A digital characterization method for the inspection results of strip surface defects

Abstract

The present invention discloses a digital characterization method for the inspection results of strip surface defects, which is as follows: Read the inspection results of cold-rolled plate defects from a strip surface defect inspector, and construct a mapping relationship between the inspection results of strip surface defects of cold-rolled plates and basic material data; Perform digital characterization on the detected inspection results of cold-rolled plate surface defects, that is, calculate the defect index of the cold-rolled plate and the distribution characteristics of the defects along the width direction of the strip; Trace the associated process parameter sequence corresponding to the digital defect inspection results of the current cold-rolled plate, and the associated process parameter sequence is arranged according to the process flow. It realizes the calculation of the distribution characteristics of a certain type of defect on the strip surface in the plate width direction and the defect index within any selected area, laying a foundation for studying and analyzing the occurrence law of strip surface defects, associating with the key process parameters in the whole process of steelmaking - refining - continuous casting - hot rolling - cold rolling, analyzing the causes of defects and improving the quality traceability of the whole process.

Description

Technical Field

[0001] The present invention belongs to the field of quality management and information technology. More specifically, the present invention relates to a digital characterization method for the inspection results of strip surface defects. Background Art

[0002] The strip production plants of major steel companies are generally equipped with strip surface defect inspection instruments to online inspect the surface defects of strips, and conduct macroscopic position positioning (positions in the length and width directions of the steel coil) of the defects on each coil of strip, take pictures of the defect pictures, calculate the length, width or area of the defects, classify and count according to the defect types, and save them in units of coils.

[0003] Process technicians analyze and study the inspection results of the surface defect inspection instrument, and establish defect maps, defect determination rules and surface quality grading determination rules. Quality supervisors make final quality determinations according to the defect maps and surface quality grading determination rules, and decide on the blocking, downgrading or release of the steel coil, reducing the quality losses caused by the outflow of unqualified steel coils.

[0004] At present, the detection data of strip surface defect inspection instruments used for surface defect detection of cold-rolled sheets and hot-rolled sheets in each steel mill mostly exist in the form of isolated islands, which can only be used for defect detection and determination of the surface quality of current cold-rolled or hot-rolled strip steel, and there is a problem of low utilization rate of the detection data of strip surface defect inspection instruments. Summary of the Invention

[0005] The present invention provides a digital characterization method for the inspection results of strip surface defects, aiming to improve the above problems.

[0006] The present invention is implemented as follows. A digital characterization method for the inspection results of strip surface defects, the method includes the following steps:

[0007] S1. Read the inspection results of cold-rolled sheet defects from a strip surface defect inspection instrument, and construct a mapping relationship between the inspection results of strip surface defects of cold-rolled sheet steel and basic material data;

[0008] S2. Digitally characterize the inspection results of strip surface defects of the detected cold-rolled sheet steel, including: calculating the defect index of the defects and the distribution along the width direction of the strip;

[0009] S3. Automatically track the digital characterization results of the defects of the current cold-rolled sheet and the corresponding associated process parameter sequences, and the associated process parameter sequences are arranged according to the process flow.

[0010] Further, preprocess the inspection results of cold-rolled sheet defects read in step S1, specifically including the following steps:

[0011] Eliminate the abnormal table inspection data from the read defect data;

[0012] Normalize the table inspection data of multiple strip surface defect detectors for different units, including: unifying the units and defect names;

[0013] Take the length direction of the long side of the strip on the OS operation side as the longitudinal coordinate axis Y, the width direction of the wide side of the strip head as the transverse coordinate axis X, and the intersection point of the long and wide sides of the strip on the OS operation side as the origin to determine the positions of each defect on the strip.

[0014] Furthermore, the construction process of the mapping relationship between the surface defect inspection results of cold-rolled strip steel and the basic material data is specifically as follows:

[0015] Take the export coil number of the previous unit equal to the import coil number of the next unit, and concatenate the basic material data according to the technological process; construct the mapping relationship between the concatenated basic material data and the surface defect detection results of cold-rolled strip steel based on the grade associated with the basic material data.

[0016] Furthermore, the method for digital characterization of defects on cold-rolled plates is specifically as follows:

[0017] Count the defects whose coordinates are located in the area of the current rectangular sliding window, read the defect type, the length and width of the defect, and traverse the statistical area on the surface of the cold-rolled plate through the rectangular sliding window to finally obtain all the defects on the statistical area of the cold-rolled plate;

[0018] The statistical area is the entire cold-rolled plate or a partially specified area of the cold-rolled plate.

[0019] Furthermore, define and calculate the defect index using the number of defects per unit area and the area of defects per unit area, and use the defect index to characterize the severity of the defects on the strip surface. Furthermore, the defect index I of defect k k , and its calculation formula is specifically as follows:

[0020]

[0021] In the formula, represents the number of defect k in the area of the i-th rectangular sliding window; represents the length of the j-th defect k in the area of the i-th rectangular sliding window; represents the width of the j-th defect k in the area of the i-th rectangular sliding window; S represents the area of the rectangular window; N represents the number of rectangular sliding windows in the statistical area.

[0022] Furthermore, the method for obtaining the distribution of defects along the width direction of the strip is specifically as follows:

[0023] A sliding window is formed based on a specified step size. The sliding window slides along the width direction of the strip steel, dividing the strip steel into several statistical regions along the width direction.

[0024] The defect index of each statistical region is calculated, that is, the distribution of defects along the width direction of the strip steel is obtained.

[0025] Furthermore, after step S3, it further includes:

[0026] S4. Based on the defect index of the strip steel, the associated process parameters are adjusted to optimize the key process parameters.

[0027] Furthermore, the adjustment method of the associated process parameters is specifically as follows:

[0028] (1) Analyze the correlation between the defect index and the key process parameters;

[0029] (2) When the correlation between the strip steel defect index and the key process parameter is strong, and the current value of this key process parameter is not sufficient to reduce the defect index, then narrow the process control window;

[0030] When the correlation between the strip steel defect index and the key process parameter is weak, then widen the process control window;

[0031] When the correlation between the strip steel defect index and the key process parameter is very weak, then cancel this key process parameter or do not include it in the cleanliness judgment rule.

[0032] The present invention automatically counts the surface defect situation of the strip steel and quickly performs digital characterization, significantly improving the statistical analysis efficiency (the time-consuming for completing the statistical analysis of one roll is increased from 20 - 30 minutes per roll to only a few seconds per roll). It can realize the graphical representation of the distribution characteristics of a certain type of defect on the strip steel surface in the plate width direction and / or the calculation of the defect index in any selected area, laying a foundation for process technicians to study and analyze the occurrence law of strip steel surface defects, correlate with the key process parameters in the whole process of steelmaking - refining - continuous casting - hot rolling - cold rolling production process, analyze the cause of defects and improve the whole process quality traceability. It improves the analysis efficiency and timeliness, meeting the needs of timely and rapid process analysis, trend management and quality consistent management. Description of the Drawings

[0033] Figure 1 It is a flow chart of the digital characterization method for the inspection results of strip steel surface defects provided by the embodiment of the present invention;

[0034] Figure 2 It is a schematic diagram of the change of the defect index of the steel quality defects of the whole roll of strip steel after multiple slabs are rolled in the continuous annealing unit provided by the embodiment of the present invention;

[0035] Figure 3The digital data display interface of the defects of each coil of strip steel provided by the embodiments of the present invention;

[0036] Figure 4 The schematic diagram of the distribution of defects along the width direction of the strip steel provided by the embodiments of the present invention;

[0037] Figure 5 The schematic diagram of the change trend of steel defects provided by the embodiments of the present invention. Specific embodiments

[0038] The following is a further detailed description of the specific embodiments of the present invention by referring to the accompanying drawings and describing the embodiments, so as to help those skilled in the art have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention.

[0039] Figure 1 The flowchart of the digital characterization method for the inspection results of strip steel surface defects provided by the embodiments of the present invention, and the method specifically includes the following steps:

[0040] S1. Read the inspection results of cold-rolled strip steel defects from the strip steel surface defect inspector, and construct the mapping relationship between the inspection results of cold-rolled strip steel surface defects and basic material data;

[0041] In the embodiments of the present invention, the inspection results of cold-rolled strip steel defects include: production date, unit, grade, coil number, strip steel length, strip steel width, strip steel thickness and surface inspection defect data, where the defect data includes defect code, defect name, lengths and widths or areas of various defects, and the positions of defects on the strip steel (the positions of defects in the length and width of the strip steel);

[0042] Preprocess the read defect data, including: (1) eliminating abnormal surface inspection data; (2) normalizing the surface inspection data of multiple surface inspection instruments of different units, including: unifying units and defect names; (3) taking the length direction of the long side of the strip steel on the OS operation side as the longitudinal coordinate axis Y, the width direction of the wide side of the strip steel head as the transverse coordinate axis X, and the intersection point of the long and wide sides of the strip steel on the OS operation side as the origin to determine the positions of defects on the strip steel, and representing them in coordinates (X, Y), indicating the positions of defects in the length and width of the strip steel, and storing the defect data in the high-frequency database based on HBase.

[0043] Read the basic material data, including information such as unit, production date, order contract number, grade, steelmaking furnace number, continuous casting billet number, hot-rolled coil number, corresponding inlet / outlet coil numbers of each unit, sub-coil number of cold-rolled steel coil, weight, etc., and store them in the MPP low-frequency database.

[0044] Based on the production process of strip steel from steelmaking to hot rolling and cold rolling, the inlet coil number and outlet coil number of each unit are defined. Taking the outlet coil number of the previous unit being equal to the inlet coil number of the next unit as the logic, the basic material data with a one-to-one correspondence of "furnace number - slab number - coil number" is automatically connected in series according to the technological process, and information such as the tapping mark, specification, and contract corresponding to each coil is associated. A mapping relationship is constructed between the concatenated basic material data and the surface defect inspection results of cold-rolled strip steel based on the grade associated with the contract number in the basic material data.

[0045] S2. Digitally characterize the surface defect inspection results of the detected cold-rolled strip steel, including: calculating the defect index of the defect and its distribution along the strip width direction;

[0046] In the embodiment of the present invention, the method for digitally characterizing the defects on the cold-rolled plate is specifically as follows:

[0047] Statistically analyze the defects whose coordinates are located in the area of the current rectangular sliding window, read the defect type, the length and width of the defect, and traverse the statistical area on the surface of the cold-rolled plate through the rectangular sliding window. Finally, all the defects on the statistical area of the cold-rolled plate are obtained. The above statistical area can be the entire cold-rolled plate or a specified area of the cold-rolled plate.

[0048] Starting from the OS operation side edge of the plate width, a "rectangular sliding window" with a size of "strip full length × step length" is translated one step length by one step along the strip width direction towards the DS drive side of the strip until the entire strip width is searched. Identify the natural position (X, Y) coordinates of the defects, and automatically statistically calculate the number of defects, defect length, and defect width of various types of defects on the strip surface within the range of the "strip full length × step length" rectangular sliding window by category. Finally, it is digitally characterized by the "defect index". The defect index calculated for each step length calculation domain is recorded in an Excel file in units of coils and stored in the big data center. Among them, the "rectangular sliding window" is also the "calculation domain", and its area S is equal to "strip full length L × step length L0"; the step length L0 refers to the "width range value" along the strip width direction and can be set according to the needs of process analysis. The setting of the step length L0 can be a fixed value, such as 50mm, 100mm, 200mm, 300mm..., or the strip width can be divided into several small segments, and each small segment after equal division is the step length. When setting the step length, the total width of the strip and the amount of calculation need to be considered. The wider the step length, the greater the amount of calculation; when using a fixed value as the step length, it is necessary to consider being able to cover the entire strip width, otherwise the remaining part will be calculated separately. After determining the step length, the area S of the rectangular sliding window, that is, the strip surface area corresponding to the calculation domain, can be determined.

[0049] In the embodiments of the present invention, the defect index is a defect quantification index, which comprehensively considers the number and area of defects, and more objectively characterizes the severity of defects on the strip surface. It can be evaluated by the number of defects per unit area and the proportion of the defect area. The specific calculation formula is as follows:

[0050]

[0051] In the formula, represents the number of defects k in the area where the i-th rectangular sliding window is located, with the unit of piece; represents the length of the j-th defect k in the area where the i-th rectangular sliding window is located, with the unit of mm; represents the width of the j-th defect k in the area where the i-th rectangular sliding window is located, with the unit of mm; S represents the area of the rectangular window, representing the strip surface area covered by the sliding window, with the unit of m 2 ; N represents the number of rectangular sliding windows in the statistical area, and I k represents the defect index of defect k.

[0052] (1) For the defect index I of all types of defects on the entire coil of strip, the sum of all defect types needs to be calculated to obtain the total number of all defects and the total defect area of all defects on the entire coil of strip, which are respectively At this time, N represents the number of rectangular sliding windows on the surface of the entire coil of strip, and N×S represents the surface area of the entire coil of strip.

[0053] (2) The defect index I of defect k on the entire coil of strip k , and its calculation formula is as shown above. At this time, N represents the number of rectangular sliding windows on the surface of the entire coil of strip, and N×S represents the surface area of the entire coil of strip.

[0054] (3) The defect index of all types of defects in the statistical area. At this time, i is the rectangular sliding window in the statistical area, N is the number of rectangular windows in the statistical area, and N×S represents the surface area of the statistical area. Calculate the total number of all defects and the total defect area of all defects in the statistical area. Assume that the statistical area is rectangular sliding windows 3 to 5, then the value of i is 3 to 5, and N = 3.

[0055] (4) The defect index of defect k in the statistical area. i is the rectangular sliding window in the statistical area, N is the number of rectangular windows in the statistical area, and N×S represents the surface area of the statistical area. Calculate the total number of defect k and the total defect area of defect k in the statistical area.

[0056] Calculate the "defect index of all types of defects in the whole coil of strip steel", the "defect index of a certain type of defect in the whole coil of strip steel", and the "defect index of steel quality defects (the sum of 4 types of defects such as inclusions, slag inclusions, bubbles, and surface peeling) in the whole coil of strip steel", and archive the calculation results of the defect index in the form of a table and store them in the big data center. Process technicians can flexibly call them, or input the date, grade, and defect category to display the digital defect data of each coil of strip steel produced on the same day on the computer screen.

[0057] The graphical representation of the distribution characteristics of defects in each fixed position area in the strip width direction is as follows: Each fixed position area in the strip width direction refers to "the area within B0 mm to the right of the OS operation side (i.e., within the range of 0 to B0 mm, where B0 can be 50, 100, 200 mm, or determined according to the actual situation of each factory)", "within the 1 / 4 strip width area (on both sides of the 1 / 4 strip width position line by area)", "within the 1 / 2 strip width area (on both sides of the 1 / 2 strip width position line by area)", "within the 3 / 4 strip width area (on both sides of the 3 / 4 strip width position line by area)" and "the area within B0 mm to the left of the DS drive side. The calculation of the defect index in each fixed position area in the strip width direction is to set the step size of the rectangular sliding window to B0, then move the sliding window to each fixed position area and align the center line of the sliding window with the center line of each fixed position area, automatically calculate their respective defect indices, archive them in the form of a table and graphically represent them, and store them in the big data center.

[0058] The digital representation of defects in any specified query position area in the strip width direction; Any desired query position area in the strip width direction refers to "on both sides of any specified query position line by area)", and its defect index calculation is to set the step size of the rectangular sliding window to B0, then move the sliding window to the any specified query position area and align the center line of the sliding window with the specified query position line, automatically calculate the defect index and display the result on the computer screen.

[0059] S3. Automatically form the digital characterization results of current cold-rolled sheet defects and their corresponding associated process parameter sequences. The associated process parameter sequences are arranged according to the process flow, i.e., information in the mapping relationship of basic material data such as grades - defect index I - steelmaking (key process parameter 1, key process parameter 2... ) - refining (key process parameter 1, key process parameter 2... ) - continuous casting (key process parameter 1, key process parameter 2... ) - hot rolling (key process parameter 1, key process parameter 2... ) - cold rolling (key process parameter 1, key process parameter 2... ). Construct a sequence for each cold-rolled coil number. The key process parameters are the process parameters in each process flow that have a greater impact on the defects of cold-rolled sheets.

[0060] S4. Adjust process parameters based on the defect index of the strip steel and optimize the thresholds of key process parameters. For the automatically formed digital characterization results of cold-rolled sheet defects and their corresponding associated process parameter sequences, analyze the correlation between the defect index and key process parameters based on big data, and then optimize the key process parameters according to the correlation. When the defect index of the strip steel has a strong correlation with certain key process parameters and the current value of this key process parameter is not sufficient to reduce the defect index, then tighten the threshold of this key process parameter (i.e., narrow the process control window); when the defect index of the strip steel has a weak correlation with certain key process parameters, then relax the threshold of this key process parameter (i.e., widen the process control window); when the defect index of the strip steel has a very weak correlation with certain key process parameters, then cancel this key process parameter or do not include it in the cleanliness determination rule. Analyze the correlation between the defect index and key process parameters based on big data as follows: Use the method based on minimizing the Residual Sum of Squares (RSS) in the regression tree to analyze the key associated process parameters of the defect index. The specific method is as follows:

[0061] (1) Calculate the initial total residual sum of squares (Total RSS) of the defect index:

[0062]

[0063] where, y i represents the actual defect index value, represents the average value of the defect index.

[0064] (2) For each key process parameter:

[0065] a. Try different splitting points and divide the data into left and right subsets.

[0066] b. Calculate the residual sum of squares (RSS) of the defect index in the left and right subsets at each splitting point.

[0067] RSS = RSSleft +RSS right ;

[0068]

[0069] where y i is the defect index value of the defect index, and are the average values of the defect indices in the left and right subsets respectively;

[0070] c. Select the best segmentation point: Select the segmentation point that maximizes the reduction in the total residual sum of squares as the best segmentation point of the key process parameter.

[0071] (3) Calculate the importance of the key process parameter: Calculate its importance based on the reduction in the total residual sum of squares at the segmentation point of each key process parameter. Divide the reduction in the total residual sum of squares of each key process parameter by the overall residual sum of squares to obtain the importance ratio of the key process parameter. The importance ratio is the correlation between the corresponding key process parameter and the defect index.

[0072] The present invention will be described in more detail with respect to the above-mentioned digital characterization method for surface inspection defects of cold-rolled sheets. The method is as follows:

[0073] Step 1.: Read the inspection results of the strip surface defect inspector of the cold rolling mill. The inspection results of each coil are stored in the cloud ftp server in the format that the first row is the basic coil information such as production date, unit, grade, coil number, strip length, strip width, strip thickness, etc., and the remaining rows are the defect details such as the defect code, defect name, length and width of various defects, and the natural position of the defect on the strip (the position of the defect in the strip length and width) corresponding to each defect.

[0074] Step 2.: Preprocess the data of the strip surface defect inspector. Process the defect data of the surface inspection system in the following manner and store the results in a high-frequency database based on HBase:

[0075] (1) Parse and convert the defect data of the surface inspection system read in Step 1 into Excel format, merge the basic information of each coil into the material basic information table, and use the unit, coil number, and basic defect information table as the defect detail data table;

[0076] (2) Determine whether the surface defect data is abnormal according to whether the coil number is empty or the coil number starts with S. Remind of abnormal and missing surface defect data;

[0077] (3) Use inclusion, slag inclusion, surface peeling, and bubble defects as the standard names of steel quality defects, and standardize the corresponding defects of the data collected by each surface inspection instrument; uniformly convert the length unit collected by the surface inspection instrument to mm;

[0078] (4) Uniformly stipulate that the length direction of the long side of the strip on the OS operation side is the longitudinal coordinate axis Y, the width direction of the wide side of the strip head is the transverse coordinate axis X, and the intersection point of the long and wide sides of the strip on the OS operation side is the origin, that is, the position coordinates of the operation side and the strip head are (0, 0).

[0079] Step 3: Basic material data collection and data concatenation. Write a program to achieve the following functions: automatically collect and concatenate data, and connect the data to the big data center:

[0080] (1) Collect the basic material information of the whole process from steelmaking, hot rolling to cold rolling, including information such as unit, furnace number, casting billet number, corresponding inlet / outlet coil numbers of each unit, material length, width, thickness specifications, steel grade, and contract number, and store it in the MPP low-frequency database; collect contract information such as contract number, order use, and customer name, and store it in the MPP low-frequency database;

[0081] (2) Based on the production process of the strip from steelmaking to hot rolling and cold rolling, define the inlet coil number and outlet coil number of each unit. Taking the outlet coil number of the previous unit = the inlet coil number of the next unit as the logic, write SQL to achieve the full-process automatic concatenation of the material information with a one-to-one correspondence of "furnace number - casting billet number - coil number", associate the tapping mark, specifications, and contract information corresponding to each coil, and further associate the basic contract information according to the contract number;

[0082] (3) According to the coil number rule of the strip surface inspection data collected, associate the surface inspection defect data with the previously concatenated data according to the surface inspection data coil number = the inlet coil number of the cold rolling unit, and construct an analysis data theme.

[0083] Step 4: Automatic defect statistics and digital characterization calculation. Build a configuration table according to the unit and step length, write a stored procedure to read the corresponding configuration to achieve the following functions, and construct a scheduled task to automatically calculate the data of the newly produced coil every 6 hours for incremental update:

[0084] Starting from the side edge of the OS operation of the plate width, move a "rectangular sliding window" of "strip full length × step length" along the plate width direction to the DS drive side of the strip one step length at a time for searching until the entire plate width is searched. Identify the natural position (X, Y) coordinates of the defects, automatically statistically calculate the number of various defects, defect length, and defect width on the strip surface within the "strip full length × step length" rectangular sliding window by category, and further calculate the defect index within each step length calculation domain. Store the calculated number of defects, defect length, defect width, and defect index in the big data center. For a certain unit, set the step length to 50 mm, and the calculation results of the defect index of the inclusion defects within each step length calculation domain are shown in Table 1.

[0085] Table 1 Calculation results of the defect index of inclusion defects within each step length calculation domain

[0086]

[0087]

[0088] Step 5: Digital characterization of defects. Based on the calculation result data in Step 4, develop a program to automatically calculate the following digital characterizations of defects. Develop a front-end web page based on Java to realize that according to the selected conditions, the digital characterizations of strip defects can be quickly displayed on the computer screen:

[0089] (1) Macroscopic characterization. The computer calculates the "defect index of all types of defects in the whole coil of strip", the "defect index of a certain type of defect in the whole coil of strip", and the "defect index of steel quality defects (the sum of 4 types of defects such as inclusions, slag inclusions, bubbles, and surface peeling) in the whole coil of strip" for each coil by combining the statistical results of each rectangular sliding window according to the calculation results in Step 4, and archives the defect index calculation results in the form of a table and stores them in the big data center. Process technicians can flexibly call them for secondary statistical analysis, or input the date, grade, and defect category, and the digital defect data of each coil of strip produced on the same day can be displayed on the computer screen. For example, process technicians call the data to analyze the "defect index of steel quality defects in the whole coil of strip" after rolling 9 consecutive slabs produced in the same furnace on the continuous annealing unit, and draw a graph as shown in Figure 2 shown in the figure, and find that the steel quality defect indexes of Coil C015 and the penultimate Coil C022 are relatively large, so defect traceability analysis can be carried out to check whether there are abnormalities in the steelmaking continuous casting process. Process technicians can also input the date, grade, and defect category to query and display the digital defect data of each coil of strip produced on the same day from the computer screen, as shown in Figure 3 shown in the figure.

[0090] (2) Graphic representation of the distribution characteristics of defects in each fixed position area in the strip width direction. Set the step size of the rectangular sliding window in Step 4 to B0 = 100 mm, and calculate all types of defects and steel quality defects in 5 fixed position areas, namely, the OS operator side area, the 1 / 4 strip width area, the 1 / 2 strip width area, the 3 / 4 strip width area, and the DS drive side area. For example, the defect indexes of steel quality defects in 5 fixed position areas of a selected coil with the same rolling order in the same furnace are as shown in Figure 4 shown in the figure. Process technicians can obviously find that the defect index in the 3 / 4 strip width area along the strip width direction is relatively large according to Figure 4 this, infer that there is a phenomenon of side flow or blockage on one side of the tundish during continuous casting, and thus provide a basis for improving the continuous casting process.

[0091] Step 6: Trend analysis of defect conditions. Select the cold rolling continuous annealing C308 unit, DC04 grade, and the production time from June to December 2022, and analyze the occurrence trend of steel quality defects monthly. The results are shown inFigure 5 Process technicians or quality management personnel can then Figure 5 find that the steel quality defect indicators began to deteriorate in November, and then organize efforts to conduct reverse checks on the processes to prevent quality control failures.

[0092] Step 7: Adjustment of key associated process parameters. For the defect index of cold-rolled coil steel quality defects, the range of key initial associated process parameters is configured as shown in Table 2. In the table, the number "1" for steel quality defects represents that this associated process parameter is used as the initial key associated parameter for steel quality defects, and the number "0" represents that this associated process parameter is not used as the initial key associated parameter for steel quality defects.

[0093] Table 2 Configuration Table of Initial Associated Process Parameters for Steel Quality Defects

[0094] Key process parameters Steel defects Actual value of converter additional blowing 1 Actual value of converter end-point oxygen 1 Actual value of converter slag stopping 1 Actual value of RH inlet oxygen 1 Actual value of RH aluminum consumption per ton of steel 1 Actual value of oxygen blowing for temperature increase 1 Actual value of RH scrap addition 1 Actual value of early slag arrival 1 Actual value of the next piece under early slag arrival 1 Actual value of superheat 1 Actual value of abnormal casting speed control 1

[0095] For the defect index of cold-rolled coil steel quality defects, select the cold rolling continuous annealing C308 unit, DC04 grade, and data from June to December 2022 for production time to analyze its initial key process parameters. The results of importance analysis are shown in Table 3, and the importance ratio is the correlation between the corresponding key process parameter and the defect index.

[0096] Table 3 Calculation Results of the Importance of Associated Process Parameters for Steel Quality Defects

[0097] Initial process parameters Importance ratio Actual value of RH inlet oxygen 0.2426 Actual value of the next piece under early slag arrival 0.1481 Actual value of oxygen blowing for temperature increase 0.1198 Actual value of converter end-point oxygen 0.1190 Actual value of RH scrap addition 0.0909 Actual value of abnormal casting speed control 0.0863 Actual value of superheat 0.0746 Actual value of early slag arrival 0.0619 Actual value of RH aluminum consumption per ton of steel 0.0491 Actual value of converter additional blowing 0.0077 Actual value of converter slag stopping 0.0000

[0098] The present invention has been described by way of example. Obviously, the specific implementation of the present invention is not limited by the above methods. As long as various non-substantive improvements are made by adopting the method concept and technical solution of the present invention, or the concept and technical solution of the present invention are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.

Claims

1. A digital characterization method for the inspection results of strip surface defects, characterized in that, The method includes the following steps: S1. Read the inspection results of cold-rolled plate defects from the strip surface defect inspector, and construct the mapping relationship between the inspection results of cold-rolled strip surface defects and the basic material data; S2. Digitally characterize the inspection results of the detected cold-rolled strip surface defects, including: calculating the defect index of the defects and their distribution along the strip width direction; S3. Automatically track the digital characterization results of the defects of the current cold-rolled plate and their corresponding associated process parameter sequences, and the associated process parameter sequences are arranged according to the process flow; Calculate the defect index using the number of defects per unit area and the defect area per unit area, and the defect index characterizes the severity of the defects on the strip surface; Defect index I of defect k k , and its calculation formula is as follows: In the formula, represents the number of defects k in the area where the i-th rectangular sliding window is located; represents the length of the j-th defect k in the area where the i-th rectangular sliding window is located; represents the width of the j-th defect k in the area where the i-th rectangular sliding window is located; S represents the area of the rectangular window; N represents the number of rectangular sliding windows in the statistical area; The method for obtaining the distribution of defects along the strip width direction is as follows: Form a sliding window based on a specified step size, and the sliding window slides along the width direction of the strip, dividing the strip into several statistical regions along the width direction; Calculate the defect index of each statistical region, that is, obtain the distribution of defects along the strip width direction.

2. The digital characterization method of the inspection result of strip surface defects according to claim 1, characterized in that Preprocess the inspection results of the cold-rolled plate defects read in step S1, specifically including the following steps: Eliminate the abnormal inspection data from the read defect data; Normalize the inspection data of multiple strip surface defect inspectors of different units, including: unifying the units and defect names; Take the long side length direction of the strip on the OS operation side as the longitudinal coordinate axis Y, the wide side width direction of the strip head as the transverse coordinate axis X, and the intersection point of the long and wide sides of the strip on the OS operation side as the origin to determine the positions of the defects on the strip.

3. The digital characterization method for the inspection results of strip surface defects according to claim 1, characterized in that The construction process of the mapping relationship between the inspection results of cold-rolled strip surface defects and the basic material data is as follows: Take the export coil number of the previous unit equal to the import coil number of the next unit, and concatenate the basic material data according to the process flow; construct the mapping relationship between the concatenated basic material data and the cold-rolled strip surface defect detection results based on the grades associated with the basic material data.

4. The digital characterization method for the inspection results of strip surface defects according to claim 1, characterized in that, The method for digitally characterizing the defects on the cold-rolled plate is as follows: Count the defects whose coordinates are located in the area where the current rectangular sliding window is located, read the defect type, the length and width of the defects, and traverse the statistical regions on the cold-rolled plate surface through the rectangular sliding window, and finally obtain all the defects on the cold-rolled plate statistical regions; The statistical region is the entire cold-rolled plate or a partial specified region of the cold-rolled plate.

5. The digital characterization method for the inspection results of strip surface defects according to claim 1, characterized in that After step S3, it further includes: S4. Adjust the associated process parameters based on the defect index of the strip and optimize the key process parameters.

6. The digital characterization method for the inspection results of strip surface defects according to claim 5, characterized in that, The adjustment method of the associated process parameters is as follows: (1) Analyze the correlation between the defect index and the key process parameters; (2) When the correlation between the strip defect index and the key process parameter is strong, and the current value of the key process parameter is not sufficient to reduce the defect index, then narrow the process control window; When the correlation between the strip defect index and the key process parameter is weak, then widen the process control window; When the correlation between the strip defect index and the key process parameter is very weak, then cancel the key process parameter.

Images