A method and system for online obstacle cleaning of a vertical plate casting nozzle for aluminum alloy plate production

Abstract

This invention belongs to the field of continuous casting and rolling production management technology for aluminum alloys, specifically relating to an online obstacle removal method and system for vertical plate casting nozzles in aluminum alloy sheet production. The method includes: acquiring images of the sheet surface; removing noise and extracting the visual width of defects using a density clustering algorithm; constructing a persistence index based on the defect change rate, casting speed, and temperature difference, combining fluid shear and thermodynamic principles, and inverting the physical hardness of the blockage; constructing a decision index based on the persistence index, product degradation loss, remaining production length, and obstacle removal cost; and intelligently determining the obstacle removal operation by comparing the decision index with a threshold. This invention, through deep coupling of physical inversion and economic game theory, achieves accurate differentiation between soft slag and hard blockage, as well as exemption management at the end of production, significantly improving finished product yield.

Description

Technical Field

[0001] This invention relates to the field of continuous casting and rolling production management technology for aluminum alloys. More specifically, this invention relates to an online obstacle removal method and system for vertical casting nozzles in aluminum alloy sheet production. Background Technology

[0002] In the aluminum alloy vertical plate casting process, slag buildup and blockage at the casting nozzle lip are key factors affecting plate surface quality. During casting, impurities in the molten aluminum easily deposit at the nozzle tip, causing varying degrees of blockage and resulting in defects such as streaks and scratches on the plate surface. With increasingly stringent market requirements for aluminum alloy plate surface quality, effectively managing the casting nozzle's operating conditions and promptly eliminating potential blockages has become a crucial aspect of production control.

[0003] Existing technologies typically employ a threshold-triggered management model, where a vision system detects defects on the panel surface. Once the detected defect area exceeds a preset threshold, the system immediately triggers a shutdown or speed reduction clearing procedure. However, this management method ignores the self-cleaning effect of fluid dynamics and cannot distinguish between loose, soft slag and sintered, hard blockages. Mechanically clearing soft slag, which can be washed away by high-speed fluid shear force, is not only excessive maintenance but may also introduce secondary scratches due to mechanical contact.

[0004] Furthermore, existing technologies lack economic considerations based on the production cycle. In actual production, if a coil of aluminum has only a few dozen meters left to be wound, the downtime, scrap, and joint costs caused by clearing obstacles often far exceed the degradation losses incurred by maintaining production as is. Existing technologies fail to comprehensively consider the dynamic relationship between remaining length and overall cost, leading to frequent and unnecessary clearing operations at the end of production, which severely reduces overall equipment efficiency (OEE) and finished product yield. Summary of the Invention

[0005] To address the technical problems in existing technologies, such as excessive maintenance due to the inability to distinguish the physical properties of blockages and resource waste at the end of production due to a lack of economic considerations regarding the production cycle, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides an online obstacle removal method for vertical plate casting nozzles in aluminum alloy sheet production, comprising: acquiring real-time image data of the surface of the aluminum alloy sheet; preprocessing the real-time image data to obtain a set of suspected defect points; performing spatiotemporal clustering screening on the set of suspected defect points using a density-based clustering algorithm to remove discrete noise points and retain defect clusters exhibiting linear characteristics; calculating the visual width of the defect clusters at the current moment; acquiring the current casting speed, the set temperature of the aluminum liquid holding furnace, and the real-time temperature of the casting nozzle tip; performing state estimation on the visual width using Kalman filtering to obtain the defect width growth rate; and calculating the defect width growth rate, casting speed, aluminum liquid holding furnace temperature, and other parameters based on the current casting speed, casting speed, and aluminum liquid holding furnace temperature. The difference between the set temperature and the real-time temperature of the casting nozzle tip is used to construct and calculate a stubbornness index characterizing the physical hardness of the blockage. The remaining length of the current coil to be produced, the price difference for downgrade losses per unit length of product, and the fixed comprehensive cost of performing the clearing operation are obtained. Based on the stubbornness index, the remaining length to be produced, the price difference for downgrade losses, and the fixed comprehensive cost, a decision index for assessing the necessity of clearing is constructed and calculated. This decision index is positively correlated with the remaining length to be produced and negatively correlated with the fixed comprehensive cost. The decision index is compared with a preset decision threshold. When the decision index is greater than the decision threshold, a control command is generated to trigger the clearing action to achieve online clearing of the vertical plate casting nozzle.

[0007] This invention utilizes a density-based spatiotemporal clustering algorithm to eliminate environmental interference, ensuring the accuracy of defect identification. A stubbornness index, constructed by combining fluid dynamics and thermodynamics principles, can assess the physical adhesion characteristics of blockages, effectively distinguishing between soft deposits and hard scale, thus avoiding excessive maintenance operations due to misjudgment. At the end of the roll material production process, the system can automatically adjust the clearing threshold based on marginal cost analysis, avoiding the cost of downtime and restart for minor defects.

[0008] Preferably, the step of using a density-based clustering algorithm to perform spatiotemporal clustering screening on the suspected defect point set includes: defining points in the suspected defect point set that are temporally adjacent and whose coordinate deviation in the width direction of the plate is within the allowable vibration range as neighborhood points; setting a minimum cluster point number threshold, and removing clusters of points with fewer than the minimum cluster point number threshold as environmental noise points; and determining clusters of points with a number of neighborhood points that meet the minimum cluster point number threshold as valid casting nozzle blockage defects.

[0009] Preferably, the stubbornness index satisfies the following relationship: ;in, Let be the stubbornness index at time t. This represents the defect width growth rate. The basic constant, At the current casting speed, This is the dimension adjustment factor. For temperature sensitivity coefficient, Set the temperature for the aluminum molten metal holding furnace. This is the real-time temperature of the casting nozzle tip.

[0010] By fully considering the velocity square term in fluid dynamics and the supercooling term in thermodynamics, the ability of blockages to resist fluid erosion can be effectively assessed, thereby distinguishing between soft and hard blockages and improving the accuracy of judgment.

[0011] Preferably, before the stubbornness index, the method further includes: smoothing the visual width using Kalman filtering, setting the state variable to include the defect width estimate and the instantaneous rate of change of the defect width, updating the covariance matrix using the posterior estimate from the previous time step and the observation from the current time step, and outputting the smoothed defect width growth rate.

[0012] Preferably, the decision index satisfies the following relationship: ;in, As a decision index, The stubbornness index, The price difference due to downgrading per unit length of product. This represents the remaining length of the current roll material to be produced. The fixed total cost of performing a single clearing operation, The fatigue risk coefficient of the equipment. This is the total planned length for this volume.

[0013] By introducing a remaining length variable into the control loop, the end-of-production stage of the roll-up process was controlled to suppress and clear obstacles, effectively reducing unnecessary downtime and waste, and significantly improving economic efficiency.

[0014] Preferably, the downgrade loss difference is obtained by acquiring the prices of premium and substandard products in the ERP system and calculating the difference between them.

[0015] Preferably, the fixed comprehensive cost includes the cost of scrap caused by downtime, the cost of restart transition materials, and the equipment depreciation amortization expenses; the equipment fatigue risk coefficient is used to characterize the mechanical operation risk cost caused by the increase in the continuous service mileage of the casting nozzle.

[0016] Preferably, when the decision index is not greater than the decision threshold, the system executes the operation of suspending the obstacle clearing command and marks and records the current defect area.

[0017] Preferably, the step of obtaining the remaining length to be produced includes: obtaining the target weight of the current coil, and the standard width, standard thickness, and standard density of the aluminum alloy sheet; calculating the theoretical total length of the current coil based on the target weight, standard width, standard thickness, and standard density; obtaining the current produced length from the PLC control system; and subtracting the current produced length from the theoretical total length to obtain the remaining length to be produced.

[0018] This allows for the calculation of the remaining length by combining weight and physical property parameters, improving the reliability of data at different winding stages.

[0019] Secondly, the present invention provides an online obstacle removal system for vertical plate casting nozzles in aluminum alloy sheet production, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned online obstacle removal method for vertical plate casting nozzles in aluminum alloy sheet production is implemented.

[0020] By adopting the above technical solution, a computer program is generated for the above-mentioned method of online obstacle removal of vertical plate casting nozzles for aluminum alloy sheet production, and stored in the memory so that it can be loaded and executed by the processor. In this way, a terminal device can be made based on the memory and the processor for convenient use.

[0021] This invention utilizes a density-based spatiotemporal clustering algorithm to eliminate environmental noise and combines fluid dynamics and thermodynamics principles to construct a stubbornness index, which characterizes the physical adhesion properties of blockages. This allows for accurate differentiation between soft deposits and hard scale, effectively avoiding misjudgments and over-maintenance caused by the inability of traditional visual inspection to distinguish the hardness of defects, thus ensuring the continuity and stability of the production process.

[0022] Furthermore, this invention establishes a dynamic game-theoretic decision-making model that includes remaining production length, quality degradation losses, and equipment operating costs, achieving economic optimization throughout the entire production cycle. Especially at the end of the roll material production process, the system can automatically adjust the obstacle removal threshold based on marginal cost analysis and implement a scientific end-of-life exemption strategy to avoid paying high downtime and restart costs for minor defects. While ensuring product quality, it significantly reduces the generation of process waste and significantly improves the overall equipment efficiency and finished product yield. Attached Figure Description

[0023] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0024] Figure 1This is a flowchart illustrating an online obstacle removal method for vertical plate casting nozzles in aluminum alloy sheet production according to the present invention.

[0025] Figure 2 This is a schematic diagram of the results of extracting casting nozzle defect features based on density clustering in an embodiment of the present invention;

[0026] Figure 3 This is a schematic diagram comparing the evolution of visual appearance and physical stubbornness of defects in an embodiment of the present invention;

[0027] Figure 4 This is a schematic diagram of the dynamic obstacle clearing decision mechanism based on the remaining length in an embodiment of the present invention. Detailed Implementation

[0028] 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, not all, of the embodiments of the present invention. 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.

[0029] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0030] This invention discloses an online obstacle removal method for vertical plate casting nozzles used in aluminum alloy sheet production, referring to... Figure 1 This includes steps S1-S4:

[0031] S1. Collect real-time image data of the aluminum alloy sheet surface, preprocess the real-time image data to obtain a set of suspected defect points, use density-based clustering algorithm to perform spatiotemporal clustering screening of the suspected defect point set, remove discrete noise points and retain defect clusters with linear characteristics, and calculate the visual width of the defect cluster at the current moment.

[0032] In an optional embodiment, images of the sheet metal surface can be acquired using a high-speed linear scan camera. For example, the camera sampling frequency is set to 5kHz to accommodate high-speed production lines. Due to the presence of significant oil mist and dust in the foundry, the vision sensor will generate randomly distributed noise. The system first performs basic image processing on the images, including binarization and edge detection, and outputs a set of suspected defect points. ,in This indicates the coordinates of the defect point along the width of the sheet material, in millimeters (mm). This indicates the system time frame sequence at which the data was collected, corresponding to the longitudinal position of the board.

[0033] Subsequently, a density-based clustering algorithm (DBSCAN) can be used to... The main focus of processing and defining distance metrics is... The difference in time between the two points Adjacent, for example, such as those differing by 1-5 frames and If the deviation is within the allowable vibration range (e.g., set to ±1mm), it is considered a neighboring point. A minimum cluster point threshold can also be set. For a continuous 10 frames, the algorithm will eliminate all discrete noise points with fewer than 10 neighboring points, retaining only point clusters that form continuous linear features as valid nozzle blockage defects. For each confirmed defect cluster, its value at the current time is calculated. Average visual width The unit is millimeters (mm).

[0034] like Figure 2 The diagram shown is a schematic representation of the defect feature extraction results of the casting nozzle based on density clustering in an embodiment of the present invention. Through DBSCAN clustering, the diagram shows that no clustering can be formed that satisfies the required parameters. Data points in clusters are identified as environmental noise and discarded; while data points that are arranged in a vertical line at a specific horizontal coordinate are determined to be valid casting nozzle blockage defects.

[0035] Thus, by using spatiotemporal density-based clustering analysis, we can effectively filter out complex environmental noise in industrial settings, accurately pinpoint real casting defects with longitudinal continuity and lateral stability, avoid false detections, and obtain more accurate defect baseline data.

[0036] S2. Obtain the current casting speed, the set temperature of the aluminum liquid holding furnace, and the real-time temperature of the nozzle tip. Use Kalman filtering to perform state estimation on the visual width to obtain the defect width growth rate. Based on the defect width growth rate, the casting speed, the difference between the set temperature of the aluminum liquid holding furnace and the real-time temperature of the nozzle tip, construct and calculate the stubbornness index characterizing the physical hardness of the blockage.

[0037] In an optional embodiment, the visual width alone cannot reflect the physical hardness of the blockage; therefore, this embodiment constructs an inversion model based on the fluid shear law and the thermal sintering law. Specifically, Kalman filtering can be used first to... Perform optimal estimation and define state variables. ,in This is the filtered defect width estimate (mm). Let be the instantaneous rate of change of the defect width (mm / s). Using the posterior estimate from the previous time step and the observed value at the current time step, the covariance matrix is ​​updated to output a smoothed value. .

[0038] Furthermore, the clogging persistence index can be calculated, which satisfies the following relationship:

[0039]

[0040] in, The stubbornness index at time t represents the ability of the blockage to resist fluid erosion; This represents the defect width growth rate of the Kalman filter output, in mm / s. For logical constraints, take the value 0 if the defect is decreasing; As a basic constant, for example, it takes the value of ; Given the current casting speed, and to ensure that the calculated stubbornness index is within the preset dimensionless evaluation range, the unit of the input value for this parameter in the actual calculation is meters per minute (m / min). To adapt to the unit dimension of casting speed, the adjustment factor is, for example, set to 1000; This is the temperature sensitivity coefficient, which, for example, is set to 2. The set temperature for the aluminum molten metal furnace is measured in degrees Celsius (°C) and is collected by a thermocouple. The temperature at the tip of the casting nozzle is measured in degrees Celsius (°C) and is obtained through a miniature thermocouple embedded in the side wall of the casting nozzle.

[0041] It is worth noting that, due to the growth rate of defect width in molecules The unit is mm / s, while the casting speed in the denominator is... To adapt to the habits of engineering sites, (m / min) was adopted as the unit of measurement, and since it involves square terms, the physical dimensions of the two are inconsistent. Therefore, Not only is it used for numerical amplification, but it also implicitly converts mm / s to... Physical constants are converted to units and made dimensionless to ensure that the final calculated stubbornness index is a dimensionless value that is easy to use for decision-making and evaluation.

[0042] Calculation example:

[0043] Suppose that during the production of a certain batch of aluminum alloy sheets, the data monitored by the system is: after Kalman filtering, the defect width growth rate... mm / s indicates that the defect is growing slowly; current casting speed The flow rate is set to 1.5 m / min, and the set temperature of the aluminum molten metal holding furnace is [not specified]. ℃, real-time temperature of the casting nozzle tip ℃, then the final calculation is The calculation results are close to the preset upper limit of 10, indicating that under the current flow rate and temperature difference conditions, even if the defect growth rate is only 0.02 mm / s, it exhibits extremely high physical stubbornness and belongs to hard blockage that is difficult to heal itself.

[0044] like Figure 3The diagram shown is a comparison of the evolution of visual appearance and physical stubbornness of defects in an embodiment of the present invention. It can be seen that the system can predict risks in advance by the surge in the stubbornness index when the visual width does not exceed the standard.

[0045] In this way, by combining the physical parameters of flow rate and temperature difference, it is possible to accurately identify potential blockages that may not be visually obvious but have a very high risk of hardening, thus avoiding the lag in judgment based solely on area size.

[0046] S3. Obtain the remaining length of the current roll material to be produced, the price difference of the downgrade loss per unit length of product, and the fixed comprehensive cost of performing the clearing operation. Based on the stubbornness index, the remaining length to be produced, the price difference of the downgrade loss, and the fixed comprehensive cost, construct and calculate a decision index for assessing the necessity of clearing. The decision index is positively correlated with the remaining length to be produced and negatively correlated with the fixed comprehensive cost.

[0047] In an optional embodiment, to accurately assess whether a clearing operation is necessary, the remaining length of the current roll to be produced, the price difference in degradation loss per unit length of product, and the fixed comprehensive cost of performing the clearing operation can be obtained. A dynamic clearing decision index is then constructed by combining this with a stubbornness index. The decision index satisfies the following relationship:

[0048]

[0049] in, As a decision index, The price difference for downgrading per unit length of product, expressed in yuan / meter, is obtained by comparing the price of premium-grade products with the price of substandard products through the ERP system. This represents the remaining length of the roll material to be produced, in meters. The data is sourced from the PLC system. The fixed comprehensive cost of performing one obstacle removal operation, in yuan, includes downtime waste, transitional materials and depreciation, and can be set to a fixed value, for example, 2000 yuan; The fatigue risk factor for the equipment is expressed in yuan per meter; for example, it can be set to 0.5. This is the total planned length of this volume, in meters.

[0050] Specifically, the steps for obtaining the remaining length to be produced include: obtaining the target weight of the current coil, as well as the standard width, standard thickness, and standard density of the aluminum alloy sheet; then calculating the theoretical total length of the current coil based on the target weight, standard width, standard thickness, and standard density; next, obtaining the current produced length from the PLC control system; and finally, subtracting the current produced length from the theoretical total length to obtain the remaining length to be produced.

[0051] For example, Assuming product downgrade losses Yuan / meter, for the initial stage of production, the remaining length If the length of the meter produced is close to 0, then... ,result The decision is clear: obstacles must be cleared. For the final stage of production, assuming production is nearing completion, the remaining length... Meters, produced length Rice, then ,result At this point, although the stubbornness is as high as 9.63, the system determines that it is not worthwhile to clear the obstacles because the remaining output loss is less than the cost of clearing and risk.

[0052] like Figure 4 The diagram shown is a schematic of the dynamic obstacle clearing decision mechanism based on the remaining length in an embodiment of the present invention. It can be seen that the decision threshold changes dynamically with the production progress and drops sharply at the end of the roll.

[0053] Thus, by introducing an economic game model that considers remaining length and equipment risk, intelligent dynamic decision-making can be achieved, especially in avoiding unnecessary shutdowns and clearing obstacles at the end of production. This allows for the scientific tolerance of non-fatal defects at the end of production while ensuring overall economic benefits.

[0054] S4. Compare the decision index with the preset decision threshold. When the decision index is greater than the decision threshold, generate a control command to trigger the obstacle clearing action to achieve online obstacle clearing of the vertical plate casting nozzle.

[0055] In an optional embodiment, a decision threshold can be preset to 1, when When the system determines that the benefits of clearing the obstruction outweigh the costs, it sends a clearing command to the actuator via the PLC, controlling the casting and rolling mill to slow down and separate the casting nozzle for cleaning; when If the system determines that the cost of clearing the obstacle exceeds the benefit, it suspends the obstacle clearing command, maintains the current production speed, and marks the product segment as downgraded in the quality report for reference by subsequent processes.

[0056] In this way, complex comprehensive assessments are transformed into clear execution instructions, realizing automated closed-loop control of the production process, reducing the arbitrariness of human intervention, and ensuring the consistency and efficiency of decision execution.

[0057] This invention also discloses an online obstacle removal system for vertical plate casting nozzles in aluminum alloy sheet production, comprising a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an online obstacle removal method for vertical plate casting nozzles in aluminum alloy sheet production according to the present invention.

[0058] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0059] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

[0060] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for online obstacle removal of the casting nozzle in the production of aluminum alloy sheets, characterized in that, include: Real-time image data of the surface of aluminum alloy sheet is collected, the real-time image data is preprocessed to obtain a set of suspected defect points, and the set of suspected defect points is spatiotemporally clustered and filtered using a density-based clustering algorithm to remove discrete noise points and retain defect clusters with linear characteristics. The visual width of the defect cluster at the current moment is calculated. The current casting speed, the set temperature of the aluminum liquid holding furnace, and the real-time temperature of the nozzle tip are obtained. The visual width is estimated using Kalman filtering to obtain the defect width growth rate. Based on the difference between the defect width growth rate, the casting speed, the set temperature of the aluminum liquid holding furnace, and the real-time temperature of the nozzle tip, a stubbornness index characterizing the physical hardness of the blockage is constructed and calculated. The remaining length of the current roll material to be produced, the price difference of the downgrade loss per unit length of product, and the fixed comprehensive cost of performing the clearing operation are obtained. Based on the stubbornness index, the remaining length to be produced, the price difference of the downgrade loss, and the fixed comprehensive cost, a decision index for assessing the necessity of clearing is constructed and calculated. The decision index is positively correlated with the remaining length to be produced and negatively correlated with the fixed comprehensive cost. The decision index is compared with a preset decision threshold. When the decision index is greater than the decision threshold, a control command is generated to trigger the obstacle clearing action to achieve online obstacle clearing of the upright plate casting nozzle.

2. The method for online obstacle removal of the casting nozzle in the production of aluminum alloy sheets according to claim 1, characterized in that, The step of using a density-based clustering algorithm to perform spatiotemporal clustering screening on the set of suspected defect points includes: Points in the set of suspected defect points that are temporally adjacent and whose coordinate deviation in the width direction of the plate is within the allowable vibration range are defined as neighborhood points; Set a minimum cluster point number threshold, and remove clusters of points whose number of neighboring points is less than the minimum cluster point number threshold as environmental interference noise points; Clusters of points whose number of neighboring points meets the minimum cluster point number threshold are identified as valid casting nozzle blockage defects.

3. The method for online obstacle removal of the casting nozzle in the production of aluminum alloy sheets according to claim 1, characterized in that, The stubbornness index satisfies the following relationship: in, Let be the stubbornness index at time t. This represents the defect width growth rate. The basic constant, At the current casting speed, This is the dimension adjustment factor. For temperature sensitivity coefficient, Set the temperature for the aluminum molten metal holding furnace. This is the real-time temperature of the casting nozzle tip.

4. The method for online obstacle removal of the casting nozzle in the production of aluminum alloy sheets according to claim 3, characterized in that, Before calculating the stubbornness index, the following is also included: The visual width is smoothed using Kalman filtering. The state variable is set to include the defect width estimate and the instantaneous rate of change of the defect width. The covariance matrix is ​​updated using the posterior estimate from the previous time step and the observation from the current time step, and the smoothed defect width growth rate is output.

5. The method for online obstacle removal of the casting nozzle in the production of aluminum alloy sheets according to claim 1, characterized in that, The decision index satisfies the following relationship: in, As a decision index, The stubbornness index, The price difference due to downgrading per unit length of product. This represents the remaining length of the current roll material to be produced. The fixed total cost of performing a single clearing operation, The fatigue risk coefficient of the equipment. This is the total planned length for this volume.

6. The method for online obstacle removal of the casting nozzle in the production of aluminum alloy sheets according to claim 5, characterized in that, The downgrade loss difference is obtained by acquiring the prices of premium and substandard products in the ERP system and calculating the difference between them.

7. The method for online obstacle removal of the casting nozzle in the production of aluminum alloy sheets according to claim 5, characterized in that, The fixed comprehensive cost includes the cost of scrap caused by downtime, the cost of restart transition materials, and the amortization of equipment depreciation; the equipment fatigue risk coefficient is used to characterize the mechanical operation risk cost caused by the increase in the continuous service mileage of the casting nozzle.

8. The method for online obstacle removal of the casting nozzle in the production of aluminum alloy sheets according to claim 1, characterized in that, When the decision index is not greater than the decision threshold, the system executes the operation of suspending the obstacle clearing command and marks and records the current defect area.

9. The method for online obstacle removal of the casting nozzle in the production of aluminum alloy sheets according to claim 5, characterized in that, The steps for obtaining the remaining production length include: Obtain the target weight of the current coil, as well as the standard width, standard thickness, and standard density of the aluminum alloy sheet; Calculate the theoretical total length of the current roll material based on the target weight, standard width, standard thickness, and standard density; Obtain the current produced length from the PLC control system; The remaining length to be produced is obtained by subtracting the currently produced length from the theoretical total length.

10. An online obstacle clearing system for vertical plate casting nozzles in aluminum alloy sheet production, characterized in that, include: The defect identification module is used to collect real-time image data of the surface of aluminum alloy sheet, preprocess the real-time image data to obtain a set of suspected defect points, use a density-based clustering algorithm to perform spatiotemporal clustering screening on the set of suspected defect points, remove discrete noise points and retain defect clusters with linear characteristics, and calculate the visual width of the defect cluster at the current moment. The physical inversion module is used to obtain the current casting speed, the set temperature of the aluminum liquid holding furnace, and the real-time temperature of the nozzle tip. The Kalman filter is used to perform state estimation on the visual width to obtain the defect width growth rate. Based on the difference between the defect width growth rate, the casting speed, the set temperature of the aluminum liquid holding furnace, and the real-time temperature of the nozzle tip, a stubbornness index characterizing the physical hardness of the blockage is constructed and calculated. The decision calculation module is used to obtain the remaining length of the current roll material to be produced, the price difference of the downgrade loss per unit length of product, and the fixed comprehensive cost of performing the obstacle removal operation. Based on the stubbornness index, the remaining length to be produced, the price difference of the downgrade loss, and the fixed comprehensive cost, a decision index for assessing the necessity of obstacle removal is constructed and calculated. The decision index is positively correlated with the remaining length to be produced and negatively correlated with the fixed comprehensive cost. The execution control module is used to compare the decision index with a preset decision threshold, and when the decision index is greater than the decision threshold, generate a control command to trigger the obstacle clearing action.

Images