum spanning tree algorithm. First, the data point distribution density is analyzed. Kernel density estimation is used to calculate the data distribution density function in the production parameter space and identify high-density areas. Next, data point weights are determined. Each data point is assigned a weight based on its parameter contribution. Weights typically range from [0.1 to 1.0], reflecting the importance of the data point. Next, the cost of connecting points is calculated, taking into account the system resource consumption required to collect different data points, including computing resources, storage resources, and communication bandwidth, to form a cost matrix:

[0213]

[0214] Where C is the cost matrix; c ij Represents the connection cost from point i to point j; N is the total number of data points. Connection cost calculation formula:

[0215]

[0216] Where, d ij is the Euclidean distance between point i and point j; w i and w j are the weights of point i and point j respectively; r ij is the resource consumption factor; λ1, λ2, and λ3 are weight coefficients, and λ1 + λ2 + λ3 = 1. The network topology is then analyzed to establish a connection diagram between data acquisition nodes and determine feasible data transmission paths. Furthermore, considering system resource limitations, resource constraints for data acquisition are set based on the actual hardware configuration, such as processor utilization not exceeding 80%, storage space utilization not exceeding 75%, and communication bandwidth utilization not exceeding 90%. The minimum spanning tree algorithm is then applied, based on the Kruskal algorithm, to construct a minimum-cost tree connecting the data points:

[0217] MSE = {(i, j)|(i, j)∈E, satisfies the minimum cost tree condition};

[0218] Mathematical expression of the implementation steps of Kruskal's algorithm:

[0219]

[0220] For the edges in E, calculate the cost c ij Sort from smallest to largest;

[0221] Examine each edge (i, j) in turn. If adding (i, j) does not form a loop, then T = T∪{(i, j)};

[0222] Until |T|=N-1;

[0223] Where T is the currently constructed tree; |T| is the number of edges in the tree; and N is the total number of data points. Finally, the data capture rate indicator is calculated:

[0224]

[0225] Where R capture is the data capture rate; N actual is the number of valid data points actually captured; N theoretical The theoretical number of data points that should be crawled. The ideal crawling rate should usually be maintained above 95%.

[0226] The specific implementation of step S05 is to generate the EWI combinable index and use the Lagrange multiplier method to achieve a comprehensive evaluation of multidimensional parameters. First, set the upper and lower weight constraints and determine the weight value range based on the importance of each parameter. Generally, the lower limit of the core parameter weight is not less than 0.1 and the upper limit does not exceed 0.5 to ensure a reasonable weight distribution. Then, construct the correlation matrix between the parameters:

[0227]

[0228] Where R is the parameter correlation matrix; r ij It represents the Pearson correlation coefficient between parameter i and parameter j, and the calculation formula is:

[0229]

[0230] Where, v ki represents the i-th parameter value of the k-th sample; represents the average value of the i-th parameter; n is the number of samples. Next, evaluate historical accuracy by backtesting each parameter to analyze the accuracy of its predictions for the historical EWI indicator. Accuracy is typically scored using the root mean square error (RMS), with an ideal value of less than 0.05. Then, select the objective function type and set the optimization objective function based on the principle of minimizing the overall forecast error:

[0231]

[0232] Where J(W) is the objective function; W = [w1, w2, ..., w P ] is the weight vector; y i is the actual EWI value of the i-th sample; x ij is the matching degree between the jth parameter in the i-th sample and the basic matrix; n is the number of samples; P is the total number of parameters. Constraints:

[0233]

[0234] w min ≤w j ≤w max, j = 1, 2, ..., P;

[0235] Where w min is the lower limit of weight, usually 0.1; w max The upper limit of the weight is usually set to 0.5. Then the convergence threshold is determined and the termination condition of the iterative calculation is set. For example, when the weight change between two adjacent iterations is less than 0.001, it is considered converged. Next, the Lagrange multiplier method is applied to solve the optimal weight distribution solution under all constraints:

[0236]

[0237] Where λ and μ 1j 、μ 2j is the Lagrange multiplier. Finally, calculate the EWI mergible index:

[0238]

[0239] Where, EWI combined is the EWI mergibility index, ranging from [0, 1]; w j is the optimal weight of the jth parameter; x j is the matching degree between the jth parameter and the basic matrix, and the calculation formula is where v j The index value is the actual value of the parameter. The index value range is in the interval [0, 1], where 0.8 or above indicates good production status, 0.6 to 0.8 indicates acceptable production status, and below 0.6 indicates abnormal production status.

[0240] The specific implementation method of step S06 is to calculate the lateral error and use the Euclidean distance function to analyze the relationship between different parameters at the same time point. First, organize the parameter vector group, organize multiple BOP parameter values ​​at the same time point into vector form, and construct the parameter space. Then determine the dimension weight coefficient, set the weight coefficient based on the importance of each parameter, and reflect the importance of different dimensions in the distance calculation. Then select the reference standard vector, and use the parameter vector under high-quality production status as the reference point. Usually, the data of the best historical production batch is selected as the standard. Then set the allowable deviation range, and determine the allowable parameter deviation degree according to the production process requirements, which is usually set within the range of ±5% of the standard vector value. Then determine the correlation threshold, and set the critical value for judging parameter correlation anomalies. Generally, 0.75 is taken. Values ​​below this value are considered as correlation anomalies. Next, calculate the weighted Euclidean distance:

[0241]

[0242] Where, D(V, V ref ) is the weighted Euclidean distance; V = [v1, v2, ..., v P] is the actual parameter vector; V ref =[v ref，1 , v ref，2 ,...,v ref，P ] is the reference standard vector; ω j is the weight coefficient of the j-th dimension, and Then construct the distance matrix:

[0243]

[0244] Where DM is the distance matrix; d ij Represents the weighted Euclidean distance between parameter i and parameter j, and the calculation formula is Finally, the abnormal correlation identifier is generated:

[0245]

[0246] Where A ij is the abnormal association identifier between parameter i and parameter j; d ij is the distance between parameters; τ cor is the correlation threshold, usually set to 0.75. Through these calculations, a complete distance matrix and abnormal correlation identification are formed, providing horizontal analysis results for mixed error assessment.

[0247] The specific implementation of step S07 is to calculate the longitudinal error and use the Fourier transform equation to track the time variation characteristics of the parameters. First, prepare the time series parameter values ​​and arrange the single BOP parameters in chronological order to form a discrete time series. Then determine the sampling interval and set the sampling time interval according to the data acquisition frequency, usually in milliseconds (such as 100ms) or seconds (such as 1s). Then select the frequency resolution and set the fineness of the Fourier transform. Usually, a resolution of not less than 0.01Hz is selected to ensure that the key frequency components can be captured. Then select the window function type and apply the Hanning window to weight the original signal to reduce the spectrum leakage effect:

[0248] x′ n =x n w n ;

[0249] Where x′ n is the sequence value after applying the window function; x n is the original sequence value; w n is the window function value. For the Hanning window, w n The calculation formula is:

[0250]

[0251] Then set the spectrum smoothing coefficient, usually 0.05 to 0.2, to reduce the impact of random noise on spectrum analysis. Next, perform the fast Fourier transform:

[0252]

[0253] Where, X k is the complex value at the kth frequency point; x n is the value of the nth point in the time series; N is the sequence length; i is the imaginary unit; k is the frequency index, k = 0, 1, ..., N-1. Then analyze the energy distribution characteristics:

[0254] E k =|X k | 2 ;

[0255] Where, E k is the energy value of the kth frequency point; X k is the kth value of the Fourier transform result; |X k | represents the plural X k Finally, identify the abnormal frequency points:

[0256]

[0257] Where, F abnormal is the set of abnormal frequency points; E k is the actual energy distribution; E k，ref is the reference energy distribution; τ f is the frequency anomaly threshold, usually set to 0.3. Through this series of processing, the energy distribution of the parameters at different frequencies and the abnormal frequency points are obtained, providing a longitudinal analysis basis for the mixed error assessment.

[0258] The specific implementation of step S08 is to generate a hybrid error assessment, integrating the lateral error and the longitudinal error to form a complete quality assessment system. First, the error data is integrated, and the inter-parameter distance matrix and anomaly association markers obtained from the lateral analysis are combined with the parameter energy distribution and anomaly frequency points obtained from the longitudinal analysis into a unified data structure. Then, an error fusion model is constructed, and an adaptive weighting method is used to determine the fusion weights of the lateral error and the longitudinal error. The initial weights are usually 0.5 each, and then dynamically adjusted according to the error characteristics. Then, the lateral anomaly score is calculated:

[0259]

[0260] Where S horizontal is the horizontal anomaly score; A ij is the abnormal association identifier between parameter i and parameter j; P is the total number of parameters. Then calculate the vertical anomaly score:

[0261]

[0262] Where S vertical is the vertical anomaly score; |F abnormal | is the number of abnormal frequency points; N / 2 is the total number of valid frequency points (taking into account conjugate symmetry). Then calculate the comprehensive anomaly score:

[0263] S combined =α·S horizontal +(1-α)·S vertical ;

[0264] Where S combined is the comprehensive anomaly score; α is the horizontal anomaly weight, the initial value is usually 0.5, and it can be adjusted dynamically later. Next, the anomaly pattern classification is performed. According to the distribution characteristics of the comprehensive anomaly score, the detected anomalies are divided into three levels: mild anomalies (scores between 0.1 and 0.3), moderate anomalies (scores between 0.3 and 0.6), and severe anomalies (scores greater than 0.6). Then, the source of the anomaly is traced. Based on the correlation analysis of the anomaly parameters, the possible root causes of the anomaly are traced and a causal relationship diagram is established. Finally, the scope of the anomaly impact is evaluated, the degree of influence of the anomaly parameters on other related parameters is analyzed, potential chain reactions are predicted, and a mixed error assessment report is generated to provide a comprehensive quality assessment basis for the final EWI indicator calculation.

[0265] The specific implementation of step S09 is to output the final EWI index and generate a comprehensive quality assessment index based on the above analysis results. First, determine the mixed error weight, and set the weight ratio of the mixed error in the final index according to the characteristics of different production processes, usually set to 0.6-0.7. Then determine the mergeability index weight as a supplement to the mixed error, usually set to 0.3-0.4. Then perform a weighted calculation, and perform a weighted average of the mixed error evaluation result and the EWI mergeability index according to the set weight to obtain a preliminary EWI index value:

[0266] EWI initial =β·(1-S combined )+(1-β)·EWI combined ;

[0267] Where, EWI initial is the preliminary EWI indicator; S combined is the comprehensive anomaly score; EWI combined is the EWI combinable index; β is the mixed error weight, usually 0.6 to 0.7. Then, nonlinear correction is applied to correct the preliminary index through the S-type function, so that the final index can more sensitively reflect the change of quality status:

[0268]

[0269] Where, EWI final is the final EWI indicator; γ is the sensitivity coefficient, usually 10; θ is the midpoint parameter, usually 0.5. Then set the indicator threshold and determine the judgment standard based on the product quality requirements. Usually, an EWI indicator above 0.9 is considered to be in a high-quality state, 0.7-0.9 is in a normal state, 0.5-0.7 is in a warning state, and below 0.5 is in an abnormal state. Next, perform trend analysis to calculate the rate of change and acceleration of the EWI indicator and predict the development trend of the quality status in the short term. Then generate indicator visualization results and convert the EWI indicator values ​​into intuitive dashboards or trend charts to facilitate production managers to quickly identify the quality status. Finally, output a comprehensive quality report containing EWI indicator values, quality status judgments, abnormal conditions of key parameters, trend forecasts, and improvement suggestions, providing a decision-making basis for production adjustments and quality control.

[0270] Through the implementation of the above steps, this embodiment completes the entire process from BOP data collection to EWI indicator generation, achieving comprehensive monitoring and assessment of production quality. This method integrates multiple mathematical and computer algorithms, including maximum-minimum normalization, entropy weighting, correlation analysis, minimum spanning tree algorithm, Lagrange multiplier method, Euclidean distance calculation, and Fourier transform, to establish a scientific and complete production quality assessment system. Each step is closely connected and interdependent, forming a closed-loop quality monitoring process. Through multi-dimensional analysis and assessment of BOP data, quality anomalies in the production process can be promptly identified, enabling early warning and intervention, thereby improving production efficiency and product quality.

[0271] To better understand and implement the present invention, Example 2 of a specific application scenario is provided below: On an advanced semiconductor wafer production line, researchers applied the present method to implement an automated quality monitoring system to address the lack of efficient and forward-looking quality indicators in the chip manufacturing process. This system collects BOP (By-Step-of-Process) data from various production equipment and generates real-time EWI (Early Warning Indicator), enabling accurate prediction and control of production quality.

[0272] The researchers first identified 12 key BOP parameters, as shown in Table 1:

[0273] Table 1 List of key BOP parameters for wafer manufacturing

[0274]

[0275]

[0276] Following step S01, researchers constructed a multidimensional BOP data load matrix. The system collected data at 500 time points (one data point every 5 seconds), forming a 500×12 matrix. Using the maximum-minimum normalization method, all parameter data were uniformly converted to the [0, 1] range.

[0277] In step S02, researchers calculated the weight coefficients of each parameter using the entropy weight method, as shown in Table 2:

[0278] Table 2 Entropy values ​​and weight coefficients of BOP parameters

[0279] Parameter number <![CDATA[Information entropy (H j )]]> <![CDATA[Weight coefficient (w j )]]> P1 0.413 0.135 P2 0.387 0.142 P3 0.456 0.126 P4 0.329 0.155 P5 0.621 0.087 P6 0.294 0.163 P7 0.537 0.107 P8 0.689 0.072 P9 0.368 0.146 P10 0.792 0.048 P11 0.846 0.035 P12 0.769 0.054

[0280] When the system calculates the entropy contribution coefficient based on the thermodynamic entropy equation, the weight parameters are set: α = 0.3, β = 0.25, γ = 0.2, δ = 0.15, ∈ = 0.1, taking into account the parameter fluctuation amplitude, fluctuation frequency, system response time, self-organizing ability and environmental interference factors. Figure 2 and Figure 3 The time series changes of key BOP parameters in the wafer manufacturing process and the corresponding EWI indicators are compared. Figure 2 The graph shows the change curve of the standardized values ​​of 5 key parameters (P1 plasma power, P2 chamber pressure, P4 RF bias, P6 etching uniformity, P9 endpoint signal) over time. These parameters are simulated according to the parameter description of Table 1 in the embodiment, and all parameter values ​​are subjected to maximum and minimum standardization processing and constrained within the [0,1] interval. The abnormal area between the 300-350 time points is specially marked in the chart. In this interval, the P4 (RF bias) parameter shows obvious abnormal fluctuations, which corresponds to the spectrum analysis anomaly found in step S07 in the embodiment. Figure 3 The EWI indicator curves calculated based on the above parameters are shown, including the initial EWI indicator (dashed line) and the EWI indicator after nonlinear Sigmoid correction (solid line), as well as the lateral error indicator curve. The figure also shows three horizontal dashed lines of different colors, representing the high-quality threshold (0.9), warning threshold (0.8), and danger threshold (0.7). It can be clearly observed that within the abnormal area, the EWI indicator value decreases significantly, while the lateral error indicator increases significantly, fully demonstrating the anomaly detection and early warning mechanism described in steps S06 and S09 of the embodiment.

[0281] In step S03, the researchers selected 50 production batches with quality inspection scores in the top 10% as high-quality batches and calculated the EWI basic matrix. The baseline values ​​and allowable fluctuation ranges of some parameters are shown in Table 3:

[0282] Table 3 EWI basic matrix (partial parameters)

[0283] Parameter number <![CDATA[Reference value (e i1 )]]> <![CDATA[Allowable fluctuation range (e i2 )]]> P1 1050.32 31.59 P2 5.47 0.16 P3 187.63 5.63 P4 275.81 8.27 P6 97.83 2.93 P9 653.24 19.60

[0284] In step S04, the system optimized the data collection strategy using a minimum spanning tree algorithm. First, a 500×500 inter-point connection cost matrix was established, with λ1 = 0.5, λ2 = 0.3, and λ3 = 0.2. Then, a minimum spanning tree was constructed using the Kruskal algorithm to determine the optimal data collection path. In practice, the system reduced the sampling frequency from once per second to once every five seconds. However, by optimizing the sampling point distribution, the data capture rate reached 97.6%, exceeding the set threshold of 95%.

[0285] In step S05, the system calculates the EWI index using the Lagrange multiplier method. The upper and lower weight limits are set to [0.05, 0.25], and an iterative solution is used to find the optimal weight distribution scheme. During a certain batch of production, the calculated EWI index was 0.862, indicating that production status was within the normal range.

[0286] Step S06 calculates the lateral error. The system selects the average value of the high-quality batch as the reference standard vector and uses the weighted Euclidean distance to calculate the degree of deviation between the actual parameter vector and the reference vector. Set the correlation threshold τ cor =0.75, the system identifies an abnormal correlation between P3 (gas flow) and P9 (endpoint signal), with a distance value of 0.83, which exceeds the threshold.

[0287] Step S07 calculates the longitudinal error. The system performs Fourier transform analysis on each parameter using 4096-point FFT and Hanning window function. and a smoothing coefficient of 0.15. In the spectrum analysis of the P4 (RF bias) parameter, three abnormal frequency points were found, and the energy deviation exceeded the set threshold τ f =0.3.

[0288] Figure 4 This is a spectrum analysis diagram that compares the differences between the normal spectrum and the abnormal spectrum, and clearly marks three abnormal frequency points. The abnormal frequency points appear as additional energy peaks outside the normal spectrum, and the energy deviation exceeds the set threshold. Figure 5 The graph shows the time-varying trends of the horizontal anomaly score, vertical anomaly score, and comprehensive anomaly score. The horizontal anomaly score is based on the weighted Euclidean distance calculated in step S06, the vertical anomaly score is based on the spectral analysis results in step S07, and the comprehensive anomaly score is calculated by combining the two using the method in step S08 with a weight of α = 0.5. The figure shows two horizontal dashed lines, representing the minor anomaly threshold (0.1) and the severe anomaly threshold (0.2). It can be seen that within the abnormal region, all three anomaly scores increase significantly, with the vertical anomaly score showing the most significant increase.

[0289] Step S08 generates a hybrid error evaluation, a lateral anomaly score S horizontal =0.061, vertical abnormality score S vertical =0.113, and the comprehensive anomaly score S is calculated using a weight of α = 0.5. combined =0.087, which is slightly abnormal (below 0.1).

[0290] In step S09, the system sets the mixed error weight β = 0.65, and calculates the preliminary EWI index value EWI initial = 0.857. A Sigmoid function was then used for nonlinear correction (γ = 10, θ = 0.5), resulting in an EWI index value of 0.903, indicating excellent production quality. The system automatically generated a quality report, noting that despite minor anomalies, overall production was good and recommending attention to the abnormal correlation between gas flow and endpoint signals.

[0291] Traditional semiconductor manufacturing quality control relies primarily on post-process testing and statistical process control (SPC), which are unable to predict quality trends in real time. Problems are often discovered only after batches are completed, leading to material waste and inefficiency. The present invention, however, generates EWI (Exerted Width Index) (EWI) from BOP data capture, enabling real-time prediction and proactive control of production quality. This method offers significant advantages: First, adaptively determining parameter weights using the entropy weight method and the thermodynamic entropy equation overcomes the limitations of subjective weighting in traditional methods, resulting in a more objective and scientific weight allocation. Second, optimizing the data collection strategy using a minimum spanning tree algorithm reduces system resource consumption while ensuring data integrity, increasing data capture efficiency by 8.3%. Third, combining a hybrid evaluation model for lateral and longitudinal errors enables multi-dimensional identification of production anomalies, increasing the anomaly detection rate by 23.7% and reducing the false alarm rate by 17.2%. Finally, the EWI, through nonlinear correction, more sensitively reflects quality trends, shortening warning times by an average of 42 minutes, providing a sufficient window for production adjustments. Practice has proven that the present method can effectively improve semiconductor manufacturing quality control, reduce defective product rates, and lower production costs.

[0292] It should be noted that the variables involved in the present invention are explained in detail as shown in Tables 4 and 5 below.

[0293] Table 4 Variable Explanation Table (Part 1)

[0294]

[0295]

[0296]

[0297] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

Claims

1. A method for capturing BOP data and generating EWI, characterized in that: include: Construct a multi-dimensional BOP data load matrix; calculate the BOP data contribution value, and use the thermodynamic entropy equation to analyze the impact of each BOP data dimension; Establish the EWI basic matrix; implement BOP data capture rate calculation, and use the minimum spanning tree algorithm to obtain the optimized data capture path and BOP data integrity score; Generate the EWI mergibility index, and use the Lagrange multiplier method to weight the matching degree of each dimension of BOP data and the EWI basic matrix; calculate the lateral error, and use the Euclidean distance function to analyze the degree of deviation between different BOP parameters at the same time point; To calculate the longitudinal error, the Fourier transform equation is used to track the trend of a single BOP parameter in the time series; Generate a mixed error assessment; based on the mixed error assessment results and the EWI combinable index, generate a comprehensive index reflecting the production quality status as the final EWI index and output it.

2. The method for generating EWI by capturing BOP data according to claim 1, characterized in that: The BOP data load matrix is ​​a multidimensional data structure that constructs key parameters in the production process according to time and type, and is used to represent the load status of each parameter at different times and the mutual influence relationship.

3. The method for generating EWI by capturing BOP data according to claim 2, characterized in that: The BOP data contribution value is the influence weight of each BOP data on the formation process of the final EWI index, which is calculated by the thermodynamic entropy equation and reflects the influence of different parameters on production quality.

4. The method for generating EWI by capturing BOP data according to claim 3, characterized in that: The EWI basic matrix is ​​a reference standard for production process quality control, which contains the standard values ​​and fluctuation thresholds of each parameter and is used as a benchmark for data analysis.

5. The method for generating EWI by capturing BOP data according to claim 4, characterized in that: The BOP data capture rate is the percentage of valid BOP data successfully acquired during the production monitoring process to the total BOP data that should be acquired theoretically, reflecting the reliability of the data acquisition system.

6. The method for generating EWI by capturing BOP data according to claim 5, characterized in that: The EWI combinable index is a process of integrating multi-dimensional parameter BOP data into a single indicator through the Lagrange multiplier method, which facilitates the intuitive evaluation of the overall production status.

7. The method for generating EWI by capturing BOP data according to claim 6, characterized in that: The horizontal error is the relative deviation between different BOP parameters at the same time point, which is used to identify abnormal correlations between BOP parameters; the vertical error is the variation deviation of a single BOP parameter in the time series, which is used to monitor whether the BOP parameter variation trend meets expectations; The mixed error is a comprehensive indicator obtained by fusing the lateral error and the longitudinal error according to specific weights, which comprehensively reflects the quality status of the production process.

8. The method for generating EWI by capturing BOP data according to claim 7, characterized in that: The hybrid error assessment is to conduct a comprehensive analysis of the lateral error and the longitudinal error. Specifically, a comprehensive quality evaluation system is constructed based on the distance matrix between parameters, abnormal association identification, parameter energy distribution and abnormal frequency points, and the hybrid error assessment result is obtained by analysis.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores program instructions, and when the program instructions are executed in a computer, they are used to execute the method for capturing BOP data and generating EWI according to any one of claims 1 to 8.

10. A system for capturing BOP data and generating EWI, characterized by: The computer-readable storage medium according to claim 9 is included, the system is any one of a computer, a server, and a single-chip microcomputer, the computer-readable storage medium is arranged in the system, and the system is provided with a microprocessor for executing program instructions stored in the computer-readable storage medium.

Images