A dynamic correction method for pre-oscillating roll gap targeting the head and tail deviation characteristics of strip

Abstract

This application discloses a dynamic correction method for pre-sway roll gap targeting the head and tail deviation characteristics of strip. The method includes acquiring strip data, preprocessing and segmenting the head and tail data; identifying deviation trends in the head and tail data; dynamically calculating the pre-sway roll gap adjustment amount to compensate for the deviation trend; sending the pre-sway roll gap adjustment amount to the rolling mill control system to dynamically correct the pre-sway roll gap setting value for the current rolling pass; and performing self-learning optimization of the adjustment amount based on the rolling results and on-site conditions of the current rolling pass for pre-sway roll gap correction under similar rolling conditions. This dynamic correction method for pre-sway roll gap targeting the head and tail deviation characteristics of strip can proactively predict and dynamically intervene before deviation occurs, achieving proactive suppression of deviation trends, especially in the head and tail regions with the highest risk. Therefore, it can improve the accuracy of strip rolling, yield, and the level of production line intelligence and feedforward.

Description

Technical Field

[0001] This invention belongs to the field of automatic control technology for hot strip rolling, and in particular relates to a dynamic correction method for pre-sway roll gap for strip head and tail deviation characteristics. Background Technology

[0002] Hot strip rolling technology is mainly used in the production of high-strength alloy strips, typically stainless steel strips. Among these materials, titanium alloy strips have become an important rolling target due to their unique properties. Titanium alloys, with their excellent strength-to-weight ratio, superior corrosion resistance, and biocompatibility, have become core materials in cutting-edge fields such as aerospace, high-end chemicals, medical devices, and defense. As a key basic product, the rolling process of titanium plates presents significant challenges. During the rolling of such strips, especially at the moment of slab entry and exit, the stress state in the head and tail areas changes drastically. Combined with the combined effects of factors such as the incoming strip shape (e.g., wedge-shaped, sickle-shaped), uneven temperature, and fluctuations in the roll system, severe lateral deviation is easily induced. This not only directly leads to increased scrap and cutting losses and a decrease in yield, but also potentially causes unplanned production line downtime and potential equipment damage, such as scratching guides and steel piling, which severely restricts the efficiency and stability of production.

[0003] Existing strip control methods often rely on operator experience. After observing deviation, manual intervention is typically performed. However, this approach is slow and inaccurate, making it difficult to keep up with the pace of high-speed rolling. Some solutions employ static models, but if these models encounter complex dynamic disturbances during rolling, such as sudden temperature changes, minor defects in the incoming material, or thermal expansion and contraction of the rolls, the preset roll gap model struggles to adapt in real time, still resulting in deviation at both ends and failing to achieve the desired effect. Summary of the Invention

[0004] To address the aforementioned issues, this invention provides a dynamic correction method for pre-sway roll gaps targeting the head and tail deviation characteristics of strip. This method can proactively predict and dynamically intervene before deviation occurs, thereby actively suppressing deviation trends, especially in the head and tail regions where the risk is highest. Therefore, it can improve the accuracy of strip rolling, yield, and the level of intelligence and feedforward in the production line.

[0005] This invention provides a method for dynamic correction of pre-sway roll gap for strip head and tail deviation characteristics, comprising:

[0006] Obtain strip data, and preprocess and segment the head and tail data;

[0007] Identify deviation trends in the head data and the tail data;

[0008] The pre-sway roller gap adjustment amount is dynamically calculated to compensate for the deviation trend;

[0009] The pre-sway roll gap adjustment amount is sent to the rolling mill control system to dynamically correct the pre-sway roll gap setting value for the current rolling pass.

[0010] Based on the rolling results and on-site conditions of the current rolling pass, the adjustment amount is optimized through self-learning to be used for pre-sway roll gap correction under subsequent similar rolling conditions.

[0011] Preferably, in the above-mentioned dynamic correction method for pre-sway roll gap for strip misalignment characteristics, the step of acquiring strip data and preprocessing the head and tail data includes:

[0012] Obtain the sequence of strip data, filter out data points with the same deviation value at the head and tail, and remove a preset number of data points from the head and tail data.

[0013] Preferably, in the above-mentioned dynamic correction method for pre-sway roll gap for strip head and tail deviation characteristics, the segmentation processing of the head data and tail data includes:

[0014] The first 10% to 20% of the header data is divided into n consecutive subarrays, and the last 10% to 20% of the tail data is divided into n consecutive subarrays.

[0015] Preferably, in the above-mentioned dynamic correction method for pre-sway roll gap based on the head and tail deviation characteristics of strip, the identification of deviation trend based on the head data and the tail data includes:

[0016] Calculate the difference between adjacent data points: d i =a i -a i-1 , i=2,3,…n;

[0017] Among them, a1, a2, ... a n Data points are either at the beginning or end of a line, in mm or d. i This represents the difference between adjacent data points, in mm.

[0018] Preferably, in the above-mentioned dynamic correction method for pre-sway roll gap based on the head and tail deviation characteristics of strip, the identification of deviation trend based on the head data and the tail data further includes:

[0019] Calculate the sign function: ;

[0020] Calculate the trend strength coefficient: ;

[0021] Calculate the trend significance index: ;

[0022] When S≥S th When a significant trend is identified, the trend direction is determined based on the sign of the trend strength coefficient T. When the trend strength coefficient T is positive, it is a positive skew; when the trend strength coefficient T is negative, it is a negative skew.

[0023] Preferably, in the above-mentioned dynamic correction method for pre-sway roll gap based on the head and tail deviation characteristics of strip, the identification of deviation trend based on the head data and the tail data further includes:

[0024] Calculate the trend stability coefficient: ;

[0025] Where α is the trend stability coefficient, dimensionless, and λ is the decay factor, dimensionless, when hour, The stability threshold is dimensionless and confirms the reliability of the trend.

[0026] Preferably, in the above-mentioned dynamic correction method for pre-sway roll gap for strip head and tail deviation characteristics, the dynamic calculation of the pre-sway roll gap adjustment amount to compensate for the deviation trend includes:

[0027] Calculate the difference between the mean of the nth subarray and the first subarray: , where μ n μ1 is the mean of the data points in the nth subarray, and μ1 is the mean of the data points in the 1st subarray.

[0028] Calculation function using pre-swing roll gap adjustment amount:

[0029] ;

[0030] in, The unit is mm; k is the control coefficient, which is dimensionless.

[0031] Calculate the pre-swing roller gap adjustment amount .

[0032] Preferably, in the above-mentioned dynamic correction method for pre-sway roll gap for strip head and tail deviation characteristics, the dynamic calculation of the pre-sway roll gap adjustment amount to compensate for the deviation trend further includes:

[0033] The stability of the pre-swing roll gap adjustment amount is corrected using the following formulas to obtain the corrected head pre-swing roll gap adjustment amount △corrected,head and the corrected tail pre-swing roll gap adjustment amount △corrected,tail:

[0034] ;

[0035] The average of the corrected head pre-swing roll gap adjustment Δcorrected,head and the corrected tail pre-swing roll gap adjustment Δcorrected,tail is used as the final pre-swing roll gap adjustment.

[0036] ;

[0037] in, The final pre-sway roll gap adjustment amount is expressed in mm.

[0038] Preferably, in the above-mentioned dynamic correction method for pre-sway roll gap based on the head and tail deviation characteristics of strip, the step of performing self-learning optimization of the adjustment amount based on the rolling results and field conditions of the current rolling pass for subsequent pre-sway roll gap correction under similar rolling conditions includes:

[0039] The self-learning adjustment amount is calculated based on the following formula: ;

[0040] ;

[0041] in, For the nth volume of tape in The self-learning adjustment range of the rack, in mm; The heritability coefficient represents... The transfer ratio of the rack's own historical adjustment amounts is dimensionless; D 11 The attenuation coefficient represents... The decay rate of historical adjustment amounts is dimensionless; Δ (n-1) represents the adjustment amount of the (n-1)th roll of strip on the F1 stand, which is dimensionless; For the nth volume of tape in The self-learning adjustment range of the rack, in mm; The heritability coefficient represents... The transfer ratio of the rack's own historical adjustment amounts is dimensionless; D 22 The attenuation coefficient represents... The decay rate of historical adjustment amounts, dimensionless; The heritability coefficient represents... rack pair The transfer ratio of the rack is dimensionless; D 21 The attenuation coefficient represents... rack pair The attenuation ratio of the rack is dimensionless.

[0042] Preferably, in the above-mentioned method for dynamic correction of pre-sway roll gap for strip misalignment characteristics, the strip is a titanium plate, stainless steel plate, copper plate, aluminum plate or nickel-based alloy plate.

[0043] As described above, the pre-sway roll gap dynamic correction method for strip head and tail deviation characteristics provided by the present invention includes: acquiring strip data; preprocessing and segmenting the head and tail data; identifying deviation trends in the head and tail data; dynamically calculating the pre-sway roll gap adjustment amount to compensate for the deviation trend; sending the pre-sway roll gap adjustment amount to the rolling mill control system to dynamically correct the pre-sway roll gap setting value of the current rolling pass; and performing self-learning optimization of the adjustment amount based on the rolling results and on-site conditions of the current rolling pass for pre-sway roll gap correction under similar rolling conditions. Therefore, this method can proactively predict and dynamically intervene before deviation occurs, achieving proactive suppression of deviation trends, especially in the head and tail regions where the risk is highest. This improves the accuracy of strip rolling, yield, and the intelligence and feedforward level of the production line. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, 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 embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0045] Figure 1 This is a schematic diagram of an embodiment of a pre-swaying roll gap dynamic correction method for strip head and tail deviation characteristics provided by the present invention. Detailed Implementation

[0046] The core of this invention is to provide a dynamic correction method for pre-sway roll gap for strip misalignment characteristics at the head and tail. This method can proactively predict and dynamically intervene before misalignment occurs, thereby actively suppressing the misalignment trend, especially for the head and tail areas where the risk is highest. Therefore, it can improve the accuracy of strip rolling, yield, and the level of intelligence and feedforward of the production line.

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] An implementation example of the pre-sway roll gap dynamic correction method for strip head and tail deviation characteristics provided by this invention is as follows: Figure 1 As shown, Figure 1This is a schematic diagram of an embodiment of a pre-sway roll gap dynamic correction method for strip head and tail deviation characteristics provided by the present invention. The method may include the following steps:

[0049] S1: Obtain strip data, and preprocess and segment the head and tail data;

[0050] It should be noted that the strip material here can be titanium, stainless steel, copper, aluminum, or nickel-based alloy. Furthermore, the strip data refers to the parameters corresponding to the previous strip, specifically the deviation value of each point relative to the offset centerline. This preprocessing method can include data analysis and filtering. Specifically, the strip data sequence can be obtained, and data points with the same deviation value at the beginning and end can be filtered out. Additionally, a predetermined number of data points from the beginning and end of the data can be removed. This is because the binocular camera is continuously working, recording data even when no strip is passing by. Therefore, it retains the data detected at the end of the previous strip until the beginning of the next strip arrives. Filtering out identical data at the beginning and end ensures that the obtained data is entirely from the current strip and does not involve data from the previous strip for pre-swing self-learning calculations. Moreover, the beginning and end of the strip may have irregular shapes when passing through the frame. Detecting these shapes can lead to excessive deviation values. Therefore, a predetermined number of these data points must be removed to prevent errors and outliers that could affect the judgment. The segmentation of the head and tail data can preferably include: dividing the first 10% to 20% of the head data into n consecutive subarrays, and dividing the last 10% to 20% of the tail data into n consecutive subarrays. This obtains key data to assist in determining deviation.

[0051] S2: Identify deviation trends in head and tail data;

[0052] Specifically, we can first calculate the difference between adjacent data points: d i =a i -a i-1 , i=2,3,…n;

[0053] Among them, a1, a2, ... a n These are head or tail data points, in mm, which can be obtained through the system. (d) i The difference between adjacent data points is expressed in mm. These data sequences can initially show some deviation trends. For example, if the difference becomes larger and larger, it indicates that the deviation is becoming more serious.

[0054] Furthermore, it may also include:

[0055] Calculate the sign function: ;

[0056] Calculate the trend strength coefficient: ;

[0057] Calculate the trend significance index: ;

[0058] S th The significance threshold is a dimensionless value determined by rolling history experience, typically ranging from 0.8 to 1.2, when S ≥ S th When a significant trend is identified, the trend direction is determined based on the sign of the trend strength coefficient T. When the trend strength coefficient T is positive, it is a positive skew; when the trend strength coefficient T is negative, it is a negative skew.

[0059] Furthermore, it may also include:

[0060] Calculate the trend stability coefficient: ;

[0061] Where α is the trend stability coefficient, dimensionless, and λ is the attenuation factor, dimensionless, obtained from historical rolling experience, generally taken as 0.05 to 0.10. hour, The stability threshold is dimensionless and obtained through historical rolling experience, typically ranging from 0.7 to 0.9, to confirm the reliability of the trend.

[0062] S3: Dynamically calculates the pre-sway roller gap adjustment amount to compensate for deviation trend;

[0063] Specifically, when a significant trend exists, the difference between the mean of the nth subarray and the first subarray can be calculated: , where μ n μ1 is the mean of the data points in the nth subarray, and μ1 is the mean of the data points in the 1st subarray.

[0064] Calculation function using pre-swing roll gap adjustment amount:

[0065] ;

[0066] in, The unit is mm; k is the control coefficient, which is dimensionless and generally ranges from 0.8 to 1.

[0067] Calculate the pre-swing roll gap adjustment amount .

[0068] Furthermore, the stability of the pre-swing roll gap adjustment is corrected using the following formulas to obtain the corrected head pre-swing roll gap adjustment Δcorrected,head and the corrected tail pre-swing roll gap adjustment Δcorrected,tail:

[0069] ;

[0070] The average of the corrected head pre-swing roll gap adjustment Δcorrected,head and the corrected tail pre-swing roll gap adjustment Δcorrected,tail is used as the final pre-swing roll gap adjustment.

[0071] ;

[0072] in, The final pre-sway roll gap adjustment is expressed in mm. This result incorporates the adjustment amounts at both the beginning and end, resulting in a more accurate adjustment.

[0073] S4: Send the pre-sway roll gap adjustment amount to the rolling mill control system to dynamically correct the pre-sway roll gap setting value of the current rolling pass;

[0074] It should be noted that this allows for automatic correction before deviation occurs, without human intervention. The correction is faster and more accurate, avoiding impact on production schedules and thus saving production costs.

[0075] S5: Based on the rolling results and on-site conditions of the current rolling pass, perform self-learning optimization of the adjustment amount for use in the pre-sway roll gap correction under subsequent similar rolling conditions.

[0076] This step may specifically include:

[0077] The self-learning adjustment amount is calculated based on the following formula: ;

[0078] ;

[0079] in, For the nth volume of tape in The self-learning adjustment amount of the frame (e.g., rolling force or roll gap adjustment amount), in mm; The heritability coefficient represents... The transfer ratio of the rack's own historical adjustment is dimensionless, 0 ≤ H. 11 ≤1; D 11 The attenuation coefficient represents... The decay rate of historical adjustment is dimensionless, 0 ≤ D 11 ≤1; Δ (n-1) represents the adjustment amount of the (n-1)th roll of strip on the F1 stand, which is dimensionless; For the nth volume of tape in The self-learning adjustment range of the rack, in mm; The heritability coefficient represents... The transfer ratio of the rack's own historical adjustment is dimensionless, 0 ≤ H.22 ≤1; D 22 The attenuation coefficient represents... The decay rate of historical adjustment amounts, dimensionless; The heritability coefficient represents... rack pair The transfer ratio of the frame is dimensionless and can be obtained through finite element simulation, 0 ≤ H. 21 ≤1; D 21 The attenuation coefficient represents... rack pair The attenuation ratio of the rack is dimensionless and can be obtained through finite element simulation, 0 ≤ D. 21 ≤1.

[0080] By establishing a predictive model, the roll gap inclination can be dynamically preset and optimized in real time before the beginning and end of the strip arrive, thus proactively suppressing the deviation trend and preventing problems before they occur.

[0081] As described above, the embodiment of the pre-sway roll gap dynamic correction method for strip head and tail deviation characteristics provided by the present invention includes: acquiring strip data; preprocessing and segmenting the head and tail data; identifying deviation trends in the head and tail data; dynamically calculating the pre-sway roll gap adjustment amount to compensate for the deviation trend; sending the pre-sway roll gap adjustment amount to the rolling mill control system to dynamically correct the pre-sway roll gap setting value of the current rolling pass; and performing self-learning optimization of the adjustment amount based on the rolling results and on-site conditions of the current rolling pass for pre-sway roll gap correction under similar rolling conditions. Therefore, it is evident that proactive prediction and dynamic intervention can be performed before deviation occurs, achieving proactive suppression of deviation trends, especially in the head and tail regions where the risk is highest. This improves the accuracy of strip rolling, yield, and the level of production line intelligence and feedforward.

[0082] By utilizing the aforementioned dynamic correction method for pre-sway roll gap targeting the head and tail deviation characteristics of strip, the trend of strip deviation at both ends can be accurately determined, and the adjustment amount can be dynamically calculated based on the trend. Through a self-learning control mechanism, the system can continuously optimize the adjustment strategy based on historical data and real-time feedback. This makes the system more adaptable and stable in complex and ever-changing production environments. By introducing parameters such as trend intensity coefficient, trend significance index, and trend stability coefficient, as well as the dynamic calculation and adaptive adjustment mechanism of the adjustment amount, the adjustment strategy becomes more scientific and reasonable, and can better cope with the problem of strip deviation under different conditions.

[0083] The following example, using the F2 frame and titanium plate head-to-tail deviation correction method, will provide a detailed explanation of the above method:

[0084] Step 1: Preprocess and segment the header and tail data:

[0085] Obtain the steel coil deviation data sequence, filter out data points with the same deviation value at the beginning and end, and remove the first 3 data points at the beginning and end after filtering. Then divide the first 20% of the data at the beginning into 10 consecutive subarrays and the last 20% of the data at the end into 10 consecutive subarrays.

[0086] Step 2: Perform trend identification on the data at the head and tail:

[0087] Calculate the difference between adjacent data points: d i =a i -a i-1 , i=2,3,…n;

[0088] Among them, a1, a2, ... a n These are head or tail data points, in mm, which can be obtained through the system. (d) i This represents the difference between adjacent data points, in mm.

[0089] The system derives the head from d2 to d 10 The data is shown in Table 1, which is a table of differences between adjacent data points at the head.

[0090] Table 1. Difference between adjacent data points at the head

[0091]

[0092] The system derives the tail portion from d2 to d 10 The data is shown in Table 2, which is a table of differences between adjacent data points at the tail.

[0093] Table 2. Difference between adjacent data points at the tail end

[0094]

[0095] Define symbolic functions:

[0096] ,

[0097] The header symbols are as follows: 1, 1, -1, -1, -1, -1, -1, -1, -1, -1,

[0098] The trailing symbols are: 1, 1, 1, 1, 1, 1, 1, 1, 1, 1.

[0099] Introducing a trend strength coefficient:

[0100] ,

[0101] Where T is the trend strength coefficient, which is dimensionless and is obtained through calculation.

[0102] Head trend strength coefficient: ,

[0103] Tail trend strength coefficient: ,

[0104] Define trend significance indicators:

[0105] ,

[0106] Where S is the trend significance index coefficient, which is dimensionless and calculated; T is the trend strength coefficient, which is dimensionless and calculated. This represents the difference between adjacent data points, expressed in mm, and is obtained through calculation.

[0107] Head indicator: There is a significant trend.

[0108] Tail indicators: There is a significant trend.

[0109] when ( The significance threshold is a dimensionless value, obtained from rolling history experience, and is set to 0.8. When a significant trend is determined, the trend direction is determined according to the sign of T (positive sign indicates positive skewness, negative sign indicates negative skewness).

[0110] Simultaneously, calculate the trend stability coefficient:

[0111] ,

[0112] Where α is the trend stability coefficient, dimensionless, obtained through calculation; λ is the attenuation factor, dimensionless, obtained from historical rolling experience, and is set to 0.1. ( The stability threshold is dimensionless and obtained through rolling history experience. When it is set to 0.7, the reliability of the trend is further confirmed.

[0113] Head trend stability coefficient: The trend is reliable;

[0114] Tail trend stability coefficient: The trend is reliable.

[0115] Step 3: Dynamic calculation of adjustment amount:

[0116] When a significant trend exists, calculate the difference between the mean of the 10th subarray and the 1st subarray.

[0117] , ,

[0118] Define the function for calculating the adjustment amount:

[0119] ,

[0120] in, is the difference between the mean of the nth subarray and the 1st subarray, in mm, obtained through calculation; T is the trend strength coefficient, dimensionless, obtained through step 2; k is the control coefficient, dimensionless, obtained through the system server, and is set to 0.8.

[0121] Calculate the adjustment amount:

[0122] ,

[0123] ,

[0124] Introduce a regulation stability correction factor:

[0125] ,

[0126] ,

[0127] The average of the head adjustment and the tail adjustment is taken as the final adjustment.

[0128] ,

[0129] Step 4: Distribute self-learning data based on site conditions and rolling data:

[0130] ,

[0131] ,

[0132] in, For the nth volume of steel The self-learning adjustment amount of the frame (such as rolling force or roll gap adjustment), in mm, is obtained through calculation; The heritability coefficient represents... The transfer ratio of the historical adjustment amount of the rack itself is dimensionless and obtained through finite element simulation, and is taken as 0.8; The attenuation coefficient represents... The decay rate of the historical adjustment amount, dimensionless, was obtained through finite element simulation and is taken as 0.9; Δ (n-1) represents the adjustment amount of the (n-1)th coil of steel on the F1 stand. It is dimensionless and obtained through rolling history data, and is taken as 0.02. For the nth volume of steel The self-learning adjustment amount of the frame (such as rolling force or roll gap adjustment), in mm, is obtained through calculation; The heritability coefficient represents... The transfer ratio of the rack's own historical adjustment amount, dimensionless, was obtained through finite element simulation and is taken as 0.8; D 22 The attenuation coefficient represents... The decay rate of the historical adjustment amount, dimensionless, was obtained through finite element simulation and is taken as 0.8; Δ (n-1) represents the adjustment amount of the (n-1)th coil of steel on the F2 stand. It is dimensionless and obtained through rolling history data, and is taken as 0.01. The heritability coefficient represents... rack pair The frame's transfer ratio, dimensionless, was obtained through finite element simulation and is taken as 0.75; D21 is the attenuation coefficient, representing... rack pair The attenuation ratio of the rack is dimensionless and obtained through finite element simulation, and is taken as 0.8.

[0133] The F1 self-learning adjustment amount is calculated as follows:

[0134] ,

[0135] .

[0136] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for dynamic correction of pre-oscillating roll gap for strip misalignment at the head and tail, characterized in that, include: Obtain strip data, and preprocess and segment the head and tail data; Identify deviation trends in the head data and the tail data; The pre-sway roller gap adjustment amount is dynamically calculated to compensate for the deviation trend; The pre-sway roll gap adjustment amount is sent to the rolling mill control system to dynamically correct the pre-sway roll gap setting value for the current rolling pass. Based on the rolling results and on-site conditions of the current rolling pass, the adjustment amount is optimized through self-learning to be used for pre-sway roll gap correction under subsequent similar rolling conditions. The acquisition of strip data, including preprocessing of the head and tail data, includes: Obtain the sequence of strip data, filter out data points with the same deviation value at the head and tail, and remove a preset number of data points from the head and tail data. The segmentation of the header and tail data includes: The first 10% to 20% of the header data is divided into n consecutive subarrays, and the last 10% to 20% of the tail data is divided into n consecutive subarrays. The identification of deviation trends in the head data and the tail data includes: Calculate the difference of adjacent data points: d i = a i - a i-1 , i = 2, 3,... n; wherein a1, a2,... a n is the head or tail data point, in mm, d i is the difference between adjacent data points, in mm; The identification of deviation trends for the head data and the tail data also includes: Calculate the sign function: ; Calculate the trend strength coefficient: ; Calculate the trend significance index: ; When S≥S th When a significant trend is identified, the trend direction is determined based on the sign of the trend strength coefficient T. When the trend strength coefficient T is positive, it is a positive skew; when the trend strength coefficient T is negative, it is a negative skew. The identification of deviation trends for the head data and the tail data also includes: Calculate the trend stability coefficient: ; Where α is the trend stability coefficient, dimensionless, and λ is the decay factor, dimensionless, when hour, The stability threshold is dimensionless and confirms the reliability of the trend. The dynamic calculation and compensation of the pre-sway roll gap adjustment includes: Calculate the difference between the mean of the nth subarray and the first subarray: , where μ n μ1 is the mean of the data points in the nth subarray, and μ1 is the mean of the data points in the 1st subarray. Calculation function using pre-swing roll gap adjustment amount: ； in, The unit is mm; k is the control coefficient, which is dimensionless. Calculate the pre-swing roller gap adjustment amount ; The dynamic calculation and compensation of the pre-sway roll gap adjustment also includes: The stability of the pre-swing roll gap adjustment amount is corrected using the following formulas to obtain the corrected head pre-swing roll gap adjustment amount △corrected,head and the corrected tail pre-swing roll gap adjustment amount △corrected,tail: ； The average of the corrected head pre-swing roll gap adjustment Δcorrected,head and the corrected tail pre-swing roll gap adjustment Δcorrected,tail is used as the final pre-swing roll gap adjustment. ； in, The final pre-sway roll gap adjustment amount is expressed in mm.

2. The method for dynamic correction of pre-oscillating roll gap for strip head and tail deviation characteristics according to claim 1, characterized in that, The self-learning optimization of the adjustment amount based on the rolling results and on-site conditions of the current rolling pass for pre-sway roll gap correction under similar rolling conditions includes: The self-learning adjustment amount is calculated based on the following formula: ; ； in, For the nth volume of tape in The self-learning adjustment range of the rack, in mm; The heritability coefficient represents... The transfer ratio of the rack's own historical adjustment amounts is dimensionless; D 11 The attenuation coefficient represents... The decay rate of historical adjustment amounts is dimensionless; Δ (n−1) represents the adjustment amount of the (n−1)th roll of strip on the F1 stand, which is dimensionless; For the nth volume of tape in The self-learning adjustment range of the rack, in mm; The heritability coefficient represents... The transfer ratio of the rack's own historical adjustment amounts is dimensionless; D 22 The attenuation coefficient represents... The decay rate of historical adjustment amounts, dimensionless; The heritability coefficient represents... rack pair The transfer ratio of the rack is dimensionless; D 21 The attenuation coefficient represents... rack pair The attenuation ratio of the rack is dimensionless.

3. The method for dynamic correction of pre-oscillating roll gap for strip head and tail deviation characteristics according to any one of claims 1-2, characterized in that, The strip is made of titanium, stainless steel, copper, aluminum or nickel-based alloy.

Images