Strip steel quality control method and device, electronic equipment and storage medium

Abstract

The invention provides a strip steel quality control method and device, electronic equipment and a storage medium, and belongs to the technical field of strip steel production control, the method comprises the following steps: obtaining strip steel bead defect data, and screening out a production parameter set corresponding to a defect period based on the defect data; extracting target parameter values corresponding to the corresponding feature parameters from the production parameter set based on a mapping relationship between the bead defect potential cause library and the feature parameters, and normalizing the target parameter values to obtain feature parameter vectors; determining a target cause based on the feature relevancy between the feature parameter vector and each potential cause in a bead defect potential cause library; and determining a target adjustment parameter and an adjustment direction corresponding to the target cause based on a mapping rule of the cause and the adjustment parameter, and adjusting the target adjustment parameter based on a first preset adjustment step length. According to the strip steel quality control method and device, the electronic equipment and the storage medium, the product degradation rate can be reduced, and the production efficiency can be improved.

Description

Technical Field

[0001] This application belongs to the field of strip steel production control technology, and more specifically, relates to a strip steel quality control method and device, electronic equipment, and storage medium. Background Technology

[0002] In cold-rolled strip steel production, protruding ridge defects are a key issue restricting product quality. Existing technologies generally collect defect information using laser displacement sensors or visual inspection systems, combine this with human experience to select process parameters from a production database for similar time periods, and after initially determining the cause of the defect based on historical cases, adjust core parameters such as rolling force and roll crown. Some solutions introduce simple data statistical methods to assist in analysis, but the overall process still revolves around "inspection-experience judgment-rough adjustment," relying heavily on the operator's skill level.

[0003] However, in existing technologies, there is often a time and space mismatch between defect data and production parameters. The lack of quantitative support for determining the cause makes it easy to misjudge. The parameter adjustment step size relies on experience, resulting in insufficient accuracy. This not only makes it difficult to completely eliminate convex edge defects, but may also cause secondary problems such as edge waviness, resulting in a high product downgrade rate and low production efficiency. Summary of the Invention

[0004] The purpose of this application is to provide a method and apparatus for strip steel quality control, electronic equipment, and storage medium to reduce product degradation rate and improve production efficiency.

[0005] A first aspect of this application provides a method for controlling the quality of strip steel, comprising: Acquire strip steel ridge defect data and filter out the production parameter set corresponding to the defect period based on the defect data; the defect period is the time period corresponding to when the strip steel ridge defect occurs. Based on the mapping relationship between the potential cause library of convex edge defects and the feature parameters, the target parameter values ​​corresponding to the feature parameters are extracted from the production parameter set, and the target parameter values ​​are normalized to obtain the feature parameter vector. The potential cause library of convex edge defects is preset based on the production process characteristics of cold-rolled strip steel, and includes multiple potential causes. Each type of potential cause corresponds to different feature parameters. The target cause is determined based on the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects. Based on the mapping rule between causes and adjustment parameters, the target adjustment parameters and adjustment direction corresponding to the target causes are determined, and the target adjustment parameters are adjusted based on the first preset adjustment step size.

[0006] A second aspect of this application provides a strip steel quality control device, comprising: The data filtering module is used to acquire strip steel ridge defect data and filter the production parameter set corresponding to the defect period based on the defect data; the defect period is the time period corresponding to when the strip steel ridge defect occurs. The vector acquisition module is used to extract the target parameter values ​​corresponding to the feature parameters from the production parameter set based on the mapping relationship between the potential cause library of convex edge defects and the feature parameters, and to normalize the target parameter values ​​to obtain the feature parameter vector. The potential cause library of convex edge defects is preset based on the production process characteristics of cold-rolled strip steel, and includes multiple potential causes. Each type of potential cause corresponds to different feature parameters. The cause determination module is used to determine the target cause based on the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects. The control module is used to determine the target adjustment parameters and adjustment direction corresponding to the target cause based on the mapping rules between the cause and the adjustment parameters, and to adjust the target adjustment parameters based on the first preset adjustment step size.

[0007] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the strip steel quality control method described above.

[0008] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the strip steel quality control method described above.

[0009] The beneficial effects of the strip steel quality control method, device, electronic equipment, and storage medium provided in this application are as follows: By acquiring convex defect data and associating it with production parameters during the defect period, this application solves the problem of spatiotemporal misalignment between traditional defects and process data, providing accurate data support for tracing the causes; by extracting and normalizing feature parameters based on a preset cause library, parameter redundancy and dimensional interference are eliminated, ensuring the scientific nature of feature correlation calculation; by locking the target cause through feature correlation, manual experience judgment is replaced, avoiding misjudgment of causes; by determining adjustment parameters and directions according to mapping rules, and combining with a first preset adjustment step size for precise control, secondary defects caused by blind adjustment are eliminated. Ultimately, rapid tracing and precise elimination of convex defects are achieved, improving the accuracy of strip steel quality control, reducing product downgrade rate, and increasing production efficiency. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 A schematic flowchart of a strip steel quality control method provided in an embodiment of this application; Figure 2 This is a structural block diagram of a strip steel quality control device provided in one embodiment of this application; Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0012] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0013] It is understood that in the embodiments of this application, data such as user information are involved. When the embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with relevant laws, regulations and standards.

[0014] It should be noted that the terms "first," "second," etc., used in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in sequences other than those illustrated or described herein.

[0015] To make the objectives, technical solutions, and advantages of this application clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0016] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a strip steel quality control method according to an embodiment of this application. The strip steel quality control method provided in this embodiment can be executed by an electronic device, and the method may include: S101: Obtain strip steel ridge defect data and filter out the production parameter set corresponding to the defect period based on the defect data; the defect period is the time period corresponding to when the strip steel has ridge defects.

[0017] In this embodiment, the strip ridge defect data refers to the relevant information of longitudinal strip-shaped raised defects appearing on the strip surface along the rolling direction. Data acquisition can be achieved using a CLD-DEF-2025 type ridge composite detection system. This system includes a laser displacement sensor array, an electrical signal amplification module, and a defect location module, installed 50mm from the strip surface at the cold rolling mill exit. The laser module emits a 650nm semiconductor laser and acquires the strip surface distance data using triangulation, with a measurement range of 0-100μm and a defect height accuracy of ±0.1μm. The electrical module… The block amplifies the analog signal differentially by 1000 times and converts it into a 24-bit digital signal; the positioning module, combined with the strip running speed collected by the encoder (accuracy ±0.1m / min), achieves a longitudinal positioning accuracy of ±5mm and a lateral positioning accuracy of ±2mm for defects; the defect data can specifically include: defect ID, steel coil batch number, defect time period (accurate to the second, such as 10:25:30-10:25:45), defect location (lateral coordinate X, longitudinal range Y1-Y2), defect height H (average of 3 consecutive sampling points), and defect length L=Y2-Y1.

[0018] The defect time period refers to the precise time interval (accurate to the second) during which the strip passes through the detection system when a protruding ridge defect appears on the surface of the strip. It is calculated by converting the timestamp of the detection system with the running speed of the strip and is used to associate the defect with the production parameters of the corresponding time period.

[0019] The production parameter set refers to the collection of process parameters covering the entire production process of cold-rolled strip steel. The production parameter set can be collected using a dual-link "industrial 5G + fiber optic" system, covering the entire process of hot rolling heating, descaling, finishing rolling, cold rolling, and annealing. Specific collection equipment and parameters include: infrared thermometers for collecting furnace temperature (accuracy ±1℃), pressure sensors for collecting descaling pressure (accuracy ±0.1MPa), displacement sensors for collecting roll crown (accuracy ±0.2μm), and piezoelectric pressure sensors for collecting rolling force (accuracy ±1kN), etc. The collection frequency is 50-300Hz, and all parameters are accompanied by a ±1ms timestamp. During screening, the parameter data within the defect period is extracted through a data association engine, and the timestamps are aligned at 1ms / item to form a production parameter set of "timestamp - actual value of 28 types of parameters".

[0020] In this embodiment, high-precision detection equipment captures protruding defects on the strip surface in real time, recording not only the physical characteristics of the defects (location, height, and length) but also simultaneously pinpointing the precise time period in which the defects occurred (based on the conversion between the detection system's timestamp and the strip's running speed). Subsequently, using this time period as a filtering criterion, all process parameters (such as rolling force, tension, and roll temperature) corresponding to that time period are extracted from a production parameter database covering the entire cold rolling process. The parameter data is then aligned according to the timestamp, ensuring that each defect corresponds to a unique production parameter sequence and avoiding tracing errors caused by spatiotemporal mismatches between parameters and defects.

[0021] S102: Based on the mapping relationship between the potential cause library of convex edge defects and the feature parameters, the target parameter values ​​corresponding to the feature parameters are extracted from the production parameter set, and the target parameter values ​​are normalized to obtain the feature parameter vector; the potential cause library of convex edge defects is preset based on the production process characteristics of cold-rolled strip steel, and the potential cause library of convex edge defects includes multiple potential causes; each type of potential cause corresponds to different feature parameters.

[0022] In this embodiment, the potential cause library for convex edge defects is a set of possible causes for convex edge defects based on the pre-set characteristics of the cold-rolled strip steel production process. It can include various causes such as uneven heating in hot rolling, excessive crown of the finishing rolls, and tension fluctuations in cold rolling, each with a clear process logic and impact mechanism. The feature parameter mapping relationship refers to the fixed correspondence between each cause in the potential cause library and specific production parameters (e.g., "cold rolling tension fluctuation" corresponds to parameters such as tension fluctuation amplitude and fluctuation duration), used to accurately extract key information related to the causes from a massive amount of production parameters. Feature parameters are production parameters directly related to the causes of convex edge defects, and are core parameters selected from the production parameter set through mapping relationships. Each potential cause corresponds to 1-5 feature parameters, for a total of 28 feature parameters. The target parameter value refers to the specific numerical value obtained after statistical processing of the feature parameters extracted from the production parameter set (e.g., the mean and fluctuation value of time-series feature parameters, and the real-time value of static feature parameters), which is quantitative data characterizing the state of the feature parameters during the defect period. The feature parameter vector refers to a one-dimensional data vector formed by arranging the 28 normalized feature parameters in a fixed order. This is used to centrally characterize the characteristic states of defect-related parameters, providing a unified data format for subsequent correlation calculations. Potential causes refer to each specific reason that may lead to a defect (such as insufficient hot-rolling descaling pressure, excessive strip wedge shape, etc.) included in the potential cause library of convex edge defects, serving as candidate objects for tracing the root cause.

[0023] In this embodiment, based on a pre-defined library of potential causes of convex edge defects, specific parameters corresponding to each type of cause (e.g., the convexity parameter of F5 / F6 stand corresponding to "excessive convexity of finishing work roll") are extracted from the selected set of production parameters through a fixed mapping relationship of "cause-feature parameter". The target parameter value is then obtained through statistical calculation (mean / fluctuation value for time-series parameters, real-time value for static parameters). Normalization then uses linear transformation to uniformly map target parameter values ​​of different dimensions and numerical ranges to the [0,1] interval, eliminating magnitude differences between parameters (e.g., temperature ℃ and tension MPa), ensuring the scientific validity and accuracy of subsequent feature correlation calculations, and ultimately forming vector data that comprehensively characterizes the features of defect-related parameters.

[0024] S103: Determine the target cause based on the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects.

[0025] In this embodiment, feature relevance refers to the quantified value of the matching degree between the feature parameter vector and each potential cause (range [0, 100]). The higher the relevance, the stronger the correlation between the cause and the current defect. The target cause refers to the core cause (usually 1) selected from the potential cause library that has the highest relevance to the current defect features and meets the preset threshold, and is the root cause of the defect.

[0026] In this embodiment, the feature parameter vector is analyzed using a pre-trained algorithm model (such as random forest or Bayesian model), and the correlation between the vector and the features of each potential cause is calculated. The higher the correlation, the more significant the anomaly of the feature parameter corresponding to that cause during the defect period. Subsequently, by setting a correlation threshold, irrelevant causes are eliminated, and finally, the cause that best matches the current defect (target cause) is selected, solving the problem of reliance on human experience and strong subjectivity in cause determination in traditional methods.

[0027] S104: Based on the mapping rule between cause and adjustment parameters, determine the target adjustment parameters and adjustment direction corresponding to the target cause, and adjust the target adjustment parameters based on the first preset adjustment step size.

[0028] In this embodiment, the mapping rule between the cause and the adjustment parameter refers to the fixed correspondence between the preset target cause and the process parameter to be adjusted (e.g., if the target cause is "excessive crown of the finishing mill work roll", the corresponding adjustment parameter is the crown of the work roll of the F5 / F6 stand), used to quickly locate the control object. The target adjustment parameter refers to the process parameter that directly corresponds to the target cause and needs to be adjusted to eliminate the defect (e.g., rolling force, tension, roll crown, etc.), and is the object of the control action. The adjustment direction refers to the changing trend of the target adjustment parameter (e.g., increase, decrease, or remain stable), determined by the influence mechanism of the target cause (e.g., if the cause is "excessive tension fluctuation", the adjustment direction is "suppress fluctuation and move closer to the target value"). The first preset adjustment step size refers to the single adjustment range of the target adjustment parameter (e.g., 3). The value of 0.5 MPa is a core indicator for ensuring the accuracy of adjustment. It is calculated based on the strip thickness (defect severity coefficient) and the degree of parameter anomaly (parameter anomaly coefficient).

[0029] In this embodiment, based on a preset mapping rule between causes and adjustment parameters, the process parameters (target adjustment parameters) and adjustment directions corresponding to the target cause are directly located (e.g., if the cause is "uneven distribution of rolling force," the adjustment direction is "reducing the stress in high-stress areas and increasing the stress in low-stress areas"). The first preset adjustment step size is quantitatively calculated based on the severity of the defect (defect height) and the degree of parameter anomaly (deviation between actual and target values), ensuring that the adjustment range is sufficient to eliminate the defect without causing new plate shape problems due to over-adjustment. Finally, the adjustment command is sent to the corresponding actuator (such as an electro-hydraulic proportional valve or a frequency converter) through the intelligent control system to complete the parameter adjustment, forming a closed-loop control of defect detection, cause tracing, and parameter adjustment.

[0030] As can be seen from the above, this embodiment solves the problem of spatiotemporal misalignment between traditional defects and process data by acquiring convex edge defect data and associating it with production parameters during the defect period, providing accurate data support for tracing the root cause. It extracts and normalizes feature parameters based on a preset cause library, eliminating parameter redundancy and dimensional interference, ensuring the scientific validity of feature correlation calculations. It locks down the target cause through feature correlation, replacing manual experience-based judgment and avoiding misjudgments. It determines adjustment parameters and directions according to mapping rules, and combines this with a first preset adjustment step size for precise control, preventing secondary defects caused by blind adjustments. Ultimately, it achieves rapid tracing and precise elimination of convex edge defects, improving the accuracy of strip steel quality control, reducing product downgrade rates, and increasing production efficiency.

[0031] In one embodiment of this application, determining the target cause based on the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects includes: Based on the feature parameter vector, the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects is calculated through a pre-trained random forest model, and an initial cause set with a feature correlation greater than a preset correlation is obtained; the initial cause set contains at least one initial cause. Calculate the prior probability and likelihood probability of each initial cause in the initial cause set; The posterior probability of each initial cause in the initial cause set is calculated based on the prior probability and the likelihood probability. The target cause is determined based on the posterior probability of each initial cause.

[0032] In this embodiment, the random forest model can consist of 100 CART decision trees, each with a depth ≤15. During training, 14 feature parameters are randomly selected for each tree to avoid overfitting. The training sample set is constructed based on 5 years of historical production data, containing 10,000 sets of 28-dimensional feature parameter vectors and labeled samples with actual cause labels. The labels are 12-dimensional vectors (1 for cause relevance, 0 otherwise). The model uses 5-fold cross-validation for training, achieving a final training accuracy ≥95%. During calculation, the normalized feature parameter vectors are input into the model. Each decision tree determines whether 12 types of causes are related (determination rule: feature matching degree ≥0.7 indicates relevance). The model votes on the determination results of the 100 trees, and the relevance score si = (number of decision trees that determine relevance / 100) × 100. The preset relevance threshold is set to 70. Causes with si ≥ 70 are selected to form the initial cause set. For example, if the relevance score of a defect is s3 = 92, s5 = 85 (the rest < 70), then the initial cause set is... ={ (Excessive crown of the finishing mill work rolls) (Cold rolling tension fluctuation)}. In practical applications, the number of decision trees in the model can also be selected as 120 or other values, but this application embodiment does not limit this.

[0033] In this embodiment, the preset relevance refers to a threshold (usually set to 70) used to screen initial causes. This threshold is determined by historical data statistics, and potential causes with a feature relevance higher than this threshold are considered relevant to the current defect. The initial cause set refers to the set of potential causes whose feature relevance is greater than the preset relevance after calculation using the random forest model. This set contains at least one cause for further screening of target causes. Each specific cause in the initial cause set is a candidate for a target cause. The prior probability is the probability that a potential cause actually occurs in all historical cases of convex defects. It is derived from historical defect data statistics and reflects the inherent frequency of occurrence of that cause. The likelihood probability is the probability of observing the current feature parameter vector when assuming a certain initial cause is true. It is calculated using the normal distribution probability density function combined with historical feature parameter data. The posterior probability is the probability that a certain initial cause is the true cause of the current defect, calculated using Bayes' theorem by combining the prior probability and the likelihood probability. It is the core quantitative indicator for determining the target cause.

[0034] In this embodiment, the causes related to the defect are initially screened out through the voting mechanism of multiple decision trees in the random forest model, and irrelevant causes are excluded to simplify subsequent analysis. Then, the prior probability (the inherent frequency of occurrence of the cause) and the likelihood probability (the degree of matching between the cause and the current defect data) are used to fuse the two into the posterior probability through Bayes' formula, so as to realize the quantitative assessment of the correlation between the cause and the defect. Finally, based on the posterior probability threshold and defect characteristics, the target cause that plays a dominant role in the defect is accurately identified.

[0035] As can be seen from the above, the initial cause set screening in this embodiment reduces 12 potential causes to 2-3, significantly reducing the subsequent computational load and improving the efficiency of tracing the source; the combination of prior probability and likelihood probability avoids the bias of single factor judgment, and the quantification of posterior probability makes the determination of the target cause more convincing, which greatly improves the accuracy compared with traditional manual judgment.

[0036] In one embodiment of this application, the posterior probability of each initial cause in the initial cause set is calculated based on the prior probability and the likelihood probability, including: Based on the prior probability and the likelihood probability, the posterior probability of each initial cause in the initial cause set is calculated using the first formula. The first formula is: ; in, This represents the posterior probability of each initial cause. This represents the likelihood probability of each initial cause. Indicates the i-th potential cause The actual probability of occurrence in all historical cases of protruding edge defects. This represents the i-th potential cause in the potential cause library of convex edge defects. Indicates the initial cause The corresponding likelihood probability, Represents the initial causal set The j-th initial cause The actual probability of occurrence in all historical cases of protruding ridge defects in cold-rolled strip steel. Represents the initial causal set The j-th initial cause in the equation.

[0037] In this embodiment, prior probability The probability of the i-th potential cause occurring in all historical cases of convex defects is calculated using statistics from a strip quality database. The statistical range includes 12,000 cases of convex defects from the past 5 years, excluding their causes. If P( is a true cause in 3000 of these cases, then P( ) = 3000 / 12000 = 0.25; Cause If P( ) is a true cause in 2400 cases, then P( =2400 / 12000=0.2. Defect cases caused by non-process factors such as sensor malfunctions are excluded during the statistical analysis to ensure the accuracy of the probability.

[0038] Likelihood probability For "given cause" "The probability of observing the current feature parameter vector D" is calculated using the normal distribution probability density function, and the formula is: Corresponding feature parameters [ ], where G is the normal distribution function, , Causes The historical mean and variance of the corresponding k-th feature parameter (based on statistics from 10,000 training samples). For example, the causes... For the corresponding feature parameters F5 and F6, the current normalized value of F5 is x5=0.95 and the normalized value of F6 is x6=0.92. Statistically, we obtain... =0.9、 =0.0025, =0.88、 =0.0016, then:

[0039] Posterior probability: Substitute the prior probability and likelihood probability into the formula, such as for... and The initial causal set, denominator ;

[0040] In this embodiment, the first formula applies only to the initial causal set. The causal calculation in this process involves evaluating the probability of relevant causes selected by the random forest model, excluding interference from irrelevant causes. Before calculation, it is necessary to ensure that the statistical standards of all parameters are consistent. For example, the statistical cases of prior probability should include strip steel cases of the same steel type and specification as the current defect, and the mean and variance of likelihood probability should be based on the historical characteristic parameter data of the corresponding cause to avoid errors caused by mixing data across different scenarios.

[0041] The calculation tools and precision control can be implemented using Python's NumPy library for formula calculation. The normal distribution function is called through the `scipy.stats.norm` module, and the calculation precision is retained to three decimal places. For the extreme case where the denominator is 0 (which is almost non-existent in practical applications because the initial cause set contains at least one cause), a default value of 0.0001 is set to avoid program errors and trigger a manual review process.

[0042] In this embodiment, the first formula is based on Bayes' theorem. It takes the prior probability (reflecting the historical occurrence pattern of the cause) and the likelihood probability (reflecting the degree of matching between the cause and the current defect) as inputs. Through the normalization of the denominator, the sum of the posterior probabilities of all initial causes is made to be 1, ensuring that the probability calculation conforms to mathematical logic. The final output posterior probability quantifies the confidence level of "under the current defect data, a certain cause is the true cause", providing a quantitative basis for determining the target cause.

[0043] As can be seen from the above, the explicit formula in this embodiment standardizes the posterior probability calculation process, and different technicians can obtain consistent results by calculating according to the formula, thereby improving the reproducibility of the technical solution; the normalized denominator ensures the rationality of the probability result and avoids the problem of probability values ​​exceeding the range of [0,1].

[0044] In one embodiment of this application, the method for determining the first preset adjustment step size includes: The severity coefficient of defects is determined based on the thickness of the strip, and the anomaly coefficient of parameters is determined based on the type of characteristic parameters. The first preset adjustment step size is determined based on the product of the defect severity and the defect severity coefficient of the strip steel and the product of the parameter anomaly degree and the parameter anomaly degree coefficient.

[0045] In this embodiment, the defect severity coefficient is a weighted coefficient determined based on the strip thickness (range 0.02-0.04). The greater the strip thickness, the larger the coefficient, reflecting the difference in the difficulty of eliminating protruding defects in thick strips. The parameter anomaly coefficient is a weighted coefficient determined based on the type of characteristic parameter (range 0.05-0.1). Different types of parameters (such as rolling force and temperature) correspond to different coefficients, reflecting the differences in parameter control sensitivity. Defect severity refers to the severity of the defect characterized by the height of the protruding defect, determined by the average of three consecutive sampling points in the defect area collected by the detection system. Parameter anomaly refers to the deviation between the actual value and the target value of the characteristic parameter, characterizing the abnormal state of the parameter.

[0046] In this embodiment, the defect severity coefficient is positively correlated with the strip thickness. Since the protruding defects of thick strip are more difficult to eliminate, a larger adjustment range is required. The specific correspondence is shown in Table 1 below: Table 1. Comparison of Defect Severity Coefficients

[0047] For example, for a strip thickness of 1.0 mm, k1 = 0.03; for a thickness of 2.0 mm, k1 = 0.04. This coefficient was obtained through statistical analysis of 1000 sets of adjustment tests on strips of different thicknesses to ensure that the coefficient matches the adjustment requirements.

[0048] The parameter anomaly coefficient is set according to the control sensitivity of the characteristic parameter. The coefficient is larger for parameters with high control sensitivity (such as roll crown) to avoid small adjustments being ineffective. The specific correspondence is shown in Table 2 below: Table 2. Comparison Table of Parameter Abnormality Coefficient k2

[0049] For example, if the target adjustment parameter is the crown of the work roll of the F5 frame (belonging to the roll crown category), the corresponding k2=0.1; if it is the descaling pressure (pressure / flow category), the corresponding k2=0.06.

[0050] The severity of a defect is represented by its height H, which is collected using a CLD-DEF-2025 detection system. The average value is taken from three consecutive sampling points in the defect area. For example, if the defect height H = 6.2... The degree of parameter anomaly is measured by the parameter deviation. The calculation formula is as follows: For example, the actual value of the F5 frame convexity is 230. If the target value is 200μm, then .

[0051] Calculate using the addition formula. For example: Strip thickness 1.0mm (k1=0.03), defect height H=6.2mm. The target adjustment parameter is the F5 frame convexity (k2=0.1), and the parameter deviation is... ,but: ; Rounded to 3 (Saving the constraint that the single adjustment step size is ≤ 5% of the target parameter value, 200μm × 5% = 10) ).

[0052] In this embodiment, the defect severity coefficient is determined by the strip thickness to reflect the difference in the difficulty of defect elimination for strips of different specifications; the parameter anomaly coefficient is determined by the characteristic parameter type to reflect the difference in the control sensitivity of different parameters; the defect height (defect severity) and parameter deviation (parameter anomaly) are weighted and summed to obtain a quantitative adjustment step size that takes into account both defect characteristics and parameter status, thus avoiding the limitation of a single factor determining the step size.

[0053] As can be seen from the above, the quantitative calculation of the first preset adjustment step size in this embodiment frees the adjustment process from dependence on manual experience, improving the adjustment accuracy to [percentage missing]. Level; step size constraints ensure that adjustments will not exceed the equipment's safety range, significantly improving upon traditional experience-based adjustments; strong adaptability to different steel types and specifications, eliminating the need for separate step size adjustments for each scenario.

[0054] In one embodiment of this application, it further includes: Filter the set of related parameters that are coupled with the target adjustment parameters; Calculate the correlation coefficient between the target adjustment parameter and each associated parameter in the associated parameter set; The associated adjustment parameters are determined based on the correlation coefficient and the preset correlation coefficient threshold, and then adjusted based on the second preset adjustment step size.

[0055] In this embodiment, coupling relationship refers to the mutual influence between the target adjustment parameter and other process parameters (e.g., adjusting the rolling force directly affects the tension). This relationship can cause the adjustment of a single parameter to trigger abnormalities in other parameters, thus generating secondary defects. The associated parameter set refers to the set of process parameters that have a coupling relationship with the target adjustment parameter, determined through a three-step method: process screening, physical coupling screening, and historical data change rate screening. The correlation coefficient is a quantitative value (range [-1, 1]) calculated using the Pearson algorithm to represent the coupling strength between the target adjustment parameter and the associated parameters; the larger the absolute value, the stronger the coupling. The preset correlation coefficient threshold is a threshold used to determine the type of associated parameter (usually set to a strong correlation threshold of 0.7 and a medium correlation threshold of 0.3), calibrated based on process optimization experience and historical adjustment data. Associated adjustment parameters are associated parameters whose absolute correlation coefficient with the target adjustment parameter is ≥ the preset strong correlation threshold (0.7). These parameters need to be adjusted synchronously with the target adjustment parameter to avoid secondary defects. The second preset adjustment step size refers to the single adjustment amplitude of the associated adjustment parameter, calculated based on the first preset adjustment step size of the target adjustment parameter, the correlation coefficient, and other factors, ensuring coordinated optimization with the target adjustment parameter.

[0056] In this embodiment, the relevant parameters closely related to the target adjustment parameters are accurately located through three steps: process attribution, physical coupling relationship, and historical data change rate. The Pearson correlation coefficient is used to quantify the coupling strength between the two to avoid the bias of subjective judgment. The relevant parameters are classified based on the correlation coefficient threshold, and only the strongly correlated parameters are adjusted synchronously to ensure the coordination of adjustment and avoid over-adjustment that increases system complexity. Finally, the synchronous optimization of the target parameter and the associated parameters is achieved, eliminating the side effects of adjusting a single parameter.

[0057] As can be seen from the above, the determination of the associated adjustment parameters in this embodiment upgrades parameter adjustment from single adjustment to coordinated adjustment, and the occurrence rate of secondary defects decreases. The quantitative calculation of the correlation coefficient makes the determination of the associated parameters more scientific. The setting of observation parameters realizes the hierarchical management of key control and secondary monitoring, taking into account both the adjustment effect and system efficiency.

[0058] In one embodiment of this application, it further includes: Based on the importance level of the associated adjustment parameters, determine the associated adjustment coefficients corresponding to the associated adjustment parameters; The second preset adjustment step size is determined based on the correlation coefficient and associated adjustment coefficient between the associated adjustment parameter and the target adjustment parameter, and the first preset adjustment step size.

[0059] In this embodiment, the importance level refers to the level based on the degree of influence of the associated adjustment parameters on the protruding edge defect (divided into two categories: core process parameters and auxiliary process parameters). Core process parameters have a greater impact on the defect and have a higher adjustment priority. The associated adjustment coefficient refers to the weight coefficient determined based on the importance level of the associated adjustment parameters (with a value range of 0.5-1.0). The coefficient corresponding to core process parameters is larger, reflecting the adjustment priority of associated parameters.

[0060] In this embodiment, the importance level of the associated adjustment parameters is classified based on their impact on strip protrusion defects. These parameters are determined by process engineers using cold rolling production experience and historical data, and are divided into two categories: core process parameters and auxiliary process parameters. The specific correspondence is shown in Table 3 below: Table 3. Comparison table of correlation adjustment coefficient k3

[0061] For example, if the associated adjustment parameter is the cold rolling exit tension (a core process parameter), k3 = 0.9; if it is the lubricating grease temperature (an auxiliary process parameter), k3 = 0.6. This coefficient can be finely adjusted according to the characteristics of the steel grade. For example, the core parameter k3 for high-strength steel (DP590) is set to the upper limit of 1.0, while that for ordinary steel (SPHC) is set to 0.8.

[0062] In this embodiment, the adjustment coefficients are classified according to importance level to clarify the adjustment priority of different associated parameters and ensure that the adjustment range of core parameters meets the coordination requirements. The correlation coefficient (coupling strength), the associated adjustment coefficient (parameter importance) and the first preset adjustment step size (baseline range) are combined to construct the second preset adjustment step size calculation logic of multiple factors, so that the adjustment range of associated parameters is both adapted to the target parameters and in line with their own process characteristics.

[0063] As can be seen from the above, the introduction of the correlation adjustment coefficient in this embodiment makes the adjustment range of the associated parameters more in line with the parameter weights in actual production, and the synergistic adjustment effect of the core parameters is greatly improved; the multi-factor calculation logic of the second preset adjustment step avoids the mechanicalness of adjusting according to a fixed ratio and adapts to associated parameters with different coupling strengths and different importance.

[0064] In one embodiment of this application, determining a second preset adjustment step size based on the correlation coefficient between the associated adjustment parameter and the target adjustment parameter, the correlation adjustment coefficient, and a first preset adjustment step size includes: The second preset adjustment step size is determined based on the second formula; The second formula is:

[0065] in, This indicates the second preset adjustment step size. This indicates the first preset adjustment step size. This represents the correlation coefficient between the associated adjustment parameter and the target adjustment parameter. Indicates the correlation adjustment factor. This represents the correction factor. This represents the deviation suppression coefficient. This indicates the deviation rate between the actual value and the target value of the associated adjustment parameter.

[0066] In this embodiment, the correction coefficient This refers to a coefficient (range 0.9-1.1) determined based on the process type and associated adjustment parameters. The upper limit is used for rolling force / tension parameters, and the lower limit for temperature / flow rate parameters, to adapt to the control sensitivity of different parameters. Deviation suppression coefficient. This refers to a coefficient (range 0.01-0.05) used to prevent excessive adjustment of the linkage parameters. Based on the calibration of the cold rolling mill's control accuracy, the larger the deviation of the linkage parameters, the stronger the inhibitory effect of this coefficient on the adjustment step size. The deviation rate between the actual value and the target value of the linkage adjustment parameter. This refers to the deviation ratio between the actual value and the target value of the associated adjustment parameter, calculated using the following formula: It is used to characterize the degree of abnormality of the associated parameters themselves.

[0067] For example, the target adjustment parameter is F5 rack convexity ( Taking the cold rolling exit tension as an example, the parameter values ​​are as follows: =0.78, k3=0.9, k4=1.05, =0.04, =6.67%, then:

[0068] Formula application constraints: 1) It must not exceed 5% of the target value of the associated parameters. For example, if the target value of the cold rolling exit tension is 150 MPa, 5% is 7.5 MPa. The above calculation result of 1.8 MPa meets the constraint; 2) If ≥20% (with severely abnormal associated parameters), then Take the upper limit of 3% of the target value and prioritize correcting the abnormality of the associated parameters themselves; 3) The cumulative adjustment amount shall not exceed 15% of the target value of the associated parameters to avoid parameter instability caused by long-term adjustment.

[0069] In this embodiment, the second formula is... Based on, through k3 and k4 respectively quantify the coupling strength and parameter importance, ensuring that the adjustment range of associated parameters is coordinated with the target parameters; k4 adapts to the process type of associated parameters, improving the adaptability of the step size; in the denominator... The term implements the "deviation suppression" function—when the associated parameter itself deviates from the target value, the denominator increases. Automatic reduction avoids exacerbating anomalies during adjustments; the final output is a scientifically adjusted step size that balances collaborative needs and its own state.

[0070] As can be seen from the above, the application of the second formula in this embodiment standardizes the calculation of the adjustment step size of the associated parameters and greatly improves the adjustment accuracy; the deviation suppression mechanism effectively reduces the risk of overshoot when the associated parameters are severely abnormal, and avoids the problem of new abnormalities caused by the coordinated adjustment; combined with the application of constraints, it ensures that the adjustment process is carried out within the safe range of the equipment, and the coordinated adjustment of the associated parameters and the target parameters greatly improves the elimination rate of strip steel protrusion defects.

[0071] Based on the same inventive concept, this application also provides a strip steel quality control device for implementing the strip steel quality control method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the strip steel quality control device provided below can be found in the limitations of the strip steel quality control method described above, and will not be repeated here.

[0072] This application provides a strip steel quality control device, such as... Figure 2 As shown, the strip steel quality control device 20 includes: a data filtering module 21, a vector acquisition module 22, a cause determination module 23, and a control module 24.

[0073] The data filtering module 21 is used to acquire strip steel protrusion defect data and filter the production parameter set corresponding to the defect period based on the defect data; the defect period is the period when the strip steel has protrusion defects. The vector acquisition module 22 is used to extract the target parameter values ​​corresponding to the feature parameters from the production parameter set based on the mapping relationship between the potential cause library of convex edge defects and the feature parameters, and to normalize the target parameter values ​​to obtain the feature parameter vector; the potential cause library of convex edge defects is preset based on the production process characteristics of cold-rolled strip steel, and includes multiple potential causes; each type of potential cause corresponds to different feature parameters; The cause determination module 23 is used to determine the target cause based on the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects; The control module 24 is used to determine the target adjustment parameters and adjustment direction corresponding to the target cause based on the mapping rules between the cause and the adjustment parameters, and to adjust the target adjustment parameters based on the first preset adjustment step size.

[0074] In one embodiment of this application, the cause determination module 23 is specifically used for: Based on the feature parameter vector, the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects is calculated through a pre-trained random forest model, and an initial cause set with a feature correlation greater than a preset correlation is obtained; the initial cause set contains at least one initial cause. Calculate the prior probability and likelihood probability of each initial cause in the initial cause set; The posterior probability of each initial cause in the initial cause set is calculated based on the prior probability and the likelihood probability. The target cause is determined based on the posterior probability of each initial cause.

[0075] In one embodiment of this application, the cause determination module 23 is further configured to: Based on the prior probability and the likelihood probability, the posterior probability of each initial cause in the initial cause set is calculated using the first formula. The first formula is: ; in, This represents the posterior probability of each initial cause. This represents the likelihood probability of each initial cause. Indicates the i-th potential cause The actual probability of occurrence in all historical cases of protruding edge defects. This represents the i-th potential cause in the potential cause library of convex edge defects. Indicates the initial cause The corresponding likelihood probability, Represents the initial causal set The j-th initial cause The actual probability of occurrence in all historical cases of protruding ridge defects in cold-rolled strip steel. Represents the initial causal set The j-th initial cause in the equation.

[0076] In one embodiment of this application, the strip quality control device 20 further includes: a first-step length determination module, specifically used for: The severity coefficient of defects is determined based on the thickness of the strip, and the anomaly coefficient of parameters is determined based on the type of characteristic parameters. The first preset adjustment step size is determined based on the product of the defect severity and the defect severity coefficient of the strip steel and the product of the parameter anomaly degree and the parameter anomaly degree coefficient.

[0077] In one embodiment of this application, the strip quality control device 20 further includes: a strip adjustment module, specifically used for: Filter the set of related parameters that are coupled with the target adjustment parameters; Calculate the correlation coefficient between the target adjustment parameter and each associated parameter in the associated parameter set; The associated adjustment parameters are determined based on the correlation coefficient and the preset correlation coefficient threshold, and then adjusted based on the second preset adjustment step size.

[0078] In one embodiment of this application, the strip quality control device 20 further includes: a first-step length determination module, specifically used for: Based on the importance level of the associated adjustment parameters, determine the associated adjustment coefficients corresponding to the associated adjustment parameters; The second preset adjustment step size is determined based on the correlation coefficient and associated adjustment coefficient between the associated adjustment parameter and the target adjustment parameter, and the first preset adjustment step size.

[0079] In one embodiment of this application, the first step length determination module is further used for: The second preset adjustment step size is determined based on the second formula; The second formula is:

[0080] in, This indicates the second preset adjustment step size. This indicates the first preset adjustment step size. This represents the correlation coefficient between the associated adjustment parameter and the target adjustment parameter. Indicates the correlation adjustment factor. This represents the correction factor. This represents the deviation suppression coefficient. This indicates the deviation rate between the actual value and the target value of the associated adjustment parameter.

[0081] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3 The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of each module / unit in the above-described device embodiments, for example... Figure 2 The functions of the data filtering module 21, vector acquisition module 22, cause determination module 23, and control module 24 are shown.

[0082] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0083] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0084] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store information such as production parameter sets.

[0085] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation method described in the strip steel quality control method provided in the embodiments of this application, or they can execute the implementation method of the electronic device described in the embodiments of this application, which will not be repeated here.

[0086] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0087] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., provided on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0088] Those skilled in the art will recognize that the modules / units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0089] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0090] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or modules / units, or it may be an electrical, mechanical, or other form of connection.

[0091] The modules / units described as separate components may or may not be physically separate. Similarly, the components shown as modules / units may or may not be physical modules / units; they may be located in one place or distributed across multiple network modules / units. Some or all of the modules / units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0092] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0093] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for quality control of strip steel, characterized in that, include: Acquire strip steel protrusion defect data, and filter out the production parameter set corresponding to the defect period based on the defect data; The defect period is the time period corresponding to when a protruding edge defect appears in the strip steel; Based on the mapping relationship between the potential cause library of convex edge defects and the feature parameters, the target parameter values ​​corresponding to the feature parameters are extracted from the production parameter set, and the target parameter values ​​are normalized to obtain the feature parameter vector. The potential cause library of convex edge defects is preset based on the characteristics of cold-rolled strip steel production process, and includes multiple potential causes. Each type of potential cause corresponds to different feature parameters. Based on the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects, the target cause is determined; Based on the mapping rule between the cause and the adjustment parameter, the target adjustment parameter and adjustment direction corresponding to the target cause are determined, and the target adjustment parameter is adjusted based on the first preset adjustment step size.

2. The strip steel quality control method as described in claim 1, characterized in that, The step of determining the target cause based on the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects includes: Based on the feature parameter vector, and by calculating the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects using a pre-trained random forest model, an initial cause set with the feature correlation greater than a preset correlation is obtained; the initial cause set contains at least one initial cause. Calculate the prior probability and likelihood probability of each initial cause in the initial cause set; The posterior probability of each initial cause in the initial cause set is calculated based on the prior probability and the likelihood probability. The target cause is determined based on the posterior probability of each initial cause.

3. The strip steel quality control method as described in claim 2, characterized in that, The calculation of the posterior probability of each initial cause in the initial cause set based on the prior probability and the likelihood probability includes: Based on the prior probability and the likelihood probability, the posterior probability of each initial cause in the initial cause set is calculated using the first formula. The first formula is: ； in, This represents the posterior probability of each initial cause. This represents the likelihood probability of each initial cause. Indicates the i-th potential cause The actual probability of occurrence in all historical cases of protruding edge defects. This represents the i-th potential cause in the potential cause library of convex edge defects. Indicates the initial cause The corresponding likelihood probability, Represents the initial causal set The j-th initial cause The actual probability of occurrence in all historical cases of protruding ridge defects in cold-rolled strip steel. Represents the initial causal set The j-th initial cause in the equation.

4. The strip steel quality control method as described in claim 1, characterized in that, The methods for determining the first preset adjustment step size include: The defect severity coefficient is determined based on the strip thickness, and the parameter anomaly coefficient is determined based on the type of the characteristic parameter. The first preset adjustment step size is determined based on the product of the defect severity of the strip steel and the defect severity coefficient, and the product of the parameter anomaly degree and the parameter anomaly degree coefficient.

5. The strip steel quality control method as described in claim 4, characterized in that, Also includes: Filter the set of related parameters that are coupled with the target adjustment parameters; Calculate the correlation coefficient between the target adjustment parameter and each associated parameter in the associated parameter set; The associated adjustment parameters are determined based on the correlation coefficient and the preset correlation coefficient threshold, and the associated adjustment parameters are adjusted based on the second preset adjustment step size.

6. The strip steel quality control method as described in claim 5, characterized in that, Also includes: Based on the importance level of the associated adjustment parameter, determine the associated adjustment coefficient corresponding to the associated adjustment parameter; The second preset adjustment step size is determined based on the correlation coefficient between the associated adjustment parameter and the target adjustment parameter, the associated adjustment coefficient, and the first preset adjustment step size.

7. The strip steel quality control method as described in claim 6, characterized in that, The step of determining the second preset adjustment step size based on the correlation coefficient between the associated adjustment parameter and the target adjustment parameter, the associated adjustment coefficient, and the first preset adjustment step size includes: The second preset adjustment step size is determined based on the second formula; The second formula is: in, This indicates the second preset adjustment step size. This indicates the first preset adjustment step size. This represents the correlation coefficient between the associated adjustment parameter and the target adjustment parameter. Indicates the correlation adjustment factor. This represents the correction factor. This represents the deviation suppression coefficient. This indicates the deviation rate between the actual value and the target value of the associated adjustment parameter.

8. A strip steel quality control device, characterized in that, include: The data filtering module is used to acquire strip steel protrusion defect data and filter out the production parameter set corresponding to the defect period based on the defect data; The defect period is the time period corresponding to when a protruding edge defect appears in the strip steel; The vector acquisition module is used to extract the target parameter values ​​corresponding to the feature parameters from the production parameter set based on the mapping relationship between the potential cause library of convex edge defects and the feature parameters, and to normalize the target parameter values ​​to obtain the feature parameter vector; the potential cause library of convex edge defects is preset based on the production process characteristics of cold-rolled strip steel, and includes multiple potential causes; each type of potential cause corresponds to different feature parameters; The cause determination module is used to determine the target cause based on the feature correlation between the feature parameter vector and each potential cause in the potential cause library of convex edge defects; The control module is used to determine the target adjustment parameters and adjustment direction corresponding to the target cause based on the mapping rules between the cause and the adjustment parameters, and to adjust the target adjustment parameters based on a first preset adjustment step size.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

Images