Method, system and equipment for analyzing plate shape quality defects of medium plate and medium

Abstract

The invention discloses a plate shape quality defect analysis method, system and device for a medium plate and a medium, and relates to the technical field of computers, and the method comprises the steps: obtaining a plurality of technological parameters and corresponding plate shape quality data in the production process of the medium plate; analyzing each process parameter and the corresponding plate shape quality data based on a correlation analysis algorithm, and determining an influence degree value corresponding to the process parameter; under the condition that the influence degree value is greater than a preset correlation threshold value, constructing a key parameter data set based on the process parameters corresponding to the influence degree value; performing feature extraction on each process parameter in the key parameter data set to obtain a statistical feature value; the statistical characteristic value is compared with a corresponding standard value in a plate shape control knowledge base, the cause of the plate shape quality defect is determined based on the comparison result, and the plate shape control knowledge base is constructed based on data with the judgment result being qualified after the historical statistical characteristic value and the corresponding plate shape quality detection result are judged in the historical production process.

Description

Technical Field

[0001] This application relates to the field of medium and heavy plate shape quality control technology, and in particular to a method, system, equipment and medium for analyzing medium and heavy plate shape quality defects. Background Technology

[0002] Currently, medium and heavy plates, as important basic materials, are widely used in key fields such as shipbuilding, bridges, pressure vessels, and building structures. Their shape quality, including flatness and cross-sectional shape, is a core indicator for evaluating product grade and performance. During the rolling process, shape quality is intricately affected by dozens of process parameters, including roll gap, pass reduction, temperature distribution, and straightening force. These parameters exhibit strong nonlinear correlations, and numerous interference sources in the production environment contribute to the extremely complex causes of shape defects.

[0003] However, existing plate shape quality defects mainly rely on manual analysis based on the experience of process technicians. This method is time-consuming, inefficient, and struggles to accurately pinpoint the key process parameters causing the defects. Summary of the Invention

[0004] The summary section introduces a series of simplified concepts, which will be further explained in detail in the detailed description section. This summary section is not intended to limit the key and essential technical features of the claimed technical solution, nor is it intended to determine the scope of protection of the claimed technical solution.

[0005] In a first aspect, embodiments of this application provide a method for analyzing the shape quality defects of a medium-thick plate, the method comprising:

[0006] Obtain multiple process parameters and corresponding plate shape and quality data during the production of medium and heavy plates;

[0007] Based on the correlation analysis algorithm, each process parameter and the corresponding plate shape quality data are analyzed to determine the degree of influence of the process parameter.

[0008] If the influence value is greater than a preset correlation threshold, a key parameter dataset is constructed based on the process parameters corresponding to the influence value.

[0009] For each process parameter in the key parameter dataset, feature extraction is performed to obtain statistical feature values;

[0010] The statistical feature values ​​are compared with the corresponding standard values ​​in the plate shape control knowledge base, and the causes of plate shape quality defects are determined based on the comparison results. The plate shape control knowledge base is constructed based on data from historical production processes that have been judged as qualified after comparing historical statistical feature values ​​with corresponding plate shape quality test results.

[0011] In one embodiment of the present invention, the step of analyzing each process parameter and the corresponding plate shape quality data based on a correlation analysis algorithm to determine the degree of influence of the process parameter includes:

[0012] Based on the maximum mutual information method, the maximum mutual information coefficient between each process parameter and the corresponding plate shape quality data is determined.

[0013] The maximum mutual information coefficient is used as the influence value corresponding to the process parameter.

[0014] In one embodiment of the present invention, the statistical feature values ​​include maximum value, minimum value, average value, variance, standard deviation, and coefficient of variation. The step of extracting features from each process parameter in the key parameter dataset to obtain the statistical feature values ​​includes:

[0015] Feature extraction is performed on each process parameter in the key parameter dataset to obtain the maximum value, minimum value, average value, variance, and standard deviation of each process parameter;

[0016] The ratio of the standard deviation to the mean is used as the coefficient of variation.

[0017] In one embodiment of the present invention, the plate shape control knowledge base is constructed in the following manner:

[0018] Obtain historical process parameters and corresponding historical plate shape quality inspection results during the historical production process;

[0019] The historical process parameters are classified based on steel grade data and specification data to obtain a data subset;

[0020] Feature extraction is performed on the process parameters in each data subset to obtain historical statistical feature values;

[0021] Based on the comparison results between the historical coefficient of variation in the historical statistical feature values ​​and the preset stability threshold, it is determined whether the corresponding historical process parameters are in a steady state.

[0022] If the historical process parameters are in a steady state and the corresponding historical plate shape quality test results are qualified, the production sample corresponding to the historical process parameters will be regarded as an excellent sample.

[0023] Based on the excellent samples, the random forest algorithm is used to predict the historical standard values ​​and historical optimal ranges of historical process parameters under the corresponding steel grade data and specification data.

[0024] Based on the historical standard values ​​and the historical optimal range, the plate shape control knowledge base is constructed.

[0025] In one embodiment of the present invention, the classification of the historical process parameters based on steel grade data and specification data to obtain a data subset includes:

[0026] Based on the steel grade data, the historical process parameters are divided into a first group;

[0027] The first group is divided based on the slab thickness data in the specification data to obtain the second group;

[0028] The second group is divided based on the steel plate thickness data in the specification data to obtain the third group;

[0029] The third group is divided based on the steel plate width data in the specification data to obtain the fourth group;

[0030] The fourth group is divided based on the steel plate length data in the specification data to obtain the data subset.

[0031] In one embodiment of the present invention, comparing the statistical feature value with the corresponding standard value in the plate shape control knowledge base, and determining the cause of the plate shape quality defect based on the comparison result, includes:

[0032] Based on the statistical feature value, the deviation value between the statistical feature value and the corresponding standard value in the plate shape control knowledge base is determined;

[0033] Based on the comparison results between the deviation value and the corresponding optimal range in the plate shape control knowledge base, the process parameters in the key parameter dataset that are in an abnormal state are determined.

[0034] Based on the process parameters that are in an abnormal state, determine the cause of plate shape quality defects.

[0035] In one embodiment of the present invention, after comparing the statistical feature value with the corresponding standard value in the plate shape control knowledge base, and determining the cause of the plate shape quality defect based on the comparison result, the process includes:

[0036] An analysis report containing parameter deviation information is generated based on the causes of plate shape quality defects, and the display device is controlled to display the analysis report in the form of charts.

[0037] Secondly, this application proposes a plate shape quality defect analysis system for medium and thick plates, the system comprising: a data analysis module, a feature extraction module, and a data comparison module;

[0038] The data analysis module is configured to: acquire multiple process parameters and corresponding plate shape quality data during the production of medium and heavy plates; analyze each process parameter and corresponding plate shape quality data based on a correlation analysis algorithm to determine the degree of influence of the process parameter;

[0039] The feature extraction module is configured to: construct a key parameter dataset based on the process parameters corresponding to the influence value when the influence value is greater than a preset correlation threshold; and extract features from each process parameter in the key parameter dataset to obtain statistical feature values.

[0040] The data comparison module is configured to compare the statistical feature value with the corresponding standard value in the plate shape control knowledge base, and determine the cause of the plate shape quality defect based on the comparison result. The plate shape control knowledge base is constructed based on data that has been judged as qualified after historical statistical feature values ​​and corresponding plate shape quality test results in the historical production process.

[0041] Thirdly, an electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program stored in the memory to implement the steps of a method for analyzing the shape quality defects of a medium-thick plate as described in any of the first aspects above.

[0042] Fourthly, this application also proposes a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the steps of the method for analyzing the shape quality defects of a medium-thick plate according to any one of the first aspects.

[0043] In summary, the method for analyzing the shape quality defects of medium and heavy plates according to embodiments of this application acquires multi-dimensional process parameters and final plate shape quality data during the production process, and uses correlation analysis algorithms to quantify the influence of each process parameter on the plate shape quality. This scientifically selects key parameter datasets from a massive number of process variables, solving the problems of long analysis cycles, low efficiency, and difficulty in accurately locating key process parameter variables that cause defects in traditional methods. Furthermore, by extracting statistically significant feature values ​​from key parameters and systematically comparing them with standard values ​​in a plate shape control knowledge base built based on historical qualified production data, the core process links and parameter deviations that cause plate shape quality defects can be quickly and accurately located, significantly improving the efficiency and accuracy of defect analysis and effectively preventing the recurrence of defects.

[0044] The method for analyzing the shape quality defects of medium and thick plates proposed in this application, along with other advantages, objectives, and features of this application, will be partly apparent from the following description and partly understood by those skilled in the art through study and practice of this application. Attached Figure Description

[0045] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit this specification. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0046] Figure 1 This is a flowchart illustrating a method for analyzing the shape quality defects of a medium-thick plate, as provided in an embodiment of this application.

[0047] Figure 2 A schematic diagram of a medium-thick plate shape quality defect analysis system provided in this application embodiment;

[0048] Figure 3 This is a schematic diagram of an electronic device for analyzing the shape quality defects of a medium-thick plate, provided as an embodiment of this application. Detailed Implementation

[0049] To better understand the technical solutions provided in the embodiments of this specification, the technical solutions of the embodiments of this specification will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of this specification and the specific features in the embodiments are detailed descriptions of the technical solutions of the embodiments of this specification, rather than limitations on the technical solutions of this specification. In the absence of conflict, the embodiments of this specification and the technical features in the embodiments can be combined with each other.

[0050] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The term "two or more" includes two or more cases.

[0051] Please see Figure 1This is a flowchart illustrating a method for analyzing the shape quality defects of a medium-thick plate, as provided in an embodiment of this application. Specifically, it includes:

[0052] S110. Obtain multiple process parameters and corresponding plate shape and quality data during the production of medium and heavy plates.

[0053] For example, the plate shape quality defect analysis process database stores control parameters related to plate shape quality control, characteristic data of key parameters for plate shape quality control, excellent samples of key parameters for plate shape quality control, and knowledge about plate shape quality control. Specifically, the plate shape quality defect analysis process database mainly includes four types of data tables: Plate shape quality control parameter data table: stores position sequence control parameters related to plate shape quality control, including final rolling thickness, final rolling temperature, ACC (Accelerated Cooling Control) set water ratio, and total pre-straightening force; Plate shape quality control key parameter characteristic value data table: stores the characteristic value calculation results of key control parameters during the production process of a steel plate, on a per-plate basis; Excellent sample library of key parameters for plate shape quality control: stores the characteristic values ​​of key control parameters for plate shape that have been judged as excellent samples by the excellent sample judgment module; Plate shape quality control knowledge base: uses steel grade and specification as index conditions to store the standard values ​​and optimal ranges of key parameters for plate shape control corresponding to different steel grades and specifications.

[0054] The process data acquisition module collects control parameters in real time during the plate shape control process, that is, it acquires multiple process parameters and corresponding plate shape quality data in the medium and heavy plate production process, which are used for further excellent sample processing, feature extraction, and knowledge base construction. The process parameters include the acquisition time, acquisition location, and data value. During the acquisition cycle of each process parameter, the process data acquisition module matches all acquired process parameters with the steel plate position based on the steel plate tracking information, forming a position sequence data to ensure that each set of process parameters corresponds to a specific production stage of the steel plate.

[0055] The acquired process parameters are cleaned, primarily addressing issues such as missing values, erroneous data, outliers, and noise that are inherently prevalent in actual data. The cleaned process parameters then undergo further cleaning and preprocessing, mainly achieved through a cleaning rule engine, which is parameterized and configurable. Specifically, methods such as ignoring or filling in missing values ​​are employed. Different handling methods are used for different attribute value gaps: for gaps in measurable attributes, deletion is used; for non-measurable attributes, they are supplemented based on domain knowledge. Erroneous data is handled by performing limit checks on the collected actual process parameters to confirm their validity; if they exceed the limit range, the limit value is used as a replacement. Data filtering is performed by combining the arithmetic mean method and the median filtering method. First, the median filtering method is used to filter out sampled values ​​biased by impulse interference, and then an arithmetic mean is calculated. This removes impulse interference and smooths the sampled values. The principle is: x1≤x2≤…≤x N Where x is the sampled value, N is the number of samples, and 3 ≤ N ≤ 5. Y = (x² + x³ + ... + x...) N-1 ) / (N-2), where y is the filtered data.

[0056] S120. Analyze each process parameter and the corresponding plate shape quality data based on the correlation analysis algorithm to determine the degree of influence of the process parameter.

[0057] For example, in the actual production of medium and heavy plates, numerous and interdependent process parameters affect the final plate shape quality, making it difficult for traditional methods to accurately identify the actual contribution of each parameter to the quality result. This application aims to solve this problem from a data-driven perspective by introducing a correlation analysis algorithm. Its core principle lies in using mathematical methods to calculate the statistical correlation strength between the data sequence of each process parameter and the final plate shape quality data sequence, thereby transforming a qualitative understanding of the process influence into a quantifiable influence value. This influence value objectively reflects the tightness of the linear or nonlinear relationship between a specific parameter and plate shape quality, providing direct numerical evidence for the subsequent scientific selection of a few key control variables from a massive number of parameters.

[0058] S130. If the influence value is greater than a preset correlation threshold, construct a key parameter dataset based on the process parameters corresponding to the influence value.

[0059] For example, after obtaining the quantified impact values ​​of each process parameter, this application sets a preset correlation threshold as a screening criterion to focus analytical resources and eliminate interference from irrelevant variables. Only when the impact value of a certain process parameter exceeds this threshold is it determined to be a key parameter with substantial impact on plate shape quality and included in the key parameter dataset for subsequent in-depth analysis. This transforms the extensive and exploratory correlation quantification results in the early stages into a clear, finite set of parameters that can be used for subsequent feature extraction and knowledge comparison, thereby achieving analytical focus from broad-spectrum correlation detection to core variable locking. Specifically: the correlation threshold is 0.4. Using 0.4 as the judgment standard, if the impact value is greater than this value, the process parameter corresponding to that impact value is considered a key parameter for plate shape quality control; otherwise, the process parameter corresponding to that impact value is not considered a key parameter.

[0060] S140. Extract features from each process parameter in the key parameter dataset to obtain statistical feature values;

[0061] For example, after selecting the key parameter dataset, in order to transform the high-dimensional original data sequence representing the specific production process into low-dimensional features that can represent the overall fluctuation state and stability of the batch production, this application extracts features from each key process parameter in the key parameter dataset to obtain statistical feature values. The core principle is to quantify the dynamic behavior of process parameters in the production process from multiple dimensions such as central tendency, dispersion, and relative volatility by calculating a series of statistical quantities such as maximum, minimum, average, variance, standard deviation, and coefficient of variation. The obtained statistical feature values, on the one hand, eliminate the temporal sequence of the original data sequence, making the data between different production batches comparable; on the other hand, these statistical feature values ​​constitute the core data foundation for subsequent determination of production steady state, construction of knowledge standards, and defect comparison, and are the key transformation link for abstracting the specific production process into an analyzable and evaluable model object.

[0062] S150. The statistical feature value is compared with the corresponding standard value in the plate shape control knowledge base, and the cause of the plate shape quality defect is determined based on the comparison result. The plate shape control knowledge base is constructed based on the data that is qualified after the historical statistical feature value and the corresponding plate shape quality test result are judged in the historical production process.

[0063] For example, the statistical characteristic values ​​of the process parameters of the current production batch are quantitatively compared with the shape control knowledge base to trace the root cause of defects. This shape control knowledge base is not a static set of experience values, but rather an "excellent sample" based on historical production data whose process parameters are stable (determined by historical statistical characteristic values) and whose final shape inspection is qualified. It is refined through intelligent algorithms. It systematically stores the standard values ​​that key process parameters should achieve for different steel grades and specifications, along with their allowable optimal ranges. When a shape defect occurs, the system compares the statistical characteristic values ​​(such as average value, volatility, etc.) extracted from the current production batch with the standard values ​​for the corresponding steel grade and specification in the knowledge base, calculating quantitative indicators such as deviation values ​​and deviation percentages. By identifying which key parameter characteristic values ​​significantly deviate from their historical optimal range, the most likely process steps and parameter anomalies causing the current shape quality defect can be objectively and quantitatively pinpointed, thereby achieving accurate and rapid diagnosis from "abnormal quality results" to "causes in the production process."

[0064] In summary, the medium-thick plate shape quality defect analysis method proposed in this application obtains multi-dimensional process parameters and final plate shape quality data during the production process, and uses correlation analysis algorithms to quantify the influence of each process parameter on plate shape quality. This scientifically selects key parameter datasets from a massive amount of process variables, solving the problems of long analysis cycles, low efficiency, and difficulty in accurately locating key process parameter variables that cause defects in traditional methods. Furthermore, by extracting statistically significant feature values ​​from key parameters and systematically comparing them with standard values ​​in a plate shape control knowledge base built based on historical qualified production data, the core process links and parameter deviations that cause plate shape quality defects can be quickly and accurately located, significantly improving the efficiency and accuracy of defect analysis and effectively preventing the recurrence of defects.

[0065] In some examples, the analysis of each process parameter and the corresponding plate shape quality data based on the correlation analysis algorithm to determine the degree of influence of the process parameter includes:

[0066] Based on the maximum mutual information method, the maximum mutual information coefficient between each process parameter and the corresponding plate shape quality data is determined.

[0067] The maximum mutual information coefficient is used as the influence value corresponding to the process parameter.

[0068] For example, traditional linear correlation coefficients can only measure the linear relationship between variables, while in actual production, there are often complex nonlinear coupling relationships between process parameters and plate shape and quality data. The MIC (Maximal Information Coefficient) method explores the mutual information between two variables across all possible grid partitions, and can fairly and robustly capture the strength of any form of association between them, whether linear, periodic, or other complex functional relationships. Its core is to calculate the maximum value of the mutual information between each process parameter and its corresponding plate shape and quality data under various grid resolutions, and then standardize this value to obtain a maximum mutual information coefficient between 0 and 1. The closer this coefficient is to 1, the stronger the statistical dependence between the process parameter and plate shape and quality, regardless of the form of dependence; therefore, it has a more comprehensive and universal characterization ability as a value of influence.

[0069] Specifically, the Maximum Inter-Information Technology (MIC) method is used to quantify the linear or nonlinear correlation between various process parameters and plate shape quality, and to measure the degree of influence. The process parameters X and plate shape quality data Y for medium-thick plate production are divided into equal parts x and y, respectively. Different intervals can be used to divide the plate into different numbers of grids. The magnitude of the mutual information is proportional to the number of grids, and the total number of grids must be limited to xy. <n 0.6 Let n be the number of samples for random variables X and Y; let the initial partition be x = 2 and y = n. 0.6 / 2, After each calculation is completed, change the values ​​of x and y to x = x + 1 and y = n respectively. 0.6 / 2, until y=2; in each partition, it falls on the (x)th... j y j The number of samples n in each grid cell ij Dividing by the total number of samples n, we obtain the probability value of this grid cell as p(xj, yj); the probability distribution under this condition can be calculated as D| xy The mutual information I(D|) of the partitioning conditions for the current values ​​of x and y can be obtained through the mutual information calculation method. xy Its maximum value is I. max [D(x,y)]=maxI(D| xy Standardizing it yields:

[0070]

[0071] Among them, M(D) xy ) represents the standardized mutual information, I max [D(x,y)] represents the maximum mutual information, and min{x,y} represents the minimum dimension of the grid partition.

[0072] Find the maximum mutual information coefficient under each partition condition;

[0073] MIC(x,y)=max{M(D xy )} (2);

[0074] Here, MIC(x,y) represents the maximum mutual information coefficient. This approach overcomes the limitations of traditional linear correlation coefficients in analyzing complex industrial process data, enabling unbiased identification of key process parameters with strong nonlinear correlations to plate shape quality. This ensures the scientific rigor and completeness of the key parameter dataset construction. This lays a reliable data foundation for subsequent feature extraction and knowledge base comparison, avoiding the omission of key process parameters due to incomplete correlation measurement methods. Ultimately, it improves the accuracy and reliability of the entire plate shape quality defect analysis process, resulting in a more comprehensive and accurate identification of defect causes.

[0075] In some examples, the statistical feature values ​​include maximum, minimum, mean, variance, standard deviation, and coefficient of variation. The step of extracting features from each process parameter in the key parameter dataset to obtain statistical feature values ​​includes:

[0076] Feature extraction is performed on each process parameter in the key parameter dataset to obtain the maximum value, minimum value, average value, variance, and standard deviation of each process parameter;

[0077] The ratio of the standard deviation to the mean is used as the coefficient of variation.

[0078] For example, feature extraction is performed on each process parameter in the key parameter dataset to obtain quantitative indicators that can comprehensively and multidimensionally characterize the dynamic behavior and stability of the parameter in a production batch. First, the original data sequence of each process parameter (e.g., final rolling thickness, final rolling temperature) in each key parameter dataset is calculated to determine its maximum and minimum values ​​during the production process of that batch, thus characterizing the extreme value fluctuation range of the parameter. The average value of each process parameter in each key parameter dataset is calculated to reflect the degree of agreement between the parameter's central tendency and the set target. The variance and standard deviation of each process parameter in each key parameter dataset are calculated to quantify the dispersion and absolute fluctuation intensity of the parameter data around the average value. These statistics collectively constitute the basic description of the process parameter behavior.

[0079] Building upon this foundation, to further eliminate the incomparability of data fluctuations caused by differences in dimensions and inherent magnitudes of different process parameters, and to establish a normalized stability evaluation index, this application introduces the coefficient of variation as a key feature. Specifically, the standard deviation calculated in the previous step is compared with the average value, and the result of this ratio is determined as the coefficient of variation of the process parameter. A preset stability judgment threshold (e.g., 0.3) is used as a benchmark. If the coefficient of variation of a parameter is less than or equal to this threshold, it indicates that the parameter exhibits relatively small fluctuations during production and is in a stable and controlled state; conversely, it indicates significant fluctuations. This step transforms the standard deviation, which characterizes absolute dispersion, into the coefficient of variation, which characterizes relative volatility. This allows for objective comparison and comprehensive evaluation of the stability of parameters with vastly different magnitudes, such as force (unit kN) and temperature (unit °C), under the same standard.

[0080] Specifically, for each process parameter in each key parameter dataset, including final rolling thickness, final rolling temperature, ACC setpoint water ratio, and total pre-straightening force, the maximum, minimum, average, variance, standard deviation, and coefficient of variation are extracted for each process parameter. To avoid incomparability in data stability due to differences in data magnitude, the coefficient of variation (cv) is introduced as a key indicator of parameter normalization stability, thus eliminating the impact of magnitude differences.

[0081]

[0082] Where σ is the standard deviation, μ is the mean, and cv is the coefficient of variation. This method compresses high-dimensional, time-series raw process data into a set of low-dimensional statistical characteristic values ​​with clear physical meaning. In particular, by introducing the coefficient of variation as a normalization index, it effectively eliminates dimensional interference between parameters, providing a precise and comparable data foundation for accurately determining the steady state of the production process, screening reproducible excellent samples, and building a high-quality knowledge base.

[0083] In some examples, the plate shape control knowledge base is constructed in the following manner:

[0084] Obtain historical process parameters and corresponding historical plate shape quality inspection results during the historical production process;

[0085] The historical process parameters are classified based on steel grade data and specification data to obtain a data subset;

[0086] Feature extraction is performed on the process parameters in each data subset to obtain historical statistical feature values;

[0087] Based on the comparison results between the historical coefficient of variation in the historical statistical feature values ​​and the preset stability threshold, it is determined whether the corresponding historical process parameters are in a steady state.

[0088] If the historical process parameters are in a steady state and the corresponding historical plate shape quality test results are qualified, the production sample corresponding to the historical process parameters will be regarded as an excellent sample.

[0089] Based on the excellent samples, the random forest algorithm is used to predict the historical standard values ​​and historical optimal ranges of historical process parameters under the corresponding steel grade data and specification data.

[0090] Based on the historical standard values ​​and the historical optimal range, the plate shape control knowledge base is constructed.

[0091] For example, firstly, the sequence of all historical process parameters recorded during historical production is obtained, along with their corresponding historical plate shape quality inspection results. Then, to establish refined process standards, the complex historical process parameters are classified at multiple levels based on steel grade and specification data (such as slab thickness, plate thickness, width, and length), forming well-defined data subsets. Each data subset represents a set of production data under a specific combination of steel grades and specifications. Next, the process parameters in each data subset undergo the same feature extraction operation as the current production batch, calculating various historical statistical feature values, including the historical coefficient of variation. The core screening logic of the construction process is as follows: based on a preset stability threshold (e.g., 0.3) and the comparison result with the historical coefficient of variation, it is determined whether the process parameters of the corresponding historical production sample are in a stable and controlled steady state; only those production samples determined to be in a steady state, and whose associated historical plate shape quality inspection results are also qualified, can be identified as excellent samples with replicability and guidance value. This step ensures that the data foundation entering the plate shape control knowledge base construction stage possesses both process stability and excellent results.

[0092] After obtaining a sufficient number of excellent samples, the construction process enters the intelligent modeling stage. For the set of excellent samples under the same steel grade and specification category, the random forest ensemble learning algorithm is used for training and prediction. This algorithm can efficiently handle multivariate and nonlinear industrial data. By constructing a large number of decision trees and summarizing their prediction results, it learns the complex mapping relationship between the feature values ​​of key process parameters in excellent samples and the final high-quality plate shape results. Based on this trained model, the system can predict and output the historical standard values ​​(such as the ideal center point of statistical feature values) and the historical optimal range of allowable fluctuations (such as the upper and lower limits of the reasonable range) that each key process parameter should achieve under a specific steel grade and specification. Finally, the parameter standard values ​​and optimal ranges under all steel grade and specification categories are systematically stored and managed, thus forming a plate shape control knowledge base that can serve online quality comparison. This plate shape control knowledge base is not static; it will continuously iterate and update with newly generated excellent samples, thereby achieving self-evolution and improvement of knowledge.

[0093] Specifically, in the step of determining whether the corresponding historical process parameter is in a steady state based on the comparison results of the historical coefficient of variation in the historical statistical feature values ​​and the preset stability threshold, the stability judgment criterion is: the 3σ principle: when about 99.7% of the data is in the range of (μ-3σ, μ+3σ), the data is considered stable, where σ is the standard deviation and μ is the mean. For this step, the stability threshold is set to 0.3 as the standard for judging data stability. cv < 0.3 indicates that the data is in a steady state, otherwise it is not in a steady state. The stability judgment result needs to be reflected in a quantitative form: the stability judgment result (S) of the historical coefficient of variation corresponding to the i-th historical process parameter in the data subset is quantitatively expressed.

[0094]

[0095] Where s is the stability judgment result after quantification, n is the number of data items of the key parameters for plate shape control, i is the i-th historical process parameter in the data subset, and cv i Let be the historical coefficient of variation corresponding to the i-th historical process parameter in the data subset.

[0096] Based on the aforementioned excellent samples, in the step of predicting the historical standard values ​​and historical optimal ranges of historical process parameters under corresponding steel grade and specification data using the random forest algorithm, numerous excellent samples already obtained for the same steel grade and specification are utilized. A nonlinear correlation model between key historical process parameters and historical plate shape quality data is constructed based on the random forest algorithm to predict the historical standard values ​​and historical optimal ranges corresponding to the historical process parameters. The random forest model parameters are optimized by adjusting the number of decision trees to 300, the maximum tree depth to 15, the minimum number of samples per tree to 3, and increasing the split threshold to 5. These historical standard values ​​and historical optimal ranges are continuously updated during the production process, achieving self-optimization and improvement of the plate shape control key parameter knowledge base. When the plate shape control knowledge base is applied to online defect analysis, it can compare the statistical feature values ​​extracted from real-time process parameters with the corresponding precise standards in the plate shape control knowledge base, thereby quickly identifying abnormal parameters deviating from the optimal range and achieving precise tracing of defect causes. This process not only greatly improves the efficiency and accuracy of analyzing complex plate shape defects, but also adapts to process changes and the production of new steel grades through continuous updates to the plate shape control knowledge base. It fundamentally establishes a long-term mechanism to prevent the recurrence of defects and achieve continuous process optimization, significantly improving the level of intelligent quality control and product consistency in the production of medium and heavy plates.

[0097] In some examples, the classification of historical process parameters based on steel grade data and specification data yields a data subset, including:

[0098] Based on the steel grade data, the historical process parameters are divided into a first group;

[0099] The first group is divided based on the slab thickness data in the specification data to obtain the second group;

[0100] The second group is divided based on the steel plate thickness data in the specification data to obtain the third group;

[0101] The third group is divided based on the steel plate width data in the specification data to obtain the fourth group;

[0102] The fourth group is divided based on the steel plate length data in the specification data to obtain the data subset.

[0103] For example, the first step in building a knowledge base for plate shape control is to refine the classification of historical process parameters to establish a process standard system that strictly corresponds to specific production conditions. This classification process employs a multi-level progressive division strategy: First, all historical process parameters are initially divided based on steel grade data, grouping historical process parameters of steel plates of the same grade or material together to form a first group based on steel grade; then, based on the first group, a second division is performed by introducing the slab thickness dimension from the specification data. According to preset slab thickness grouping rules (e.g., grouping slabs with thicknesses ranging from 40mm to 500mm by a standard of ±1mm difference in thickness), historical process parameters under the same steel grade are initially subdivided to obtain a second group; subsequently, within the second group, a third division is performed based on the steel plate thickness from the specification data, following the steel plate thickness grouping rules (e.g., for steel plate thicknesses from 5mm to 400mm, also grouped by...). The data is further divided into a third group (with an interval of ±1mm). Next, based on the steel plate width in the specification data, the historical process parameters within the third group are divided a fourth time, forming a fourth group according to the preset steel plate width grouping rules (e.g., for a width range of 1250mm to 4250mm, the range is divided into intervals of 0.1 meters). Finally, based on the historical process parameters within the fourth group, a fifth division is performed according to the steel plate length in the specification data, following the preset steel plate length grouping rules (e.g., for lengths from 2m to 52m, the range is divided into groups of 3m). This completes the final subdivision, resulting in a clearly defined and uniquely conditional data subset that precisely corresponds to a specific combination of steel grade, slab thickness, steel plate thickness, steel plate width, and steel plate length.

[0104] This allows the historical statistical feature values ​​(such as average values ​​and optimal ranges) extracted from each data subset to accurately reflect the ideal process state under specific production conditions (steel grade and precise dimensions). The resulting plate shape control knowledge base is no longer a general set of empirical values, but a "personalized" standard strictly matched to specific production orders. When plate shape defects occur in online production, the system can first quickly match the steel grade and specification information of the current steel plate with the classification index of the knowledge base, locating the most relevant standard data subset, and then comparing the feature values. This greatly improves the targeting and accuracy of the comparison, avoiding misjudgments or omissions caused by broad standards, ensuring that the deviation in the identified defect cause parameters truly stems from process control failure under the current specific production conditions, rather than interference from other irrelevant conditions. This achieves a qualitative leap in defect tracing from "rough benchmarking" to "precise matching."

[0105] In some examples, comparing the statistical feature values ​​with the corresponding standard values ​​in the plate shape control knowledge base, and determining the cause of plate shape quality defects based on the comparison results, includes:

[0106] Based on the statistical feature value, the deviation value between the statistical feature value and the corresponding standard value in the plate shape control knowledge base is determined;

[0107] Based on the comparison results between the deviation value and the corresponding optimal range in the plate shape control knowledge base, the process parameters in the key parameter dataset that are in an abnormal state are determined.

[0108] Based on the process parameters that are in an abnormal state, determine the cause of plate shape quality defects.

[0109] For example, firstly, based on the extracted statistical feature values ​​(especially the average value representing the central tendency of the parameters), an arithmetic operation is performed with the standard values ​​of the corresponding key parameters retrieved from the plate shape control knowledge base that precisely match the current steel plate grade and specifications, to calculate the deviation value between the two. Here, the standard value is a central reference value in the knowledge base, learned through a random forest algorithm based on historical excellent samples, representing the ideal state of the process parameters under specific production conditions. The purpose of calculating the deviation value is to quantitatively align the current production state with the historical best process benchmark, transforming abstract quality differences into numerical deviations of specific parameters, providing a direct and measurable data basis for subsequent anomaly judgment, thereby overcoming the ambiguity of qualitative descriptions such as "too large" or "too small" in traditional experience comparisons.

[0110] After obtaining the deviation values ​​of each key parameter, the system performs the core abnormal state determination. This step compares the deviation value with the optimal range corresponding to the same parameter recorded in the plate shape control knowledge base. The optimal range is also derived from statistical learning of historical excellent samples, defining a reasonable fluctuation range around the standard value (usually with clear upper and lower limits). The determination logic is as follows: if the absolute value of the deviation of a parameter, or its deviation direction and magnitude relative to the standard value, causes the actual characteristic value of the parameter to fall outside its optimal range, the system determines that the parameter is in an abnormal state; conversely, if the characteristic value falls within the optimal range, it is determined to be in a normal state. By traversing the process parameters in all key parameter datasets and completing the above comparison, the system can accurately identify one or more process parameters in an abnormal state. Finally, based on the process parameters in an abnormal state, the system determines the cause of the plate shape quality defect. The reasoning logic is that those parameters that significantly deviate from their historical optimal process range are most likely the direct process variables that cause the current plate shape quality to fail to meet the requirements. The system can output all abnormal state parameters and their deviation values ​​as core evidence to determine the cause of plate shape defects, thus completing the diagnostic closed loop from quantitative deviation to qualitative attribution.

[0111] By executing a standard process of "calculating deviation values ​​- comparing to the optimal range - determining abnormal states" for each process parameter, it is possible to automatically and quickly identify which process control steps have deviated from historically validated best practices. This allows complex plate shape defects to be directly traced back to specific, controllable process parameter anomalies. This not only greatly improves the efficiency and accuracy of defect analysis, achieving cause localization within minutes or even seconds, but also ensures that the analysis conclusions are entirely data-driven, eliminating subjective assumptions. This makes subsequent process adjustments and optimizations clear and evidence-based, effectively preventing the recurrence of similar defects.

[0112] In some examples, after comparing the statistical feature values ​​with the corresponding standard values ​​in the plate shape control knowledge base, and determining the cause of the plate shape quality defect based on the comparison results, the process includes:

[0113] An analysis report containing parameter deviation information is generated based on the causes of plate shape quality defects, and the display device is controlled to display the analysis report in the form of charts.

[0114] For example, based on the determined causes of plate shape quality defects, a structured analysis report is automatically generated. The core content of this report is a detailed data list containing parameter deviation information. This list systematically lists each key process parameter determined to be in an abnormal state, clearly recording its specific statistical characteristic value (such as the actual average value), the standard value retrieved from the plate shape control knowledge base, the calculated deviation value, the deviation percentage, and the judgment conclusion regarding whether the parameter exceeds its corresponding optimal range. The logical basis for generating the report lies in summarizing, organizing, and formatting the discrete abnormal parameter judgment results and quantified deviation data obtained in the aforementioned comparative analysis steps, transforming them into a complete document that process engineers can read and use for decision-making.

[0115] Subsequently, the system performs a visualization presentation, controlling the display device to present the analysis report in chart form. This "chart form" may include, but is not limited to: visually comparing abnormal parameters and their deviation values ​​using bar charts or radar charts, highlighting parameters that exceed the optimal range; plotting the characteristic value curves of key parameters for the current production batch against the standard value range curves in the knowledge base on the same trend chart for historical comparison; and generating a comprehensive analysis interface containing key data tables and a summary of conclusions. The purpose of controlling the display device to present the data-driven analysis conclusions is to convey them to the user in the most intuitive and easily understandable way, completing the final link from back-end data analysis to front-end decision support. This step ensures the operability and usability of the analysis results, enabling process engineers to quickly focus on the problem and understand the root cause of the defect.

[0116] The present invention will be described in detail below with reference to the embodiments, but these should not be construed as limiting the scope of protection of the present invention.

[0117] Example:

[0118] The main parameters of the medium-thick plate production line selected in the embodiments of this invention are as follows:

[0119] Slab thickness range: 40~500mm;

[0120] Steel plate thickness range: 5~400mm;

[0121] Steel plate width range: 1250~4250mm;

[0122] Steel plate length range: 2–52m;

[0123] The products are mainly steel for pipelines, marine engineering, and shipbuilding, and have high requirements for the shape and quality of the steel plates.

[0124] The analysis of the shape quality defects of this production line lacks systematic support, the quality analysis cycle is long, which is likely to cause batch accidents; there is a lack of a way to trace the quality throughout the process and analyze genetic influences, making it inconvenient for each process to collaborate in improving product quality and precise slitting; the tracing of quality problems is slow, and customer complaints cannot be handled in a timely manner.

[0125] All kinds of detection instruments and meters equipped on the production line send the actual values of various process parameters, including shape control parameters, in the production process of medium and heavy plates to the on-site L1-level control system. The process data acquisition and storage module of this method communicates with the on-site L1-level shape control system of medium and heavy plates using the TCP / IP protocol, and continuously collects the actual production process data in real time at a frequency of 100 ms. The specific data acquisition items are shown in Table 1 and Table 2:

[0126]

[0127]

[0128] Table 1

[0129] Serial Number Process Data Object 1 Roughing rolling passes [-] 2 Finishing rolling passes [-] 3 Rolling length [mm] 4 Final rolling target width [mm] 5 Mother plate rolling width [mm] 6 Final rolled thickness [mm] 7 Target final rolled thickness [mm] 8 Standard value for final rolling temperature [°C] 9 Average final rolling temperature [°C] 10 Standard temperature value after water cooling [°C] 11 Actual temperature after water cooling [°C] 12 Average final cooling temperature [°C] 13 Total straightening force before straightening [kN] 14 Pre-straightening inlet roll gap [mm] 15 Pre-straightening exit roll gap [mm] 16 Pre-correction secondary calculation of straightening force 17 ACC water ratio setting [-] 18 Hot straightening temperature [°C] 19 Thermal straightening force [kN] 20 Heat straightening inlet setting roll gap [mm] 21 Actual roll gap at the hot straightening inlet [mm] 22 Straightening path number [-] 23 Heat straightening exit roll gap setting [mm] 24 Actual roll gap at the exit of heat straightening [mm]

[0130] Table 2

[0131] In each data acquisition cycle, each data acquisition item includes the acquisition time, acquisition location, and data value of the data. This module matches all the process parameters collected this time with the steel plate position according to the steel plate tracking information, forms position sequence data, and stores it in the shape quality control process data table.

[0132] The process parameters stored in the shape quality control process data table directly come from the on-site control link. In order to eliminate the clutter, repetition, and incompleteness of these data, guided by the shape quality control process knowledge, reorganize and process the process parameters to provide clean, accurate, and more targeted process data for the subsequent analysis of shape quality defects, thereby improving the efficiency and accuracy of the analysis of shape quality defects. The means of processing include handling missing values, error data, dimension matching, and filtering processing, etc.

[0133] Use the maximum mutual information technology MIC to quantitatively calculate the linear or non-linear correlation between each process parameter and the shape quality data, and measure the degree of influence. Obtain the key parameters with a greater degree of influence on shape control from among the numerous control parameters affecting shape quality and construct an object set. Take 0.4 as the threshold for parameter selection, and use the parameters whose absolute value of correlation is greater than the given threshold as the objects for further analysis and research. As shown in Table 3, thus find the key control parameter objects that play a key role in shape quality.

[0134] Serial Number Key control parameters Correlation 1 Roughing rolling passes [-] 0.431 2 Finishing rolling passes [-] 0.481 3 Final rolled thickness [-] 0.528 4 Final rolling temperature [-] 0.637 5 Temperature after water cooling [-] 0.64 6 Final cooling temperature [-] 0.471 7 Total corrective force before straightening [-] 0.435 8 Pre-straightening inlet roll gap [-] 0.481 9 Pre-straightening exit roll gap [-] 0.417 10 ACC water ratio setting [-] 0.441 11 Thermal straightening temperature [-] 0.455 12 Total straightening force during thermal correction [-] 0.627 13 Actual roll gap at the hot straightening inlet [-] 0.481 14 Actual roll gap at the hot straightening exit [-] 0.581 15 Straightening path number [-] 0.438

[0135] Table 3

[0136] After a piece of material is produced, for each parameter in the key parameter dataset, typical characteristics are calculated using the corresponding data for the current material. These characteristics are used to characterize the data features of that key parameter during the material's production process. The calculation results of the characteristic values ​​for a certain steel plate are shown in Table 4 (Steel Plate 1) and Table 5 (Steel Plate 2).

[0137] Serial Number Key control parameters Maximum value Minimum value average value variance Standard deviation coefficient of variation 1 Roughing rolling passes [-] 7 7 7 0.00 0.00 0 2 Finishing rolling passes [-] 9 9 9 0.00 0.00 0 3 Final rolled thickness [mm] 18.86 16.91 17.67 378.48 19.45 1.10 4 Final rolling temperature [°C] 804 742 768.76 483801.83 695.56 0.90 5 Temperature after water cooling [°C] 650 586 618.36 270034.07 519.65 0.84 6 Final cooling temperature [°C] 598 532 579.12 232938.12 482.64 0.83 7 Total straightening force before straightening [kN] 861 777 821.83 4946851.58 2224.15 2.71 8 Pre-straightening inlet roll gap [mm] 19.81 18.99 19.12 174.05 13.19 0.69 9 Pre-straightening exit roll gap [mm] 19.72 18.95 19.23 242.62 15.58 0.81 10 ACC water ratio setting [-] 1.4 1 1.1 0.16 0.40 0.37 11 Hot straightening temperature [°C] 577 509 547.81 613665.89 783.37 1.43 12 Total straightening force for thermal straightening [kN] 1373.13 975.38 1089.35 1735034.56 1317.21 1.21 13 Actual roll gap at the hot straightening inlet [mm] 19.52 18.98 19.21 218.79 14.79 0.77 14 Actual roll gap at the exit of heat straightening [mm] 19.22 18.95 19.01 121.57 11.03 0.58 15 Straightening path [-] 1 1 1 0.00 0.00 0

[0138] Table 4

[0139] Serial Number Key control parameters Maximum value Minimum value average value variance Standard deviation coefficient of variation 1 Roughing rolling passes [-] 7 7 7 0.00 0.00 0 2 Finishing rolling passes [-] 9 9 9 0.00 0.00 0 3 Final rolled thickness [mm] 16.86 16.76 16.81 3.546 1.883 0.112 4 Final rolling temperature [°C] 777 732 747.11 1031.76 32.12 0.043 5 Temperature after water cooling [°C] 662 601 630.71 9805.25 99.02 0.157 6 Final cooling temperature [°C] 637 586 609.11 2938.81 54.21 0.089 7 Total straightening force before straightening [kN] 831 771 800.73 29362.95 171.36 0.214 8 Pre-straightening inlet roll gap [mm] 17.06 16.95 16.99 0.35 0.59 0.035 9 Pre-straightening exit roll gap [mm] 16.99 16.89 16.9 9.46 3.08 0.182 10 ACC water ratio setting [-] 1.4 1 1.2 0.01 0.09 0.077 11 Hot straightening temperature [°C] 602 563 589.53 21032.07 145.02 0.246 12 Total straightening force for thermal straightening [kN] 1274.63 1074.31 1199.64 3743.19 61.18 0.051 13 Actual roll gap at the hot straightening inlet [mm] 16.95 16.88 16.9 5.44 2.33 0.138 14 Actual roll gap at the exit of heat straightening [mm] 16.95 16.86 16.89 21.26 4.61 0.273 15 Straightening path number [-] 1 1 1 0.00 0.00 0

[0140] Table 5

[0141] After calculating statistical characteristic values ​​for the steel plates produced off the production line, the results are used to determine whether a sample is considered excellent. This process is used to eliminate unstable data during production and obtain stable sample data with replicable value, providing reliable process data for further building a plate shape control knowledge base. Using a preset stability threshold of 0.3, steel plate 1, as described above, does not meet the criteria for excellent samples, while steel plate 2 does.

[0142] The key parameter feature values ​​of the aforementioned steel plate 2 are used as excellent samples. Taking into account numerous excellent samples already obtained for steel plates of the same type and specification, a random forest intelligent algorithm is used to update the standard values ​​and optimal ranges of the key parameters for shape control for the same type and specification, thus constructing a knowledge base for key parameters for shape control. Table 6 shows the standard values ​​and optimal ranges of key parameters for shape control corresponding to a certain type and specification.

[0143] Serial Number Key control parameters Standard value Maximum value Minimum value 1 Roughing rolling passes [-] 7 7 7 2 Finishing rolling passes [-] 9 9 9 3 Final rolled thickness [mm] 16.85 18.031 15.7829 4 Final rolling temperature [°C] 760.12 791.4833 729.7955 5 Temperature after water cooling [°C] 625.81 660.0081 600.9893 6 Final cooling temperature [°C] 597.58 633.8991 558.3812 7 Total straightening force before straightening [kN] 807.91 841.8503 756.3317 8 Pre-straightening inlet roll gap [mm] 16.91 18.0047 16.1801 9 Pre-straightening exit roll gap [mm] 16.77 17.8731 16.0391 10 ACC water ratio setting [-] 1.26 1.3911 1.144 11 Hot straightening temperature [°C] 580.81 598.773 568.6331 12 Total straightening force for thermal straightening [kN] 1209.89 1278.9177 1104.3091 13 Actual roll gap at the hot straightening inlet [mm] 16.83 17.419 16.1887 14 Actual roll gap at the exit of heat straightening [mm] 16.21 17.51 16.0021 15 Straightening path number [-] 1 1 1

[0144] Table 6

[0145] After a steel plate comes off the production line, defects appear in its shape. It is necessary to automatically analyze the causes of these defects. The plate shape control key parameter feature extraction module is invoked to calculate the data characteristics of each key parameter for the current steel plate's shape control. The standard values ​​of historical process parameters for the corresponding steel grade and specification are retrieved from the shape control knowledge base and compared with these values. The deviations and percentages of deviation are calculated to determine the causes of the shape defects. The comparison results are shown in Table 7. The data in the table indicate that the final rolling temperature, the total pre-straightening force, and the total hot straightening force are the main causes of the current steel plate's shape defects.

[0146]

[0147]

[0148] Table 7

[0149] like Figure 2 As shown, this application proposes a plate shape quality defect analysis system, which includes: a data analysis module 21, a feature extraction module 22, and a data comparison module 23.

[0150] The data analysis module 21 is configured to: acquire multiple process parameters and corresponding plate shape quality data during the production of medium and heavy plates; analyze each process parameter and corresponding plate shape quality data based on a correlation analysis algorithm to determine the degree of influence of the process parameter;

[0151] The feature extraction module 22 is configured to: construct a key parameter dataset based on the process parameters corresponding to the influence value when the influence value is greater than a preset correlation threshold; and extract features from each process parameter in the key parameter dataset to obtain statistical feature values.

[0152] The data comparison module 23 is configured to compare the statistical feature value with the corresponding standard value in the plate shape control knowledge base, and determine the cause of the plate shape quality defect based on the comparison result. The plate shape control knowledge base is constructed based on data that has been judged as qualified after historical statistical feature values ​​and corresponding plate shape quality test results in the historical production process.

[0153] The effects of applying the aforementioned method in the above system can be found in the description of the aforementioned method embodiments, and will not be repeated here.

[0154] like Figure 3 As shown, this application embodiment also provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor. When the processor 320 executes the computer program 311, it implements the steps of any of the above-described methods for analyzing the shape quality defects of medium and thick plates.

[0155] Since the electronic device described in this embodiment is the device used to implement the medium-thick plate shape quality defect analysis device in the embodiments of this application, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the method described in the embodiments of this application. Therefore, how the electronic device implements the method in the embodiments of this application will not be described in detail here. Any device used by those skilled in the art to implement the method in the embodiments of this application is within the scope of protection of this application.

[0156] In practical implementation, when the computer program 311 is executed by the processor, it can achieve the following: Figure 1 Any of the corresponding implementation methods in the embodiments.

[0157] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0158] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-readable program code.

[0159] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0160] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0161] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0162] This application also provides a computer program product, which includes computer software instructions that, when executed on a processing device, cause the processing device to execute the LDPC decoding method of a solid-state drive controller.

[0163] A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0164] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0165] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

[0166] Although preferred embodiments have been described in this specification, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this specification.

[0167] Obviously, those skilled in the art can make various modifications and variations to this specification without departing from its spirit and scope. Therefore, if such modifications and variations fall within the scope of the claims and their equivalents, this specification is also intended to include such modifications and variations.

Claims

1. A method for analyzing the shape quality defects of medium-thick plates, characterized in that, The method includes: Obtain multiple process parameters and corresponding plate shape and quality data during the production of medium and heavy plates; Based on the correlation analysis algorithm, each process parameter and the corresponding plate shape quality data are analyzed to determine the degree of influence of the process parameter. If the influence value is greater than a preset correlation threshold, a key parameter dataset is constructed based on the process parameters corresponding to the influence value. For each process parameter in the key parameter dataset, feature extraction is performed to obtain statistical feature values; The statistical feature values ​​are compared with the corresponding standard values ​​in the plate shape control knowledge base, and the causes of plate shape quality defects are determined based on the comparison results. The plate shape control knowledge base is constructed based on data from historical production processes that have been judged as qualified after comparing historical statistical feature values ​​with corresponding plate shape quality test results.

2. The method for analyzing the shape quality defects of medium-thick plates according to claim 1, characterized in that, The step of analyzing each process parameter and the corresponding plate shape quality data based on a correlation analysis algorithm to determine the degree of influence of each process parameter includes: Based on the maximum mutual information method, the maximum mutual information coefficient between each process parameter and the corresponding plate shape quality data is determined. The maximum mutual information coefficient is used as the influence value corresponding to the process parameter.

3. The method for analyzing the shape quality defects of medium-thick plates according to claim 1, characterized in that, The statistical feature values ​​include maximum value, minimum value, average value, variance, standard deviation, and coefficient of variation. The step of extracting features from each process parameter in the key parameter dataset to obtain statistical feature values ​​includes: Feature extraction is performed on each process parameter in the key parameter dataset to obtain the maximum value, minimum value, average value, variance, and standard deviation of each process parameter; The ratio of the standard deviation to the mean is used as the coefficient of variation.

4. The method for analyzing the shape quality defects of medium-thick plates according to claim 1, characterized in that, The plate shape control knowledge base is constructed in the following way: Obtain historical process parameters and corresponding historical plate shape quality inspection results during the historical production process; The historical process parameters are classified based on steel grade data and specification data to obtain a data subset; Feature extraction is performed on the process parameters in each data subset to obtain historical statistical feature values; Based on the comparison results between the historical coefficient of variation in the historical statistical feature values ​​and the preset stability threshold, it is determined whether the corresponding historical process parameters are in a steady state. If the historical process parameters are in a steady state and the corresponding historical plate shape quality test results are qualified, the production sample corresponding to the historical process parameters will be regarded as an excellent sample. Based on the excellent samples, the random forest algorithm is used to predict the historical standard values ​​and historical optimal ranges of historical process parameters under the corresponding steel grade data and specification data. Based on the historical standard values ​​and the historical optimal range, the plate shape control knowledge base is constructed.

5. The method for analyzing the shape quality defects of medium-thick plates according to claim 4, characterized in that, The historical process parameters are classified based on steel grade data and specification data to obtain a data subset, including: Based on the steel grade data, the historical process parameters are divided into a first group; The first group is divided based on the slab thickness data in the specification data to obtain the second group; The second group is divided based on the steel plate thickness data in the specification data to obtain the third group; The third group is divided based on the steel plate width data in the specification data to obtain the fourth group; The fourth group is divided based on the steel plate length data in the specification data to obtain the data subset.

6. The method for analyzing the shape quality defects of medium-thick plates according to claim 1, characterized in that, The step of comparing the statistical feature values ​​with the corresponding standard values ​​in the plate shape control knowledge base, and determining the cause of plate shape quality defects based on the comparison results, includes: Based on the statistical feature value, the deviation value between the statistical feature value and the corresponding standard value in the plate shape control knowledge base is determined; Based on the comparison results between the deviation value and the corresponding optimal range in the plate shape control knowledge base, the process parameters in the key parameter dataset that are in an abnormal state are determined. Based on the process parameters that are in an abnormal state, determine the cause of plate shape quality defects.

7. The method for analyzing the shape quality defects of medium-thick plates according to claim 1, characterized in that, The step of comparing the statistical feature values ​​with the corresponding standard values ​​in the plate shape control knowledge base, and determining the cause of the plate shape quality defects based on the comparison results, includes: An analysis report containing parameter deviation information is generated based on the causes of plate shape quality defects, and the display device is controlled to display the analysis report in the form of charts.

8. A system for analyzing the shape and quality defects of medium-thick plates, characterized in that, The system includes: a data analysis module, a feature extraction module, and a data comparison module; The data analysis module is configured to: acquire multiple process parameters and corresponding plate shape quality data during the production of medium and heavy plates; analyze each process parameter and corresponding plate shape quality data based on a correlation analysis algorithm to determine the degree of influence of the process parameter; The feature extraction module is configured to: construct a key parameter dataset based on the process parameters corresponding to the influence value when the influence value is greater than a preset correlation threshold; and extract features from each process parameter in the key parameter dataset to obtain statistical feature values. The data comparison module is configured to compare the statistical feature value with the corresponding standard value in the plate shape control knowledge base, and determine the cause of the plate shape quality defect based on the comparison result. The plate shape control knowledge base is constructed based on data that has been judged as qualified after historical statistical feature values ​​and corresponding plate shape quality test results in the historical production process.

9. An electronic device, comprising: The memory and processor are characterized in that the processor is used to execute a computer program stored in the memory to implement the steps of the method for analyzing the shape quality defects of a medium-thick plate as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for analyzing the shape quality defects of a medium-thick plate as described in any one of claims 1-7.

Images