Strip shape defect feature recognition and quantitative control method based on 26-zone sensor

Abstract

The invention relates to the technical field of reversible cold-rolled sheet strip shape control of a 20-roller multi-roller cold-rolling mill, in particular to a strip shape defect feature recognition and quantitative control method based on a 26-area sensor, which comprises the following steps: dividing the 26-area strip shape sensor into five functional areas: a left area 1-3, a left 1 / 4 area 6-9, a middle area 12-15, a right 1 / 4 area 18-21 and a right area 24-26; the method comprises the following steps of: setting acquisition periods, acquiring sensor data of each area in real time in each acquisition period, calculating a middle wave characteristic value delta Icenter, a bilateral wave characteristic value delta Iedge, a unilateral wave characteristic value delta Iasym, a left 1 / 4 wave characteristic value delta IQ1 and a right 1 / 4 wave characteristic value delta IQ3, comparing each characteristic value with a corresponding preset threshold value, and judging whether a defect exists or not.

Description

Technical Field

[0001] This invention relates to the field of reversible cold-rolled strip shape control technology for 20-roll multi-roll cold rolling mills, and particularly to a method for identifying and quantifying strip shape defect features based on a 26-zone sensor. Background Technology

[0002] Existing 20-roll multi-roll cold rolling mill shape control systems face several technical challenges in the ultra-thin precision strip rolling process. Firstly, optimizing multi-objective conflicts is difficult. Traditional PID control systems can only handle single-objective control, making it challenging to achieve an effective balance between multiple interdependent objectives such as shape accuracy control, rolling force stability, surface quality assurance, and energy consumption optimization. Shape accuracy needs to be controlled within ±2-3I, while ensuring rolling force fluctuations are within a reasonable range, maintaining stable surface roughness Ra values, and optimizing energy consumption while ensuring quality.

[0003] Currently, industrial plate shape analyzers typically employ 20-30 measurement zones, with a 26-zone configuration being the most common. However, existing technology lacks a clear technical solution for systematically identifying and quantitatively controlling defects using data from these 26-zone sensors. Faced with complex plate shape curves, operators often struggle to quickly and accurately determine defect types and adjustment amounts, resulting in inconsistent control performance.

[0004] Secondly, there are unique challenges in controlling small-diameter work rolls. The work rolls of a 20-roll mill have a diameter of only 21-45mm, smaller than those of traditional mills. This results in a relatively large elastic flattening value, directly affecting the accuracy of strip shape control. Small-diameter work rolls wear relatively quickly, requiring frequent adjustments to control parameters, and are highly sensitive to pressure changes, necessitating improvements in system control stability.

[0005] Thirdly, there are technological challenges in identifying and controlling plate shape defects. The defect identification system based on the 26-point plate shape detector has difficulties in handling four typical defects: the middle wave is manifested as an abnormality in the sensor 12-15 area, which requires appropriate adjustment of the rolling force; the double-sided wave is manifested as an abnormality in the sensor 1-3 and 24-26 areas, which requires appropriate adjustment of the bending roll force; the single-sided wave is manifested as excessive asymmetry, which requires adjustment in conjunction with tilting rolls and cross rolls; the 1 / 4 wave is manifested as an abnormality in the sensor 6-9 and 18-21 areas, which requires adjustment of the intermediate roll displacement.

[0006] The limitations of existing technical systems are mainly reflected in: Defect identification lacks quantitative standards: Existing technologies mostly use qualitative descriptions (such as "middle wave" and "edge wave"), lacking quantitative identification formulas based on sensor data. Operators need to rely on visual observation of the plate shape curve to determine the defect type, with an accuracy rate of about 80% and a response time of 3-5 seconds.

[0007] The lack of clear formulas for control adjustments: Existing technologies fail to provide specific control formulas and parameters for identified plate shape defects. For example, there are no clear guidelines on how many kN the bending roll force or the rolling force should be adjusted after discovering intermediate waviness. The adjustment range can vary by up to ±50% between different operators.

[0008] The sensor zoning function is unclear: among the 26-zone sensor, which zones are used to monitor edge waves, which zones are used to monitor middle waves, and which zones are used to monitor quarter waves are not systematically defined in the current technology.

[0009] Complex defects are difficult to handle: When complex defects such as intermediate waves and single-sided waves occur simultaneously, existing technologies lack systematic identification methods and coordinated control strategies, requiring repeated trial and error, with a success rate of only about 40% for a single correction.

[0010] Lack of parameter self-optimization mechanism: The influence coefficient and control threshold are usually fixed and cannot be automatically optimized according to the actual control effect. Summary of the Invention

[0011] The technical problem to be solved by this invention is: how to construct a multi-objective collaborative optimization control system suitable for the characteristics of small-diameter work rolls in 20-roll mills, to achieve intelligent balance of plate shape accuracy, rolling force, surface quality and energy consumption efficiency, and to establish an effective identification and control mechanism for defects of four typical plate shapes (middle wave, double wave, single wave and 1 / 4 wave).

[0012] The technical solution adopted in this invention is as follows: A 26-zone plate sensor is divided into five functional areas: the left side (zones 1-3), the left quarter (zones 6-9), the middle area (zones 12-15), the right quarter (zones 18-21), and the right side (zones 24-26) (each area corresponds to one sensor, such as a laser sensor). A data acquisition cycle is set, and in each acquisition cycle, sensor data for each area is acquired in real time (normally once per cycle). The characteristic values ​​of the middle wave (ΔI_center), the bilateral wave (ΔI_edge), the single-sided wave (ΔI_asym), the left quarter wave (ΔI_Q1), and the right quarter wave (ΔI_Q3) are calculated. Each characteristic value is compared with its corresponding preset threshold to determine whether a defect exists. If no defect exists, wait for the next acquisition cycle to perform the same judgment. If a defect exists, determine the defect type, select the corresponding adjustment method and adjustment formula according to the defect type, calculate the adjustment amount, and output control commands to adjust according to the adjustment amount. After each adjustment output control command, wait for a set time (e.g., 500ms), recalculate each feature value, and compare each feature value with a preset threshold. If the same defect type still exists, perform iterative correction. For the same defect type, the number of consecutive corrections is less than or equal to 5. For each correction, the adjustment method cannot exceed the corresponding single adjustment range, and for consecutive corrections, the adjustment method cannot exceed the cumulative adjustment range. For each adjustment, record the relevant parameters of each adjustment process.

[0013] The 26-zone plate-shaped sensors are divided into five functional areas: Left side: Sensors 1-3, monitoring the left side waves; Left 1 / 4 zone: Sensors 6-9, monitoring the left 1 / 4 waves; Middle zone: Sensors 12-15, monitoring the middle waves; Right 1 / 4 zone: Sensors 18-21, monitoring the right 1 / 4 waves; Right side: Sensors 24-26, monitoring the right side waves; Transition zone: Sensors 4-5, 10-11, 16-17, and 22-23, for auxiliary judgment.

[0014] When the feature value is within the range of the corresponding preset threshold to twice the corresponding preset threshold (including twice the corresponding preset threshold), the severity is mild; when the feature value is within the range of twice the corresponding preset threshold to three times the corresponding preset threshold (including three times the corresponding preset threshold), the severity is moderate; when the feature value is above three times the corresponding preset threshold, the severity is severe.

[0015] When ΔI_center > 3 (I-units), it is determined to be a central wave defect, where ΔI_center = I_center - I_edge, I_center = (I_12 + I_13 + I_14 + I_15) / 4, I_edge = (I_1 + I_2 + I_3 + I_24 + I_25 + I_26) / 6, I_center is the average value of the central region, I_edge is the average value of the edge region, I_n is the value of the nth region, n is the sensor number, which is a natural number from 1 to 26, and I-units is the plate shape defect quantification index with a value of 10. 5 ΔI_center>0 indicates that the extension of the middle part is relatively large compared to the edge, showing a middle wave; ΔI_center<0 indicates that the extension of the middle part is relatively small compared to the edge, showing a tight middle wave (in this case, it is treated as a two-sided wave).

[0016] When ΔI_edge > 3 (I-units) and ΔI_symmetry < 4 (I-units), it is judged as a double-sided wave defect, where ΔI_edge = I_edge_avg - I_center, ΔI_symmetry = |I_left - I_right|, I_edge_avg = (I_left + I_right) / 2, I_left = (I_1 + I_2 + I_3) / 3, I_right = (I_24 + I_25 + I_26) / 3, ΔI_symmetry is the edge symmetry judgment value, I_edge_avg is the comprehensive average value of the edge, I_left is the average value of the left side, and I_right is the average value of the right side; When -ΔI_asym>+4, it is determined to be a left-side single-sided wave defect, where ΔI_asym=I_left-I_right; When -ΔI_asym < -4, it is determined to be a right-side single-sided wave defect; When -ΔI_Q1>4, it is determined to be a defect in the left 1 / 4 wave, where ΔI_Q1=I_Q1-I_ref_left, I_Q1=(I_6+I_7+I_8+I_9) / 4, I_ref_left=(I_left+I_center) / 2, where I_Q1 is the average value of the left 1 / 4 zone, and I_ref_left is the reference value of the left 1 / 4 zone; When -ΔI_Q3>4, it is determined to be a right 1 / 4 wave defect, where ΔI_Q3=I_Q3-I_ref_right, I_Q3=(I_18+I_19+I_20+I_21) / 4, I_ref_right=(I_right+I_center) / 2, where I_Q3 is the average value of the right 1 / 4 zone, and I_ref_right is the reference value of the right 1 / 4 zone; When two or more defect criteria are met simultaneously, it is judged as a composite defect. Composite defects include intermediate wave + unilateral wave defect, bilateral wave + unilateral wave defect, 1 / 4 wave + intermediate wave defect, and 1 / 4 wave + unilateral wave defect.

[0017] When the defect is in the middle wave, the adjustment method is to adjust the bending roll force. The bending roll force adjustment formula is: ΔF_bend=-K_bend×K_steel×ΔI_center×B×h, where K_bend is the bending roll force influence coefficient, with a value range of 8-12kN / [(I-unit)*m*mm)], K_steel is the steel grade correction coefficient, B is twice the strip width, and h is the strip thickness; When there is a double-sided wave defect, the adjustment method is to first adjust the bending roll force, and then adjust the rolling force. The bending roll force adjustment formula is: ΔF_bend=K_bend×K_steel×ΔI_edge×B×h, and the rolling force adjustment formula is: ΔF_roll=-K_roll×K_steel×ΔI_edge×B, where ΔF_bend is the increase in bending roll force, K_steel is the steel grade correction coefficient, B is twice the strip width, and h is the strip thickness. When the defect is a single-sided wave defect, i.e., a left-sided wave defect or a right-sided wave defect, the adjustment method is to first adjust the tilting roller angle, and then adjust the cross-roller displacement. The tilting roller angle adjustment formula is Δθ_titl=K_titl×K_steel×ΔI_asym÷B, and the cross-roller displacement adjustment formula is: ΔS_shift=-K_shift×K_steel×ΔI_asym, where Δθ_titl is the tilting angle, K_titl is the tilting roller influence coefficient, ΔS_shift is the cross-roller displacement, and K_shift is the cross-roller influence coefficient. When the defect is a 1 / 4 wave defect, i.e., a left 1 / 4 wave defect or a right 1 / 4 wave defect, the adjustment method is to first adjust the displacement of the intermediate roller, and then perform local cooling control. For a left 1 / 4 wave defect, the intermediate roller displacement adjustment formula is ΔS_IR=-K_IR×K_steel×ΔI_Q1, and the local cooling control formula is ΔQ_cool=K_cool×K_steel×ΔI_Q1, where ΔS_IR is the displacement of the intermediate roller for the left 1 / 4 wave defect, and K_IR is the influence coefficient of the intermediate roller for the left 1 / 4 wave defect. For a right 1 / 4 wave defect, the intermediate roller displacement adjustment formula is ΔS_IL=-K_IL×K_steel×ΔI_Q3, where ΔS_IL is the displacement of the intermediate roller for the right 1 / 4 wave defect, and K_IL is the influence coefficient of the intermediate roller for the right 1 / 4 wave defect. For composite defects, each defect is identified, and different adjustment types are adjusted simultaneously. For the same adjustment type, adjustments are made sequentially according to the size of the eigenvalue. Each adjustment is multiplied by the coupling correction coefficient K_coupling, which ranges from 0.80 to 0.95.

[0018] The correction factor for ordinary austenitic steel is 1.00, the correction factor for molybdenum-containing austenitic steel is 1.05-1.08, and the correction factor for ferritic steel is 0.88-0.92. Detailed Implementation

[0019] This invention employs a 26-zone plate-shaped sensor, evenly distributed along the width of the strip. Taking a 650mm bandwidth as an example, the sensor spacing is approximately 52mm.

[0020] The sensor functional zones are shown in the table below:

[0021] Explanation of zoning principles: Three sensors were used in each of the edge regions, covering approximately 12% of the bandwidth, which is sufficient to reflect the shape characteristics of the edge plate. Four sensors were used in each of the 1 / 4 regions, covering approximately 15% of the bandwidth, to accurately capture the characteristics of the 1 / 4 wave. Four sensors were used in the central region, covering approximately 15% of the bandwidth, to accurately characterize the intensity of the waves in the middle. Transition zone sensors are used to assist in judgment and trend analysis, but do not directly participate in feature value calculation.

[0022] Eigenvalue calculation formula For the middle wave Average value for the central region: I_center=(I_12+I_13+I_14+I_15) / 4; Average value of the peripheral region: I_edge=(I_1+I_2+I_3+I_24+I_25+I_26) / 6; intermediate wave eigenvalues: ΔI_center = I_center - I_edge; Physical meaning: ΔI_center>0 indicates that the extension of the middle part is relatively large compared to the edge part, presenting a wave-like shape in the middle; ΔI_center<0 indicates that the extension of the middle part is relatively small compared to the edge part, resulting in a tight middle (in this case, it is treated as a double-sided wave).

[0023] bilateral wave eigenvalues Left side average: I_left = (I_1 + I_2 + I_3) / 3; Right side average: I_right=(I_24+I_25+I_26) / 3; Average value of the edges: I_edge_avg=(I_left+I_right) / 2; bilateral wave eigenvalues: ΔI_edge = I_edge_avg - I_center Edge symmetry: ΔI_symmetry=|I_left-I_right|; Judgment criteria: When ΔI_edge>3 and ΔI_symmetry<4, it is judged as a two-sided wave.

[0024] For unilateral waves Unilateral wave characteristic values: ΔI_asym = I_left - I_right; Judgment rules: -ΔI_asym>+4: Left-sided wave (left side extends more than right side); -ΔI_asym<-4: Right-side single wave (the right side extends more than the left side).

[0025] For the 1 / 4 wave characteristic Average value of the left 1 / 4 zone: I_Q1=(I_6+I_7+I_8+I_9) / 4; Average value of the right 1 / 4 zone: I_Q3=(I_18+I_19+I_20+I_21) / 4; Reference baseline value: I_ref_left=(I_left+I_center) / 2; I_ref_right=(I_right+I_center) / 2; 1 / 4 wave characteristic value: ΔI_Q1 = I_Q1 - I_ref_left; ΔI_Q3 = I_Q3 - I_ref_right; Judgment rules: -ΔI_Q1>4: Left 1 / 4 wave; -ΔI_Q3>4: Right 1 / 4 wave.

[0026] Determination of composite defects When two or more defect criteria are met simultaneously, it is determined to be a composite defect.

[0027] Common types of composite defects: Intermediate wave + unilateral wave, double-sided wave + unilateral wave, 1 / 4 wave + intermediate wave, 1 / 4 wave + unilateral wave.

[0028] Limiting rules If the calculated adjustment exceeds the single limit, the limit value (set value) shall be used. If the cumulative adjustments reach the limit, the adjustment of that institution will be stopped, and a backup institution will be activated. An alarm will be triggered if any institution reaches its cumulative limit.

[0029] Control cycle setting: Data acquisition: 100ms Feature calculation: 50ms Decision making: 50ms Command output: 100ms Effect evaluation: 500ms Full cycle: 800ms Iterative convergence criterion: Convergence condition: All feature values ​​< trigger threshold, or two consecutive changes < 0.5 (I-units); Divergence condition: The system fails to converge after 5 iterations, or the eigenvalues ​​increase in the opposite direction. Divergence processing: Stop automatic control and trigger an alarm; Lock the current location of the executing agency; The system prompts for manual inspection of process parameters or roller condition.

[0030] Learning Objectives By analyzing historical control data, the influence coefficient K and trigger threshold are optimized.

[0031] Record each control process: Control the front panel shape data and characteristic values; Control the action and adjust the amount; Control the back plate shape data and feature values; Evaluation of control effectiveness.

[0032] Actuator (Adjustment Method) Coordination Control Parameter Specification

[0033] Mid-wave control example Process conditions: Steel grade: 304, thickness: 0.12mm, width: 675mm, rolling speed: 7.5m / s.

[0034] Plate shape data (unit: I-units):

[0035] Recognition calculation: I_center=(I_12+I_13+I_14+I_15) / 4=(8+9+10+9) / 4=9.0 (I-units); I_edge=(I_1+I_2+I_3+I_24+I_25+I_26) / 6=(2+1+3+1+2+1) / 6≈1.67 (I-units); ΔI_center=I_center-I_edge=9.0-1.67=7.33 (I-units).

[0036] Judgment result: ΔI_center=7.33 (I-units), which is in the range of 6-10 and is judged as a moderate intermediate wave.

[0037] When the defect is in the middle wave, the adjustment method is to adjust the bending roll force. The bending roll force adjustment formula is: ΔF_bend=-K_bend×K_steel×ΔI_center×B×h, where K_bend is the bending roll force influence coefficient, with a value range of 8-12kN / [(I-unit)*m*mm)], K_steel is the steel grade correction coefficient, B is twice the strip width, and h is the strip thickness; From the table, the steel grade correction factor K_steel is 1.

[0038] F_bend=-K_bend×K_steel×ΔI_center×B×h=-10×1×7.33×1.35×0.12=-11.87kN; Adjustment action: Adjust the bending roller force by 12kN (approximate value) in the direction of the negative bending roller.

[0039] Effect verification: After adjustment, the plate shape data was re-collected:

[0040] ΔI_center = 2.5 (I-units), 2.5 < 3, the intermediate wave is eliminated, and the control is successful.

[0041] Example of bilateral wave control Process conditions: Steel grade: 316L, thickness: 0.15 mm, width: 625 mm, rolling speed: 6.0 m / s.

[0042] Plate shape data (unit: I-units):

[0043] Recognition calculation: When there is a double-sided wave defect, the adjustment method is to first adjust the bending roll force, and then adjust the rolling force. The bending roll force adjustment formula is: ΔF_bend=K_bend×K_steel×ΔI_edge×B×h, and the rolling force adjustment formula is: ΔF_roll=-K_roll×K_steel×ΔI_edge×B, where ΔF_bend is the increase in bending roll force, K_steel is the steel grade correction coefficient, B is twice the strip width, and h is the strip thickness.

[0044] I_left=(I_1+I_2+I_3) / 3=9.0 (I-units); I_right=(I_24+I_25+I_26) / 3=9.0 (I-units); I_edge_avg=(I_left+I_right) / 2=9.0 (I-units); I_center = 1.5 (I-units); ΔI_edge=I_edge_avg-I_center=7.5 (I-units); ΔI_symmetry=|I_left-I_right|=0; Moderate bilateral wave.

[0045] Control calculation: Step 1: Adjusting the bending roller force From the table, the correction factor K_steel for 316L steel is 1.05. ΔF_bend=K_bend×K_steel×ΔI_edge×B×h=10×1.05×7.5×1.25×0.15≈15kN.

[0046] Adjustment action: Adjust the bending roller force by 15 kN in the direction of the positive bending roller.

[0047] Effect verification: After adjustment, the plate shape data was re-collected:

[0048] ΔI_edge = 3.08 (I-units).

[0049] Judgment result: ΔI_edge is still slightly higher than the threshold of 3, and further adjustment is needed.

[0050] ΔF_roll=-K_roll×K_steel×ΔI_edge×B=-65×1.05×3.08×1.25≈270kN; Since the single-cycle limit of 200 kN is exceeded, the actual adjustment is -200 kN.

[0051] Adjustment action: Reduce rolling force by 200 kN.

[0052] Effect verification: After adjustment, the plate shape data was re-collected:

[0053] ΔI_edge = 1.08 (I-units); ΔI_symmetry==0; Judgment result: Both sides of the wave were eliminated, and control was successful.

[0054] Single-sided wave control example Process conditions: Steel grade: 301, thickness: 0.10 mm, width: 1400 mm, rolling speed: 8.0 m / s.

[0055] Plate shape data (unit: I-units):

[0056] Recognition calculation: I_left=(I_1+I_2+I_3) / 3=(11+12+10) / 3=11(I-units); I_right=(I_24+I_25+I_26) / 3=(1+2+2) / 3≈1.67 (I-units); ΔI_asym=I_left-I_right=11.0-1.67=9.33 (I-units); Judgment result: ΔI_asym > 4 and is a positive value, so it is judged as a moderate left-side unilateral wave (8-12 range).

[0057] Control calculation: When a single-sided wave defect occurs, i.e., a left-sided or right-sided wave defect, the adjustment method is to first adjust the tilting roller angle, and then adjust the cross-roller displacement. The tilting roller angle adjustment formula is Δθ_titl=K_titl×K_steel×ΔI_asym÷B, and the cross-roller displacement adjustment formula is: ΔS_shift=-K_shift×K_steel×ΔI_asym, where Δθ_titl is the tilting angle, K_titl is the tilting roller influence coefficient, ΔS_shift is the cross-roller displacement, and K_shift is the cross-roller influence coefficient. Step 1: Adjusting the tilting roller angle

[0058] Δθ_titl=K_titl×K_steel×ΔI_asym÷B=0.020×1.0×9.33÷1.40=0.133mrad Since the single-transaction limit was exceeded by 0.05 mrad, the actual adjustment was +0.05 mrad.

[0059] Adjustment action: Adjust the tilting roller angle by 0.05 mrad in the downward pressing direction on the operating side.

[0060] Effect verification: After adjustment, the plate shape data was re-collected:

[0061] I_left ≈ 7.33 (I-units); I_right≈2.33 (I-units); ΔI_asym=I_left-I_right=5.0 (I-units); Result: ΔI_asym is still higher than the threshold of 4 and needs further adjustment.

[0062] Step 2: Adjusting the displacement of the rollers

[0063] ΔS_shift=-K_shift×K_steel×ΔI_asym=-1.2×1.0×5.0=-6.0mm Due to exceeding the single-time limit of -5 mm, the actual adjustment is -5 mm.

[0064] Adjustment action: Move the work roller 5 mm toward the drive side.

[0065] Effect verification: After adjustment, the plate shape data was re-collected:

[0066] I_left ≈ 4.33 (I-units); I_right=3.0 (I-units); ΔI_asym=I_left-I_right=1.33 (I-units); Judgment result: ΔI_asym < 4, unilateral wave elimination, control successful.

[0067] 1 / 4 wave control example Process conditions: Steel grade: 430, thickness: 0.18 mm, width: 600 mm, rolling speed: 5.5 m / s.

[0068] Plate shape data (unit: I-units):

[0069] Recognition calculation: I_left=(I_1+I_2+I_3) / 3≈2.33 (I-units); I_right=(I_24+I_25+I_26) / 3≈2.33 (I-units); I_center=(I_12+I_13+I_14+I_15) / 4=2.5 (I-units); I_Q1=(I_6+I_7+I_8+I_9) / 4=9.75 (I-units); I_Q3=(I_18+I_19+I_20+I_21) / 4=3.5 (I-units); I_ref_left=(I_left+I_center) / 2=2.415(I-units); I_ref_right=(I_right+I_center) / 2=2.415(I-units); ΔI_Q1=I_Q1-I_ref_left=7.335; ΔI_Q3=I_Q3-I_ref_right=1.085; Judgment result: ΔI_Q1>4, ΔI_Q3<4, judged as a moderate left 1 / 4 wave (7-11 interval).

[0070] Control calculation: When the defect is a left 1 / 4 wave, the formula for adjusting the intermediate roller displacement is ΔS_IR=-K_IR×K_steel×ΔI_Q1, and the formula for local cooling control is ΔQ_cool=K_cool×K_steel×ΔI_Q1, where ΔS_IR is the displacement of the intermediate roller for the left 1 / 4 wave, and K_IR is the influence coefficient of the intermediate roller for the left 1 / 4 wave.

[0071] Step 1: Adjusting the displacement of the intermediate roller From the table, the intermediate roll coefficient for 430 ferritic steel is 1.0.

[0072] ΔS_IR=-K_IR×K_steel×ΔI_Q=-0.8×7.335=-5.87mm Due to exceeding the single-time limit of -3 mm, the actual adjustment is -3 mm.

[0073] Adjustment action: Move the intermediate roller 3 mm toward the drive side.

[0074] Effect verification: After adjustment, the plate shape data was re-collected:

[0075] I_Q1=(I_6+I_7+I_8+I_9) / 4=6.5 (I-units); ΔI_Q1=I_Q1-I_ref_left=4.085 (I-units); Judgment result: ΔI_Q1 is still slightly higher than the threshold of 4, and further adjustment is needed.

[0076] Step 2: Localized Cooling Control

[0077] ΔQ_cool=K_cool×K_steel×ΔI_Q1=8.17L / min Adjustment action: Increase the flow rate in cooling zone 3-4 by 8 L / min.

[0078] Effect verification: After adjustment, re-collect plate shape data (cooling adjustments will take 10-30 seconds to take effect):

[0079] I_Q1=(I_6+I_7+I_8+I_9) / 4=4.5 (I-units); ΔI_Q1=I_Q1-I_ref_left=42.085 (I-units); Judgment result: ΔI_Q1 < 4, 1 / 4 wave eliminated, control successful.

Claims

1. A method for identifying and quantifying plate shape defect features based on a 26-zone sensor, characterized in that: The 26-zone plate sensor is divided into five functional areas: left side (zones 1-3), left quarter (zones 6-9), middle (zones 12-15), right quarter (zones 18-21), and right side (zones 24-26). A data acquisition cycle is set, and in each cycle, sensor data for each area is collected in real time. The characteristic values ​​ΔI_center (center wave), ΔI_edge (double-sided wave), ΔI_asym (single-sided wave), ΔI_Q1 (left quarter wave), and ΔI_Q3 (right quarter wave) are calculated. Each characteristic value is compared with its corresponding preset threshold to determine if a defect exists. If no defect is found... Then wait for the next acquisition cycle to perform the same judgment. If a defect exists, determine the defect type, select the corresponding adjustment method and adjustment formula according to the defect type, calculate the adjustment amount, and output control instructions to adjust according to the adjustment amount. After each adjustment output control instruction, wait for a set time, recalculate each feature value, and compare each feature value with the preset threshold. If the same defect type still exists, perform iterative correction. For the same defect type, the number of consecutive corrections is less than or equal to 5. For each correction, the adjustment method cannot exceed the corresponding single adjustment range. For consecutive corrections, the adjustment method cannot exceed the cumulative adjustment range.

2. The control method according to claim 1, characterized in that: When ΔI_center > 3 (I-units), it is determined to be a central wave defect, where ΔI_center = I_center - I_edge, I_center = (I_12 + I_13 + I_14 + I_15) / 4, I_edge = (I_1 + I_2 + I_3 + I_24 + I_25 + I_26) / 6, I_center is the average value of the central region, I_edge is the average value of the edge region, I_n is the value of the nth region, n is the sensor number, which is a natural number from 1 to 26, and I-units is the plate shape defect quantification index with a value of 10. 5 ; When ΔI_edge > 3 (I-units) and ΔI_symmetry < 4 (I-units), it is judged as a double-sided wave defect, where ΔI_edge = I_edge_avg - I_center, ΔI_symmetry = |I_left - I_right|, I_edge_avg = (I_left + I_right) / 2, I_left = (I_1 + I_2 + I_3) / 3, I_right = (I_24 + I_25 + I_26) / 3, ΔI_symmetry is the edge symmetry judgment value, I_edge_avg is the comprehensive average value of the edge, I_left is the average value of the left side, and I_right is the average value of the right side; When -ΔI_asym>+4, it is determined to be a left-side single-sided wave defect, where ΔI_asym=I_left-I_right; When -ΔI_asym < -4, it is determined to be a right-side single-sided wave defect; When -ΔI_Q1>4, it is determined to be a defect in the left 1 / 4 wave, where ΔI_Q1=I_Q1-I_ref_left, I_Q1=(I_6+I_7+I_8+I_9) / 4, I_ref_left=(I_left+I_center) / 2, where I_Q1 is the average value of the left 1 / 4 zone, and I_ref_left is the reference value of the left 1 / 4 zone; When -ΔI_Q3>4, it is determined to be a right 1 / 4 wave defect, where ΔI_Q3=I_Q3-I_ref_right, I_Q3=(I_18+I_19+I_20+I_21) / 4, I_ref_right=(I_right+I_center) / 2, where I_Q3 is the average value of the right 1 / 4 zone, and I_ref_right is the reference value of the right 1 / 4 zone; When two or more defect criteria are met simultaneously, it is judged as a composite defect. Composite defects include intermediate wave + unilateral wave defect, bilateral wave + unilateral wave defect, 1 / 4 wave + intermediate wave defect, and 1 / 4 wave + unilateral wave defect.

3. The control method according to claim 2, characterized in that: When the defect is in the middle wave, the adjustment method is to adjust the bending roll force. The bending roll force adjustment formula is: ΔF_bend=-K_bend×K_steel×ΔI_center×B×h, where ΔF_bend is the increase in bending roll force, K_bend is the bending roll force influence coefficient, K_bend is in kN / [(I-unit)•m•mm)], and the value range is 8-12, K_steel is the steel grade correction coefficient, B is twice the strip width, and h is the strip thickness. When there is a double-sided wave defect, the adjustment method is to first adjust the bending roll force, and then adjust the rolling force. The bending roll force adjustment formula is: ΔF_bend=K_bend×K_steel×ΔI_edge×B×h, and the rolling force adjustment formula is: ΔF_roll=-K_roll×K_steel×ΔI_edge×B, where K_roll is in kN / [(I-unit)•m)], and the value range of K_roll is 50-80. K_steel is the steel grade correction coefficient, B is twice the strip width, and h is the strip thickness. When the defect is a single-sided wave defect, i.e., a left-sided wave defect or a right-sided wave defect, the adjustment method is to first adjust the tilting roller angle, and then adjust the cross-roller displacement. The tilting roller angle adjustment formula is Δθ_titl=K_titl×K_steel×ΔI_asym÷B, and the cross-roller displacement adjustment formula is: ΔS_shift=-K_shift×K_steel×ΔI_asym, where Δθ_titl is the tilting angle, K_titl is the tilting roller influence coefficient, the unit is mrad•m / (I-unit), and the value range is 0.015-0.025; ΔS_shift is the cross-roller displacement, K_shift is the cross-roller influence coefficient, K_shift is the mm / (I-unit), and the value range is 0.8-1.5; When the defect is a 1 / 4 wave defect, i.e., a left 1 / 4 wave defect or a right 1 / 4 wave defect, the adjustment method is to first adjust the displacement of the intermediate roller, and then perform local cooling control. For a left 1 / 4 wave defect, the intermediate roller displacement adjustment formula is ΔS_IR=-K_IR×K_steel×ΔI_Q1, and the local cooling control formula is ΔQ_cool=K_cool×K_steel×ΔI_Q1, where ΔS_IR is the displacement of the intermediate roller for the left 1 / 4 wave, K_IR is the influence coefficient of the intermediate roller for the left 1 / 4 wave, and the unit of K_IR is mm / (I-unit), with a value range of 0.6-1.

0. For a right 1 / 4 wave defect, the intermediate roller displacement adjustment formula is ΔS_IL=K_IL×K_steel×ΔI_Q3, where ΔS_IL is the displacement of the intermediate roller for the right 1 / 4 wave, K_IL is the influence coefficient of the intermediate roller for the right 1 / 4 wave, and the unit of K_IL is mm / (I-unit), with a value range of 0.6-1.

0. For composite defects, each defect is identified, and different adjustment types are adjusted simultaneously. For the same adjustment type, adjustments are made sequentially according to the size of the eigenvalue. Each adjustment is multiplied by the coupling correction coefficient K_coupling, which ranges from 0.80 to 0.

95.

4. The control method according to claim 3, characterized in that: The correction factor for ordinary austenitic steel is 1.00, the correction factor for molybdenum-containing austenitic steel is 1.05-1.08, and the correction factor for ferritic steel is 0.88-0.

92.

5. The control method according to claim 3, characterized in that: The single adjustment range of ΔF_bend is (-50kN, 50kN), and the cumulative adjustment range is (-300kN, 300kN); the single adjustment range of ΔF_roll is (-200kN, 200kN), and the cumulative adjustment range is (-15% of the set value, 15% of the set value); the single adjustment range of Δθ_titl is (-0.05mrad, 0.05mrad), and the cumulative adjustment range is (-0.20mrad, 0.20mrad); the single adjustment range of ΔS_shift is (-5mm, 5mm), and the cumulative adjustment range is (-20mm, 20mm); the single adjustment range of ΔS_IR is (-3mm, 3mm), and the cumulative adjustment range is (-15mm, 15mm).

Images