Visual correction inhibition system and method based on finish rolling wedge control

Abstract

The invention provides a vision deviation correction restraining system and method based on finish rolling wedge shape control, and aims to solve the problems of strip steel deviation and wedge shape defects in the hot continuous rolling finish rolling process. The system comprises a strip steel deviation detection device, an image processing server cluster, a PLC core control unit, an AGC control system and the like, the strip steel deviation amount is obtained in real time through non-contact high-precision visual detection, the roll gap deviation value of a downstream rack is calculated in combination with the rolling mill bounce principle, deviation correction and wedge-shaped control are executed in parallel, and the deviation correction accuracy is improved. And a tail predictive control and dynamic distribution coefficient optimization strategy is introduced to realize closed-loop feedforward control. The method comprises the steps of steel biting triggering, image acquisition and processing, deviation amount calculation, roll gap adjustment, wedge-shaped filtering, dead zone judgment and the like, can stably run in a high-temperature severe environment, effectively inhibits strip steel deviation and tail swinging, improves production stability, finished product quality and safety, and is suitable for the fields of steel and other industrial production.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology, specifically to a general-purpose high-temperature resistant device and control system for visual correction based on wedge control in steel mill finishing. Background Technology

[0002] In the finishing rolling process of hot-rolled thin strip steel in the steel industry, strip deviation and tailing have always been persistent problems plaguing the industry. Once out of control, the direct consequences include: strip deviation leading to steel pile-up and scrapping; tailing causing the strip tail to strike the work rolls, resulting in roll surface imprints, requiring unplanned roll replacements, increasing roll consumption and extending downtime, while also necessitating additional tail trimming, significantly reducing yield. Currently, the two commonly used countermeasures both have obvious shortcomings: Manual intervention: Operators visually inspect the strip shape or the difference in rolling force on the finishing mill stand, and then manually correct the roll gap deviation. Because the "observation → judgment → operation" chain is long and varies from person to person, the adjustment is lagging, the accuracy is low, the consistency is poor, and it is difficult to stably correct deviations.

[0003] Trend prediction: A few production lines attempted to use the changing trend of the rolling force difference between upstream stands to predict the potential force difference in downstream stands and automatically pre-adjust the roll gap accordingly. However, field data confirmed that changes in the rolling force difference could originate from either strip misalignment or strip wedge-shaped deformation; the symptoms are similar, but the mechanisms are opposite. If a wedge-shaped deformation is misjudged as strip misalignment, adjusting the roll gap in the opposite direction will only exacerbate the misalignment. Therefore, this approach has a high misjudgment rate and cannot reliably eliminate strip misalignment and tail-wagging.

[0004] In summary, existing technologies, whether relying on human experience or based on calculations of rolling force trends, are insufficient to achieve rapid, accurate, and stable strip centering control.

[0005] Therefore, there is an urgent need in the market for a visual correction and suppression system and method based on precision rolling wedge control that is not affected by external interference, can directly identify the true deviation direction of the strip steel and make immediate closed-loop corrections. Summary of the Invention

[0006] In view of the deficiencies in the prior art, the purpose of this invention is to provide a visual correction and suppression system and method based on precision rolling wedge control.

[0007] According to the present invention, a visual correction and suppression system based on fine rolling wedge control is provided, comprising: a strip deviation detection device, an image processing server cluster, a switch, a data logic server, a multi-functional control system, an L1 system, and an AGC control system; The strip deviation detection device includes at least one set of industrial area array cameras installed along the centerline of the rolling mill for acquiring images of the strip edge. The image processing server cluster corresponds one-to-one with the camera, performing image preprocessing and deviation calculation; The switch enables Ethernet communication between each image processing server and the data logic server; The data logic server receives the deviation data and calculates the roll gap deviation value of the downstream frame, achieves control interlock with the L1 system, and sends roll gap adjustment instructions to the AGC control system; The multi-functional control system continuously collects wedge-shaped data of the finished strip steel and feeds it back to the data logic server to realize real-time monitoring and closed-loop correction of strip steel deviation.

[0008] Preferably, the industrial area array camera is encapsulated within a protective cover equipped with a vortex tube cooling system; The cooling system includes nozzles, a vortex chamber, a separation orifice plate, cold / hot end pipes, and control valves, and is used to ensure stable operation of the equipment in a hot continuous rolling environment with high temperature and high dust.

[0009] Preferably, the strip deviation detection device is configured in a cascaded manner, with detection devices deployed between multiple adjacent stands of the finishing mill to achieve segmented deviation correction control.

[0010] Preferably, it also includes a strip tracking model for identifying the tail position of the strip and triggering a tail prediction control mode to achieve pre-compensation roll gap adjustment by integrating the visual deviation trend and the rolling force differential trend.

[0011] Preferably, the data logic server has a built-in process database that stores wedge adjustment allocation coefficient groups corresponding to different steel grades and specifications, and can dynamically optimize the allocation coefficients by combining historical data and machine learning algorithms. A visual correction and suppression method based on precision rolling wedge control, provided by the present invention, includes: Step S1: The PLC determines the strip biting state by monitoring the rolling force signal or the hot metal detector signal, and triggers the strip deviation detection device to start detection; Step S2: The strip deviation detection device acquires images of the strip, and calculates the strip deviation amount Δw through preprocessing, edge extraction, and linear fitting. Step S3: The PLC receives the deviation amount △w and calculates the roll gap deviation value △G of the downstream stand based on the rolling mill bounce principle; Step S4: Calculate the time delay T of the strip from the detection device to the downstream frame to achieve timing matching of the feedforward control; Step S5: Parallel execution of deviation correction control and wedge control: After waiting for the delay time T, adjust the roll gap of the downstream frame, and at the same time filter and judge the wedge value of the strip, calculate and adjust the roll gap on one side of the specified frame. Step S6: Determine if the downstream frame has thrown steel. If it has not thrown steel, return to step S2 to continue closed-loop control. If it has thrown steel, reset the system and wait for the next strip.

[0012] Preferably, in step S2, the image preprocessing includes median filtering or Gaussian filtering for noise reduction, threshold segmentation to separate the steel strip region, Canny operator or Sobel operator for edge extraction, and the camera parameters can be dynamically optimized according to the image brightness.

[0013] Preferably, the formula for calculating the delay time T is: T = S / V Where S is the fixed physical distance from the center line of the camera of the detection device to the center line of the downstream frame, and V is the exit speed of the strip leaving the upstream frame.

[0014] Preferably, it further includes a tail prediction control step: When the tail of the strip is about to reach the detection device, switch to the tail prediction mode. By integrating the visual deviation trend prediction value and the rolling force differential trend value, pre-compensate the roll gap of the downstream stand to suppress tail instability.

[0015] Preferably, the allocation coefficient of the wedge control can be dynamically retrieved from the process database according to the steel type and specifications, and continuously optimized through machine learning in combination with historical data.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention effectively ensures the stable operation of strip steel during the finishing rolling process through high-precision, real-time deviation detection and closed-loop control, significantly reducing the number of production accidents such as folding and tailing caused by deviation.

[0017] 2. This invention improves the transverse thickness uniformity of the strip steel and enhances the quality of the finished product by controlling the wedge shape in parallel.

[0018] 3. The control method of the present invention makes the production process more stable, reduces the accident handling time, and allows the production line to operate at a higher speed, thereby increasing output.

[0019] 4. This invention eliminates the inefficiency caused by labor-intensive work, reduces the amount of manual operation, realizes full automation of detection and control, reduces the amount of manual work, and achieves the goal of fewer people. Attached Figure Description

[0020] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the installation structure of a visual inspection device provided in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the principle of strip edge detection in an embodiment of the present invention. Figure 3This is a schematic diagram of the functional modules of the main interface of a finishing mill correction system provided in an embodiment of the present invention; Figure 4 A flowchart of a finishing mill correction control method provided in an embodiment of the present invention; Figure 5 A schematic diagram of the network connection structure of a deviation correction control system provided in an embodiment of the present invention. Figure 6 A schematic diagram illustrating the principle of calculating the roll gap deviation value that needs to be adjusted for the downstream frame, provided for an embodiment of the present invention. Detailed Implementation

[0021] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0022] This invention can be applied not only to the steel industry, but also to other industrial production fields, requiring only adjustments based on different processes.

[0023] It should be noted that the direct cause of lateral strip movement is the change in wedge ratio. The difference in strip thickness between the operating side (WS) and the drive side (DS) is called the wedge shape. The wedge ratio is defined as the wedge shape divided by the thickness at the center point of the strip width direction. The difference between the wedge ratio of the strip tail after being de-tailed from the previous stand and the wedge ratio of the strip rolled on the next stand results in a difference in the reduction elongation. If the strip has already been de-tailed from the previous stand, this causes lateral rotation of the strip tail as it enters the next mill, because the strip tail is unrestrained due to tension loss. When the laterally shifted tail enters the rolls, the lateral movement of the strip is amplified. Through analysis of the measured wedge shape data of the strip entering / exiting after finishing scrap, the following relationship exists between the calculated wedge shape and the strip thickness for each stand:

[0024] In the formula, for The inlet strip wedge value of the frame, for The wedge value of the strip at the outlet of the frame. for The thickness of the steel strip at the frame entrance. for The thickness of the strip steel at the frame exit. Described The extent to which the frame affects the strip wedge shape inheritance under zero leveling conditions. During commissioning, due to... It is difficult to observe, and individual frames are generally not adjusted separately. Instead of focusing on the value, the seven-stand rolling mill is considered as a whole, and the overall frame is adjusted based on the actual site conditions. The value is adjusted.

[0025] Example 1 This embodiment provides a visual correction and suppression system and method based on wedge control in finishing rolling. Specifically, the system uses high-precision, non-contact online detection to acquire the lateral position deviation and cross-sectional wedge shape of the strip during hot continuous rolling finishing in real time, and performs rapid closed-loop correction control accordingly to improve the stability of the production process, the quality of finished products, and safety.

[0026] like Figure 5 As shown in the figure, this diagram illustrates the network connection structure of the strip deviation correction control system in this embodiment. In one embodiment of the invention, the entire system uses a programmable logic controller (PLC) as the core logic control unit. The PLC can be a Siemens S7 series or TMIEC controller, used to execute the control quantities given by the vision control system. The 1 to 6 CCD cameras of the strip deviation detection device are respectively connected to 1 to 6 image processing servers to complete the acquisition and preprocessing of strip images. These image processing servers achieve data aggregation and transmission through switches, sending the calculated strip deviation data to the data logic server (i.e., the PLC core unit) on one hand, and exchanging information with the multi-functional control system on the other. After receiving the deviation data, the logic control unit calculates the roll gap deviation value of the downstream stand based on the mill bounce principle, simultaneously achieving control interlocking with the L1 system and issuing roll gap adjustment commands to the AGC control system. The multi-functional control system continuously collects the wedge-shaped data of the finished strip and feeds it back to the logic control unit, realizing real-time monitoring and closed-loop correction of strip deviation.

[0027] like Figure 1As shown in the figure, the physical installation of the strip misalignment detection device is detailed. The orange devices at the top represent six misalignment detection units, below which are seven stands (F1-F7). The red line in the middle represents the rolled strip. Taking a typical seven-stand finishing mill as an example, it is assumed that the device is deployed above the space between the fifth stand (F5) and the sixth stand (F6). Specifically, the visual correction and suppression system of this invention has a dedicated detection platform at the top of the finishing mill, on which the strip misalignment detection devices are installed. The strip misalignment detection devices are installed along the mill centerline. This device is a visual inspection system integrating image acquisition and processing functions. The main body is a high-precision industrial area array camera equipped with a corresponding area array lens, used to capture real-time images of the strip running area between adjacent stands. The industrial area array camera is enclosed in a protective cover with a cooling system to prevent damage to the system from the high temperature of the slab, ensuring that the camera and lens, and other precision optical and electronic equipment, can operate stably for a long time in the harsh environment of high-temperature radiation, water vapor, and dust in the hot continuous rolling mill. The cooling system employs vortex tube cooling, comprising nozzles, a vortex chamber, a separation orifice plate, tubes, and control valves. The vortex chamber, centrally located, divides the tube into cold and hot ends. Nozzles are arranged tangentially to the vortex chamber, guiding the high-pressure airflow tangentially into it. The orifice plate is positioned between the vortex chamber and the cold tube, while a control valve is installed at the outlet of the hot-end tube. Once the strip enters the camera's field of view, the system automatically triggers the image acquisition process. First, based on ambient lighting conditions, strip surface characteristics, and movement speed, the system automatically and dynamically optimizes key parameters such as camera exposure time, gain, focal length, and shooting frequency to ensure image clarity, stability, and consistency. Subsequently, the industrial area scan camera continuously scans and acquires images of the strip according to the set acquisition mode, obtaining high-resolution strip edge image data. The acquired images are processed by the image preprocessing module for denoising, enhancement, and edge extraction to extract the strip edge position information. Combined with the pixel calibration model, the image coordinates are converted into actual physical offsets, thereby achieving high-precision identification and quantitative calculation of strip deviation. Then, combined with the rolling line control model, the AGC momentum of the rolling line roll gap is controlled to achieve online steady-state control and real-time adjustment.

[0028] like Figure 4 As shown, the visual correction and suppression method based on precision rolling wedge control of the present invention includes the following steps: Step S1: Before rolling begins, the system is in standby mode. When a new strip enters the finishing mill, the PLC monitors the rolling force signal or hot metal detector signal between two adjacent stands to determine the strip biting status. Once it is confirmed that all stands have bitten the strip, the PLC sends a trigger signal to the image processing server of the strip misalignment detection device via Ethernet, thereby initiating the subsequent detection and control process.

[0029] Step S2: The triggered strip misalignment detection device begins real-time detection of the strip directly below it to obtain the strip misalignment value Δw. Specifically, the industrial camera continuously captures images of the strip at a preset high frame rate (e.g., 50 frames per second). To address the impact of varying ambient lighting conditions, inconsistent iron oxide scale on the strip surface, and fluctuations in rolling speed on image quality, the vision inspection system incorporates a parameter self-optimization function. This function dynamically and in real-time adjusts camera parameters such as exposure time, gain, and electronic shutter speed based on the brightness and contrast information of each frame to ensure consistently clear and stable strip images.

[0030] After acquiring the image, the image processing server executes a series of algorithms to calculate the deviation, the principle of which is as follows: Figure 2 As shown. Within the camera's entire field of view, the acquired raw strip image is first preprocessed, for example, using median filtering or Gaussian filtering algorithms to remove random noise. Next, taking advantage of the strip's luminescence at high temperatures and the relatively dark background, threshold-based image segmentation or blob analysis algorithms are used to accurately separate the bright strip region from the background, discarding the remaining areas with interference. Then, on the separated strip image, edge detection algorithms (such as the Canny or Sobel operators) are used to accurately extract the coordinates of a large number of pixels on the left and right edges of the strip. To obtain robust edge positions, the system uses least squares linear regression analysis on the extracted edge point set to fit fitted edge lines representing the left and right edges of the strip. These two fitted edge lines allow for accurate calculation of the strip's centerline. Finally, the calculated strip centerline is compared with the pre-calibrated rolling mill centerline; the lateral distance between the two is the current deviation amount Δw. It should be noted that the rolling mill centerline here is a baseline that is precisely calibrated in the camera's field of view using precision measuring tools and stored in the system during the system installation and commissioning phase. The entire process from image acquisition to the output of the deviation amount Δw is completed in an extremely short time (e.g., 20 milliseconds), thereby achieving real-time, high-frequency monitoring of deviation.

[0031] Step S3: The PLC receives the deviation amount △w (denoted as ERROR in the formula) sent by the strip deviation detection device and calculates the roll gap deviation value △G that needs to be adjusted downstream. Figure 6 As shown, this calculation utilizes the principle of mill bounce for reverse inference, which is the core of the deviation correction control in this embodiment. The calculation formula is as follows:

[0032] In the formula, ERROR represents the strip offset; Hgain is the preset strip thickness influence coefficient; W is the set strip width for the current rolling specification; Wgain is the preset strip width influence coefficient; FORCE is the target rolling force of the downstream stand; and KM is the mill stiffness coefficient of the downstream stand, which is a physical constant characterizing the mill's resistance to deformation and is measured at the factory or during offline calibration. The physical meaning of this formula is to estimate the rolling force imbalance caused by strip misalignment and, in turn, calculate the roll gap tilt adjustment ΔG required to counteract this imbalance, thereby correcting the strip back to its center position.

[0033] Step S4: Calculate the time it takes for the strip to travel from the strip misalignment detection device to the downstream stand, i.e., the delay time T. This aims to achieve feedforward control and ensure that the correction action is accurately applied to the strip section where misalignment has occurred. The calculation formula is as follows: T = S / V Where S is the physical distance from the camera centerline of the strip deviation detection device to the downstream frame centerline, which is a fixed measurement constant; V is the exit speed of the strip when it leaves the upstream frame.

[0034] Step S5: After calculating the delay time T, the system enters the parallel control execution phase. On one hand, it executes the deviation correction control step. After waiting for time T, the roll gap of the downstream frame is adjusted according to the roll gap deviation value ΔG. On the other hand, the system executes the wedge control step in parallel.

[0035] The wedge control steps include: First, a multi-function instrument located at the final outlet of the finishing mill continuously collects wedge data of the finished strip steel, including cross-sectional thickness data, and calculates the wedge value at a sampling period of 200ms. After receiving this series of wedge data, the PLC delays for 1 second before processing the data to avoid measurement errors caused by uneven temperature and unstable shape at the strip head. Subsequently, a moving average filtering method is used to filter the collected wedge data. Specifically, the system maintains a first-in-first-out queue of length N (e.g., N=5). In each acquisition cycle, the latest wedge value is pushed into the queue and the oldest data is removed. Then, the arithmetic mean of the N data in the queue is calculated, and this average is used as the effective wedge value for current control. That is, when the queue length is less than N, the filter does not output; when the length is N, the arithmetic mean of the queue is output; when the next acquisition cycle arrives, a new data is added, an old data is removed, and the arithmetic mean of the queue is output again, and so on.

[0036] The system performs dead-zone judgment on the filtered effective wedge value. To prevent frequent control system actions due to minor noise or unnecessary fluctuations, a control dead zone (e.g., ±5µm) is set. If the absolute value of the effective wedge value is less than this dead-zone value, no adjustment is made, and the process jumps to step S6; if the absolute value of the effective wedge value is greater than or equal to this dead-zone value, wedge adjustment calculation is initiated. The system selects several stands (e.g., F1 to F4) in the front section of the finishing mill as the execution stands for wedge adjustment, because the wedge adjustment effect is more significant in the early stage of rolling. Based on the principle of flow rate balance, the single-side roll gap adjustment amount required for each adjustment stand i is calculated. The formula is as follows:

[0037] in, This represents the single-sided roll gap adjustment amount (µm) of the i-th frame. This represents the measured wedge shape value (µm) of the strip. Indicates the mill stiffness coefficient. Indicates the plastic deformation coefficient. This indicates the thickness (mm) at the exit of the i-th rack. Indicates the F7 outlet thickness (mm). This indicates the width (mm) of the finished steel strip. This represents the allocation coefficient for the i-th rack, initially determined. =0.3, =0.3, =0.2, =0.2, This represents the coefficient of influence for a moderate impact. This represents the width influence coefficient (initial value is 0.2). This represents the thickness influence coefficient (initial value is 0.2).

[0038] The moderate influence coefficient is calculated from the strip centering value C (the distance between the strip center and the rolling line center) measured by the multi-function instrument using the following formula:

[0039] In the formula, This represents the centering value of the strip (mm). The value of this moderate influence coefficient ranges from 0 to 1. When the calculated value is greater than or equal to 1, it is taken as 1, and when it is less than or equal to 0, it is taken as 0. Its function is to appropriately reduce the wedge adjustment amount when the strip deviates far from the center, so as to avoid conflict with the deviation control.

[0040] Step S6: The system determines whether the downstream stand has rejected the strip. If the strip is still being rolled, the process returns to step S2, continuously looping the detection, calculation, and parallel control of strip deviation and wedge shape. This closed-loop control continues until the tail of the strip leaves the current downstream stand, and the PLC receives the rejection signal. At this point, the PLC resets the roll gap deviation value ΔG of the current downstream stand to zero, restoring it to a horizontal state. Finally, this control cycle ends, the system resets, and waits for the arrival of the next strip.

[0041] Operators can monitor this process throughout using the main interface on the human-machine interface. For example... Figure 3 As shown, the main interface is typically divided into several functional areas: the image display area displays the strip image captured by the strip misalignment detection device in real time, and can overlay dynamic graphics of the detected strip edge and misalignment amount △w; the misalignment data area displays the change of misalignment amount △w over time and key values ​​in the form of a curve graph; the control status area displays the current system mode (automatic or manual) and the issued control command values; the alarm indication area issues an alarm to the operator by flashing, changing color or sound when the misalignment amount exceeds the safety threshold or when the system malfunctions.

[0042] Based on the hot rolling mill process, the specific steps of image detection are as follows: Before the system is put into production, it learns the dimensions of various specifications of the slab and stores the learning results in the server. The images are then optimized until the system is mature and can automatically image. During automated production, the slab is transported to the detection area by a high-speed roller conveyor. The PLC system detects when the mill bites the slab and triggers communication with the high-speed CCD vision system to begin detection. The high-speed CCD forms a fixed-station model to capture the surface features of the slab. Due to the high temperature and red light on the slab surface, after processing by the system controller, waveforms with unevenness and different peak values ​​appear. Each waveform value is a feature value for contour detection. A Gaussian filtering algorithm is used to optimize the curves, resulting in a final optimized value. The system then calculates the interface coordinates of the slab's X and Y axes based on the optimal feature values ​​extracted from both sides of the strip. Within the CCD vision, the entire coordinate point is composed of pixels. The obtained X and Y coordinates are used to obtain an actual coordinate value based on the pixel ratio of X and Y. The actual coordinate value is compared with the benchmark coordinate to calculate the X and Y deviation value of the slab, thereby achieving precise positioning of the strip deviation.

[0043] The purpose of this method is to achieve feedforward control of the current rolling process based on the inspection results of the strip threading steady state in the finishing mill, through a correction control strategy and a wedge control setpoint calculation model. It communicates with the PLC via Ethernet TCP communication and achieves control interlocking with the L1 system.

[0044] Example 2 This embodiment, based on Embodiment 1, provides an optimized control strategy, particularly enhancing the control of the strip tail and the dynamic adaptability of the wedge control. The hardware system configuration, network connection, and basic control flow of this embodiment are the same as in Embodiment 1.

[0045] As an improvement, this embodiment introduces a dedicated tail control strategy. In actual production, the tail of the strip becomes extremely unstable due to the loss of back tension, representing a high-risk stage for "tail-wagging" accidents. When the tail of the strip passes the strip deviation detection device, visual information is lost, and the control method of Embodiment 1 will cease to function. The method of this embodiment, however, overcomes this deficiency.

[0046] Specifically, when the PLC learns from the strip tracking model that the tail of the strip is about to reach the strip misalignment detection device, the system automatically switches to tail prediction control mode. In this mode, the system performs the following operations: Data trend analysis: The system records and analyzes the trend of deviation Δw in the last few seconds (e.g., 2-3 seconds) before the visual detection fails. Through linear extrapolation or a more complex trend prediction model, a vision-based deviation trend prediction value is obtained.

[0047] Rolling force differential analysis: Simultaneously, the system closely monitors the difference in rolling force between the downstream stand operating side and the drive side. This difference is a sensitive indirect indicator of strip misalignment. The system will analyze the changing trend of this rolling force difference.

[0048] Fusion Prediction and Pre-Swing Roll Gap: The system weighted and fused the aforementioned vision-based deviation trend prediction with the trend based on differential rolling force to obtain a comprehensive prediction of the future deviation behavior of the strip tail. Based on this prediction result, the PLC calculates a pre-compensated roll gap deviation value and presets the roll gap of the downstream stand the instant the strip tail leaves the upstream stand. This proactive pre-compensation control can effectively suppress instability and swaying of the strip tail, thereby significantly reducing the probability of tail-swing accidents.

[0049] Another improvement in this embodiment lies in the dynamic allocation strategy of the wedge control. In Embodiment 1, the allocation coefficient of the wedge adjustment amount between stands F4 and F7 is fixed. However, different steel grades and specifications of products have different wedge inheritance patterns and responses to roll gap adjustments. To make the control more refined, this embodiment designs the allocation coefficient to be dynamically adjustable. As an optional implementation, different sets of recommended allocation coefficient values ​​can be established in the process database for different steel grades, thicknesses, and width specifications. Before rolling each strip, the PLC queries and downloads the corresponding allocation coefficient set based on the current strip specification information. For example: For harder, high-strength steel, which has greater resistance to plastic deformation, the system can employ a set of coefficients that allocate more adjustment to the stands with higher rolling forces in the front section, such as... =0.4, =0.3, =0.2, =0.1.

[0050] For automotive steel sheets with extremely high surface quality requirements, to avoid defects such as indentations introduced by large roll gap adjustments in the later stages of rolling, the system can employ a set of coefficients that concentrate more of the adjustment on the first two stands, such as... =0.5, =0.4, =0.1, =0. Furthermore, this dynamic allocation strategy can also incorporate historical data for self-learning optimization. More advanced L2 systems can continuously collect wedge control effect data for each steel piece and analyze it in specific situations. Control efficiency under allocation. Through machine learning algorithms, the system can periodically analyze the database. The recommended values ​​are fine-tuned to achieve continuous self-improvement of the wedge control strategy.

[0051] Furthermore, this embodiment also provides an optional hardware configuration scheme, namely, installing multiple strip deviation detection devices in the finishing mill to achieve cascaded control. For example, in addition to installing one device between stands F5 and F6, another can be added between stands F2 and F3. In this way, the detection device between F2 and F3 can provide deviation correction control for stand F3, performing preliminary correction of large deviations in the early stage of rolling; while the detection device between F5 and F6 performs more precise adjustments based on this. This cascaded architecture of multi-point detection and segmented control can suppress the development of deviation more quickly and smoothly, which is of great significance for improving rolling speed and handling steel grades with severe deviation tendency.

[0052] Compared to Example 1, the technical solution provided in this example has stronger adaptability to working conditions, higher control accuracy and better production safety through improvements to tail control and wedge dynamic distribution. Its technical advantages are more obvious, especially when dealing with extreme specification products and complex rolling conditions.

[0053] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0054] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A visual correction and suppression system based on precision rolling wedge control, characterized in that, include: Strip steel deviation detection device, image processing server cluster, switch, data logic server, multi-functional control system, L1 system and AGC control system; The strip deviation detection device includes at least one set of industrial area array cameras installed along the centerline of the rolling mill for acquiring images of the strip edge. The image processing server cluster corresponds one-to-one with the camera, performing image preprocessing and deviation calculation; The switch enables Ethernet communication between each image processing server and the data logic server; The data logic server receives the deviation data and calculates the roll gap deviation value of the downstream frame, achieves control interlock with the L1 system, and sends roll gap adjustment instructions to the AGC control system; The multi-functional control system continuously collects wedge-shaped data of the finished strip steel and feeds it back to the data logic server to realize real-time monitoring and closed-loop correction of strip steel deviation.

2. The visual correction and suppression system based on precision rolling wedge control according to claim 1, characterized in that, The industrial area array camera is encapsulated in a protective cover with a vortex tube cooling system. The cooling system includes nozzles, a vortex chamber, a separation orifice plate, cold / hot end pipes, and control valves, and is used to ensure stable operation of the equipment in a hot continuous rolling environment with high temperature and high dust.

3. The visual correction and suppression system based on precision rolling wedge control according to claim 1, characterized in that, The strip deviation detection device adopts a cascaded configuration, with detection devices deployed between multiple adjacent stands of the finishing mill to achieve segmented deviation correction control.

4. The visual correction and suppression system based on precision rolling wedge control according to claim 1, characterized in that, It also includes a strip tracking model to identify the tail position of the strip and trigger the tail predictive control mode, which achieves pre-compensation roll gap adjustment by integrating the visual deviation trend and the rolling force differential trend.

5. The visual correction and suppression system based on precision rolling wedge control according to claim 1, characterized in that, The data logic server has a built-in process database that stores wedge adjustment allocation coefficient groups corresponding to different steel grades and specifications, and can dynamically optimize the allocation coefficients by combining historical data and machine learning algorithms.

6. A visual correction and suppression method based on precision rolling wedge control, characterized in that, include: Step S1: The PLC determines the strip biting state by monitoring the rolling force signal or the hot metal detector signal, and triggers the strip deviation detection device to start detection; Step S2: The strip deviation detection device acquires images of the strip, and calculates the strip deviation amount Δw through preprocessing, edge extraction, and linear fitting. Step S3: The PLC receives the deviation amount △w and calculates the roll gap deviation value △G of the downstream stand based on the rolling mill bounce principle; Step S4: Calculate the time delay T of the strip from the detection device to the downstream frame to achieve timing matching of the feedforward control; Step S5: Parallel execution of deviation correction control and wedge control: After waiting for the delay time T, adjust the roll gap of the downstream frame, and at the same time filter and judge the wedge value of the strip, calculate and adjust the roll gap on one side of the specified frame. Step S6: Determine if the downstream frame has thrown steel. If it has not thrown steel, return to step S2 to continue closed-loop control. If it has thrown steel, reset the system and wait for the next strip.

7. The visual correction and suppression method based on precision rolling wedge control according to claim 6, characterized in that, In step S2, image preprocessing includes median filtering or Gaussian filtering for noise reduction, threshold segmentation to separate the steel strip region, Canny operator or Sobel operator for edge extraction, and camera parameters can be dynamically optimized according to image brightness.

8. The visual correction and suppression method based on precision rolling wedge control according to claim 6, characterized in that, The formula for calculating the delay time T is: T = S / V Where S is the fixed physical distance from the center line of the camera of the detection device to the center line of the downstream frame, and V is the exit speed of the strip leaving the upstream frame.

9. The visual correction and suppression method based on precision rolling wedge control according to claim 6, characterized in that, It also includes tail prediction control steps: When the tail of the strip is about to reach the detection device, switch to the tail prediction mode. By integrating the visual deviation trend prediction value and the rolling force differential trend value, pre-compensate the roll gap of the downstream stand to suppress tail instability.

10. The visual correction and suppression method based on precision rolling wedge control according to claim 6, characterized in that, The allocation coefficients for wedge control can be dynamically retrieved from the process database based on steel type and specifications, and continuously optimized through machine learning in conjunction with historical data.

Images