Online identification method for flatness of hot-rolled strip steel based on curve similarity discrimination

Abstract

The invention discloses a hot-rolled strip flatness online identification method based on curve similarity discrimination, and the method comprises the following steps: S1, obtaining the thickness distribution data of the full breadth of the cross section of a plate strip from an online system, namely the corresponding thickness value of the width of the plate strip after equal-distance dispersion; s2, acquiring a to-be-measured data group in the thickness value data corresponding to the tail; s3, extracting the distance between each data in the to-be-measured data set and the tail starting point, and performing normalization processing; s4, drawing a normalized data histogram and a width distribution histogram of data in a database; s5, the normalized data histogram is sequentially compared with a width distribution histogram of data in a database to calculate the Bhattacharyya distance, and the minimum value is compared with a required value k for judgment; the method has the advantages that the ratio sample is obtained based on the curve similarity discrimination method, and meanwhile, the database is perfected, so that the identification accuracy of the flatness of the tail part of the hot-rolled strip steel is relatively high.

Description

Technical Field

[0001] The invention relates to the field of metallurgical automation, and in particular to an online recognition method for the flatness of a hot-rolled strip steel based on curve similarity discrimination. Background Art

[0002] Hot-rolled strip tails are widely used in the industrial field, and their flatness is an important indicator for judging the quality of a batch of hot-rolled strips. Under the influence of factors such as uneven rolling temperature, material characteristics, and control system errors, the flatness of the hot-rolled strip tail along the width direction may be in various situations: the initial thickness is too high and then drops to flat; the thickness is flat at the beginning and the end, but the thickness is too high in the middle; the initial thickness is relatively flat and rises suddenly at the tail; the plate and strip are overall thicker; the plate and strip are overall thinner, etc. Defective hot-rolled strips do not meet the high-quality production requirements and will have a significant impact on the quality of the project. The current detection method mostly relies on manual detection, that is, manual inspection based on experience based on the thickness curve of the plate and strip cross section displayed instantly by the multi-function instrument. In addition to being easily affected by human factors, manual detection is difficult to accurately determine the existing thickness data, and it is very easy to make errors in judgment.

[0003] At present, scholars focus more on the judgment of defects in the flatness detection of hot-rolled strip, while few people conduct direct research on the judgment of the overall flatness of hot-rolled strip. For example, the patent "Strip Defect Evaluation Method" with the publication number CN114897795A mainly uses optical means to evaluate the degree of imbalance in the vertical direction of the roller-marked area in the strip image and the degree of imbalance in the horizontal direction of the roller-marked area to judge defects. The "Online Identification Method of Local High (Concave) Points in the Cross-Section of Plates and Strips" with the announcement number CN107470375B uses a linear fitting method to calculate the strip thickness curve, and obtains the height and width of the local high (concave) point from the difference distribution curve according to certain rules, and calculates the area of ​​the difference curve that exceeds the local high (concave) point threshold line to judge whether there is a defect in its high (concave) point. "An Intelligent Detection Method for Strip Defects" with publication number CN115035106A uses the first grayscale histogram of the strip surface for judgment. Its judgment method using the histogram is similar to that of this patent, but its focus is on the grayscale image of the strip surface rather than the thickness distribution curve of the strip cross section. At the same time, it does not establish a database for judgment.

[0004] Therefore, it is necessary to establish an online recognition method for the flatness of the tail of hot-rolled strip based on curve similarity judgment. When making curve judgments, the database can be continuously improved to make the recognition of the flatness of the tail of hot-rolled strip more and more accurate. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an online identification method for the flatness of a hot-rolled strip based on curve similarity judgment, which is used to solve the technical problem that the prior art only considers the clock problem without introducing other factors.

[0006] To achieve the above object, the present invention provides an online identification method for hot-rolled strip flatness based on curve similarity discrimination, comprising the following steps:

[0007] S1: Obtain the thickness distribution data of the full width of the strip cross section from the online system, that is, the thickness value corresponding to the strip width after being dispersed at equal distances;

[0008] S2: Cut the width of the plate strip after equidistant dispersion into three parts: head, middle and tail at equal intervals, and obtain the data group to be measured in the thickness value data corresponding to the tail;

[0009] S3: extract the distance between each data in the data group to be tested and the starting point of the tail, and perform normalization processing;

[0010] S4: Draw the normalized data histogram and the width distribution histogram of the data in the database;

[0011] S5: Compare the normalized data histogram with the width distribution histogram of the data in the database in turn to calculate the Bhattacharyya distance, and take the minimum value to compare with the required value k for judgment.

[0012] Further: The specific method of S2 is:

[0013] The width of the plate strip is cut into three parts at equal intervals: the head, the middle and the tail. Only the tail data to be judged is selected, and 1% of the tail end data is selected as the dividing point. The first 99% of the tail data is taken as the data group to be tested, that is, the last n data points in the tail data are eliminated, and the value range of n is 0<n<the number of thickness value data corresponding to the tail*1%;

[0014] Since the sudden increase in thickness at the end of the plate and strip is an inevitable phenomenon in processing, it will be sheared and discarded as waste in production and will not be taken into consideration.

[0015] Further: the specific method of S3 is:

[0016] The distance between each data in the test data group and the starting point of the tail is extracted, that is, the distance between the test data group and the tail head truncation plane, and normalized, as shown in the following formula:

[0017] l1=[L×p1];

[0018] l2=[L×p2];

[0019] l3=(L-l1-l2)*0.99;

[0020]

[0021]

[0022] Wherein, l1 is the number of thickness value data corresponding to the head intercepted in S2;

[0023] l2 is the number of thickness value data corresponding to the middle part intercepted in S2;

[0024] l3 is the number of data in the data group to be tested in the thickness value data corresponding to the tail intercepted in S2;

[0025] p1 is the percentage of the header data in the total data length;

[0026] p2 is the percentage of the middle data in the total length of the data;

[0027] k i The distance between the tail data and the tail head truncation plane, in mm;

[0028] Normalizing the distances of each data in the data group to be tested from the tail and head truncation surfaces can avoid the influence of the determination result on the data to be tested due to its own strip length.

[0029] Further: In said S4, when drawing the width distribution histogram of the normalized data and the data in the database, the group distance of the width distribution histogram needs to be determined, as shown in the following formula:

[0030]

[0031] Among them, R is the range of the full width of the plate cross section; K is the number of groups in the histogram, and n is the number of thickness value data corresponding to the tail in S2 that are eliminated.

[0032] Further: the formula for calculating the Bhattacharyya distance in S5 is as follows:

[0033]

[0034] Among them, y i h represents the cross-sectional width of the plate corresponding to the ith data point in the data group to be tested after normalization, mm; ij Indicates the cross-sectional width of the plate corresponding to the i-th data point of the j-th group of data in the database, mm; B j It represents the Bhattacharyya coefficient calculated from the data set to be tested and the jth data set in the database after normalization; D Bjrepresents the Bhattacharyya distance calculated between the test data group and the jth data group in the database after normalization; D B It represents the judgment value, which is the minimum value of the Bhattacharyya distance calculated between the normalized data set and the data in the database.

[0035] Further: the required value k has a value range of 0.2≥k≥0.01;

[0036] The required value k needs to be selected according to the actual situation, that is, the similarity J between the data in the test data set and the data in the database is measured by using the Jaccard similarity coefficient analysis, and the value range of J is 0 to 1. k is determined according to the similarity J, that is, k=0.2J.

[0037] The invention has the beneficial effects of obtaining ratio samples based on a curve similarity discrimination method and establishing a complete database, so that the recognition accuracy of the flatness of the tail of the hot-rolled strip is higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention and the technical solutions in the prior art, the drawings required for use in the embodiments and the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some implementation methods recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 is a flow chart of the present invention;

[0040] Figure 2 A schematic diagram of the thickness distribution of a hot-rolled steel strip cross section according to a specific embodiment of the present invention;

[0041] Figure 3 A schematic diagram of the cross-sectional thickness distribution of a hot-rolled steel strip after normalization according to a specific embodiment of the present invention;

[0042] Figure 4 A normalized hot-rolled steel strip cross-section thickness distribution histogram according to a specific embodiment of the present invention;

[0043] Figure 5 A schematic diagram of the cross-sectional thickness distribution of hot-rolled strip steel in databases 1, 2, and 3 of a specific embodiment of the present invention;

[0044] Figure 6 A database 1 is a histogram of the cross-sectional thickness distribution of hot-rolled strip steel according to a specific embodiment of the present invention;

[0045] Figure 7 A database 2 is a hot-rolled strip cross-sectional thickness distribution histogram according to a specific embodiment of the present invention;

[0046] Figure 8 Database 3 is a histogram of the cross-sectional thickness distribution of hot-rolled strip according to a specific embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0048] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can be applied to other similar scenarios based on these drawings without any creative work.

[0049] In addition, it can be understood that although the efforts made in this development process may be complex and lengthy, for ordinary technicians in the field related to the contents disclosed in this application, some design, manufacturing or production changes based on the technical contents disclosed in this application are just conventional technical means and should not be understood as the contents disclosed in this application are insufficient.

[0050] In order to better understand the solutions of the embodiments of the present application, some relevant terms and concepts that may be involved in the embodiments of the present application are first introduced below.

[0051] Unless otherwise specified, all embodiments and optional embodiments of the present application can be combined with each other to form a new technical solution.

[0052] See also Figure 1 , the flow chart of the present invention comprises the following steps:

[0053] S1: Obtain the thickness distribution data of the full width of the strip cross section from the online system, that is, the thickness value corresponding to the strip width after being dispersed at equal distances;

[0054] S2: Cut the width of the plate strip after equidistant dispersion into three parts: head, middle and tail at equal intervals, and obtain the data group to be measured in the thickness value data corresponding to the tail;

[0055] S3: extract the distance between each data in the data group to be tested and the starting point of the tail, and perform normalization processing;

[0056] S4: Draw the normalized data histogram and the width distribution histogram of the data in the database;

[0057] S5: Compare the normalized data histogram with the width distribution histogram of the data in the database in turn to calculate the Bhattacharyya distance, and take the minimum value to compare with the required value k for judgment.

[0058] Specifically:

[0059] Obtaining the thickness distribution data of the full width of the strip cross section from the online system, which is a multifunctional instrument and control system for detecting the strip;

[0060] In production, the thickness of the end of the hot-rolled strip is inevitably increased due to uneven temperature distribution. It will be sheared and removed as waste in production. Therefore, the width of the strip is cut into three parts at equal intervals, the head, the middle and the tail. The tail includes the end of the strip, so only the tail data that needs to be judged is selected. Usually, 1% of the tail end data is selected as the dividing point, and the first 99% of the tail data is taken as the data group to be tested, that is, the end n data points in the tail data are removed, and the value range of n is 0<n<the number of thickness value data corresponding to the tail * 1%;

[0061] The remaining tail length is normalized, that is, the distance between each data in the test data group and the starting point of the tail is extracted, that is, the distance between the test data group and the tail head truncation plane is normalized, as shown in the following formula:

[0062] l1=[L×p1];

[0063] l2=[L×p2];

[0064] l3=(L-l1-l2)*0.99;

[0065]

[0066]

[0067] Among them, l1 is the number of thickness value data corresponding to the head intercepted in S2; l2 is the number of thickness value data corresponding to the middle intercepted in S2; l3 is the number of data in the data group to be tested in the thickness value data corresponding to the tail intercepted in S2; p1 is the percentage of the head data in the total data length; p2 is the percentage of the middle data in the total data length; k i The distance between the tail data and the tail head truncation plane, in mm;

[0068] Get the width distribution histogram of the processed data, and make histograms of the data in the database in turn, that is, draw the width distribution histograms of the normalized data and the data in the database respectively. When drawing, determine the group distance of the width distribution histogram, as shown in the following formula:

[0069]

[0070] Among them, R is the range of the full width of the strip cross section; K is the number of groups in the histogram, and n is the number of thickness value data corresponding to the tail in S2 that are eliminated;

[0071] Finally, the histogram of the processed data is compared with the histogram of each group of data in the database to calculate the Bhattacharyya distance. The minimum value of the calculated results is taken as the judgment standard and compared with the similarity k between the data group to be tested and the database. The similarity k is the required value.

[0072] The formula for calculating Bhattacharyya distance is as follows:

[0073]

[0074] Among them, y i h represents the cross-sectional width of the plate corresponding to the ith data point in the data group to be tested after normalization, mm; ij Indicates the cross-sectional width of the plate corresponding to the i-th data point of the j-th group of data in the database, mm; B j It represents the Bhattacharyya coefficient calculated from the data set to be tested and the jth data set in the database after normalization; D Bj represents the Bhattacharyya distance calculated between the test data group and the jth data group in the database after normalization; D B represents the judgment value, which is the minimum value of the Bhattacharyya distance calculated between the normalized data set and the data in the database;

[0075] The value k is required to be in the range of 0.2 ≥ k ≥ 0.01;

[0076] The required value k needs to be selected according to the actual situation, that is, the similarity J between the data in the test data set and the data in the database is measured by using the Jaccard similarity coefficient analysis, and the value range of J is 0 to 1. k is determined according to the similarity J, that is, k=0.2J.

[0077] For an embodiment of the present invention, please refer to Figure 2 , is a schematic diagram of the cross-sectional thickness distribution of a hot-rolled strip according to a specific embodiment of the present invention, wherein the x-axis in the figure represents the distance between each cross-sectional area of ​​the observed strip and the head (starting point) of the strip, and the y-axis represents the thickness of the strip corresponding to the position;

[0078] See also Figure 3 , is a schematic diagram of the cross-sectional thickness distribution of hot-rolled strip steel after normalization according to a specific embodiment of the present invention, wherein the x-axis in the figure represents the distance of the sampling point after normalization, and the y-axis represents the plate and strip thickness corresponding to the position;

[0079] See also Figure 4,, is a normalized hot-rolled strip cross-sectional thickness distribution histogram of a specific embodiment of the present invention, wherein the x-axis represents the strip thickness interval, and the y-axis represents the frequency of occurrence of each thickness interval;

[0080] See also Figure 5 , is a schematic diagram of the cross-sectional thickness distribution of hot-rolled strips in databases 1, 2, and 3 of a specific embodiment of the present invention, wherein the x-axis represents the distance of each cross section of the strip in databases 1, 2, and 3 (1#, 2#, and 3#) from the head (starting point) of the strip, and the y-axis represents the thickness of the strip corresponding to the position;

[0081] See also Figure 6 , is a histogram of the cross-sectional thickness distribution of hot-rolled strip steel in database 1 according to a specific embodiment of the present invention, wherein the x-axis represents the thickness interval of the strip in database 1, and the y-axis represents the frequency of occurrence of each thickness interval of the strip in database 1;

[0082] See also Figure 7 , is a histogram of the cross-sectional thickness distribution of hot-rolled strip steel in database 2 according to a specific embodiment of the present invention, wherein the x-axis represents the thickness interval of the strip in database 2, and the y-axis represents the frequency of occurrence of each thickness interval of the strip in database 2;

[0083] See also Figure 8 , which is a histogram of the cross-sectional thickness distribution of hot-rolled strip in database 3 according to a specific embodiment of the present invention, wherein the x-axis represents the thickness interval of the strip in database 3, and the y-axis represents the frequency of occurrence of each thickness interval of the strip in database 3.

[0084] The above is a very detailed description of one or more embodiments of the present invention, but the content is only a specific example of the present invention and cannot be considered to limit the scope of application of the present invention. Any other methods and changes proposed based on the content of the present invention should fall within the scope of patent protection of the present invention.

Claims

1. An online identification method for hot-rolled strip flatness based on curve similarity discrimination, characterized in that: The steps include: S1: Obtain the thickness distribution data of the full width of the strip cross section from the online system, that is, the thickness value corresponding to the strip width after being dispersed at equal distances; S2: Cut the width of the plate strip after equidistant dispersion into three parts: head, middle and tail at equal intervals, and obtain the data group to be measured in the thickness value data corresponding to the tail; S3: extract the distance between each data in the data group to be tested and the starting point of the tail, and perform normalization processing; S4: Draw the normalized data histogram and the width distribution histogram of the data in the database; S5: Compare the normalized data histogram with the width distribution histogram of the data in the database in turn to calculate the Bhattacharyya distance, and take the minimum value to compare with the required value k for judgment.

2. The method for online flatness identification of hot-rolled strip based on curve similarity discrimination according to claim 1 is characterized in that: The specific method of S2 is: The width of the plate and strip is cut into three parts at equal intervals, and only the tail data that needs to be judged is selected. 1% of the tail end data is selected as the dividing point, and the first 99% of the tail data is taken as the data group to be tested, that is, the last n data points in the tail data are eliminated, and the value range of n is 0<n<the number of thickness value data corresponding to the tail*1%.

3. The method for online flatness identification of hot-rolled strip steel based on curve similarity discrimination according to claim 1, characterized in that: The specific method of S3 is: The distance between each data in the data group to be tested and the starting point of the tail is extracted, that is, the distance between the data group to be tested and the tail-head truncation plane is normalized, as shown in the following formula: l1=[L×p1]; l2=[L×p2]; l3＝(L-l1-l2)*0.99； k i ={k1,k2,k3,......,k l3 }; Wherein, l1 is the number of thickness value data corresponding to the head intercepted in S2; l2 is the number of thickness value data corresponding to the middle part intercepted in S2; l3 is the number of data in the data group to be tested in the thickness value data corresponding to the tail intercepted in S2; p1 is the percentage of the header data in the total data length; p2 is the percentage of the middle data in the total length of the data; k i It is the distance between the tail data and the tail head truncation plane, in mm.

4. The method for online flatness identification of hot-rolled strip steel based on curve similarity discrimination according to claim 1, characterized in that: In S4, when drawing the width distribution histogram of the normalized data and the data in the database, the group distance of the width distribution histogram must be determined, as shown in the following formula: Among them, R is the range of the full width of the plate cross section; K is the number of groups in the histogram, and n is the number of thickness value data corresponding to the tail in S2 that are eliminated.

5. The method for online flatness identification of hot-rolled strip steel based on curve similarity discrimination according to claim 1, characterized in that: The formula for calculating the Bhattacharyya distance in S5 is as follows: Among them, y i h represents the cross-sectional width of the plate corresponding to the ith data point in the data group to be tested after normalization, mm; ij Indicates the cross-sectional width of the plate corresponding to the i-th data point of the j-th group of data in the database, mm; B j It represents the Bhattacharyya coefficient calculated from the data set to be tested and the jth data set in the database after normalization; D Bj represents the Bhattacharyya distance calculated between the test data group and the jth data group in the database after normalization; D B It represents the judgment value, which is the minimum value of the Bhattacharyya distance calculated between the normalized data set and the data in the database.

6. The method for online flatness identification of hot-rolled strip steel based on curve similarity discrimination according to claim 1, characterized in that: The required value k is in the range of 0.2≥k≥0.01; The required value k needs to be selected according to the actual situation, that is, the similarity J between the data in the test data set and the data in the database is measured by using the Jaccard similarity coefficient analysis, and the value range of J is 0 to 1. k is determined according to the similarity J, that is, k=0.2J.

Images