Automatic cleaning method for steel ladle collector nozzle

Abstract

The invention relates to the technical field of steel ladle hot repair, in particular to an automatic cleaning method for a steel ladle collector nozzle. The automatic cleaning method for the steel ladle collector nozzle comprises the steps that after a steel ladle is controlled to tip over, a field image at the steel ladle collector nozzle is collected; sequentially carrying out pixel comparison with the field image through a preset reference template so as to extract a drain port area image and position the working coordinates of the drain port area image; extracting a lower nozzle boundary in the lower nozzle area image through edge detection, and segmenting an area to be cleaned; a robot arm is controlled to pick up an oxygen lance on a preset oxygen lance frame, and the oxygen lance is controlled to spray oxygen-enriched gas to a water outlet based on the working coordinates; wherein in the oxygen blowing process, control parameters of the oxygen lance are adjusted based on the area of the to-be-cleaned area, and the control parameters at least comprise air pressure and an oxygen blowing distance; through automatic technologies such as image recognition, robotic arm control and oxygen lance spraying, automatic operation of cleaning the water outlet is achieved, and manual intervention is reduced.

Description

Technical Field

[0001] The invention relates to the technical field of ladle hot repair, and in particular to an automatic cleaning method for a ladle water outlet. Background Art

[0002] In the steel smelting process, the ladle is an important container for carrying and transporting high-temperature molten steel. Its working condition is directly related to production efficiency and product quality. After the ladle is used, a certain amount of molten iron usually remains at the drain. This residual molten iron not only wastes resources, but may also block the drain due to cooling and solidification, affecting the subsequent smooth discharge of molten steel, and even damage the overall structure of the ladle, shortening its service life.

[0003] Traditionally, cleaning the molten iron stuck to the ladle outlet mainly relies on regular manual use of mechanical tools to remove it, but this method has many shortcomings. First, manual cleaning is labor-intensive, the working environment is harsh, and high-temperature work poses a serious threat to workers' health; second, although mechanical tool cleaning can reduce labor intensity, it is often difficult to completely remove the tightly stuck molten iron, and it may cause damage to the edge of the outlet, affecting the sealing.

[0004] In view of this, the present invention is provided. Summary of the invention

[0005] To solve the above problems, the automatic cleaning method for the ladle outlet provided in this application aims to achieve efficient, safe and environmentally friendly cleaning operations, optimize the steel smelting process, improve production efficiency and product quality, and at the same time ensure the health and safety of workers.

[0006] The first aspect of the present application provides an automatic cleaning method for a ladle water outlet, comprising: after controlling the ladle to tip over, collecting a field image at the ladle water outlet; performing pixel comparisons with the field image in sequence through a preset reference template to extract a water outlet area image and locate the working coordinates of the water outlet area image; extracting the water outlet boundary from the water outlet area image through edge detection and segmenting the area to be cleaned; controlling a robot arm to pick up an oxygen gun from a preset oxygen gun rack, and controlling the oxygen gun to spray oxygen-rich gas to the water outlet based on the working coordinates; wherein, during the oxygen blowing process, the control parameters of the oxygen gun are adjusted based on the area of ​​the area to be cleaned, and the control parameters include at least air pressure and oxygen blowing distance.

[0007] In a further scheme of the present application, a preset reference template is used to perform pixel comparison with the on-site image in sequence to extract the drain area image and locate the working coordinates of the drain area image, including: real-time acquisition of the on-site image at the drain of the ladle; preprocessing the on-site image, and performing pixel-by-pixel comparison with each area of ​​the preprocessed on-site image through a preset basic template to calculate the regional similarity; sorting the regional similarities and taking the area corresponding to the highest regional similarity as the reference area; extracting the drain area image and locating the working coordinates of the drain in the on-site graphics based on the positional relationship between the reference area and the drain.

[0008] During the matching process, the method provided above calculates the similarity between each feature area and the preset template and sorts them. By finding the area with the highest similarity as the reference area, the accuracy of recognition can be further ensured. This sorting mechanism helps to select the area that best matches the preset template as the recognition result when there are multiple similar areas; since multiple feature areas are used for matching and the similarity is calculated for sorting, this method can significantly reduce the possibility of misidentification. Even if there are objects or interference factors similar to the sewer in the on-site image, this method can eliminate these interferences through feature matching and similarity sorting to ensure the accuracy of recognition.

[0009] In a further solution of the present application, the scene image is preprocessed, and a pixel-by-pixel comparison is performed between each region of the preprocessed scene image and a preset basic template, and the calculation of the regional similarity includes: preprocessing the scene image by Gaussian filtering; inputting the basic template and the scene graphic, and calculating the regional similarity using the following formula:

[0010] Where T(i, j) is the pixel value of the base template, I(x+i, y+j) is the pixel value corresponding to each area of ​​the image at position (x, y), and (x, y) are the average values ​​of the base template and the image area, respectively.

[0011] In a further scheme of the present application, the automatic cleaning method also includes: obtaining the area of ​​the area to be cleaned, and when the area of ​​the area to be cleaned is greater than a preset area threshold, triggering the control robot arm to pick up the oxygen gun from the oxygen gun rack; and / or triggering the control robot arm to pick up the oxygen gun from the oxygen gun rack at every preset time value.

[0012] The present application can adopt two triggering mechanisms. One is to accurately calculate the area of ​​the area to be cleaned and promptly trigger the robot arm to pick up the oxygen gun when the area exceeds the threshold, so as to ensure that the cleaning work is carried out when it is most needed, thereby improving the timeliness and efficiency of cleaning; the other is a triggering mechanism with a preset time interval to ensure the regular execution of the cleaning work. Regardless of the actual size of the area to be cleaned, the continuity and regularity of the cleaning can be maintained, which helps to maintain the long-term stable operation of the equipment. The two methods can be carried out simultaneously, and only one of them needs to be met; or only one triggering method can be selected.

[0013] In a further scheme of the present application, the diameter of the water outlet is consistent with the end nozzle of the oxygen gun, and the robot arm is controlled to pick up the oxygen gun from the oxygen gun rack, and the oxygen gun is controlled to spray oxygen-rich gas to the water outlet at the water outlet, including: controlling the robot arm to pick up the oxygen gun from the oxygen gun rack; controlling the oxygen gun to be connected to the water outlet according to the working coordinates, and determining whether it is installed in place according to the torque feedback of the mechanical sensor installed at the water outlet; when installed in place, triggering the oxygen gun to spray oxygen-rich gas to the water outlet.

[0014] The diameter of the drain port is designed to be consistent with the nozzle at the end of the oxygen lance, and the installation of the oxygen lance and the injection of oxygen-rich gas are precisely controlled by a robotic arm, which brings about precise matching, intelligent installation, reduced manual intervention, improved automation level, enhanced cleaning effect and extended equipment life.

[0015] In a further solution of the present application, during the oxygen blowing process, adjusting the control parameters of the oxygen gun based on the area of ​​the area to be cleaned includes: mapping the corresponding gas pressure according to the area of ​​the area to be cleaned.

[0016] In a further scheme of the present application, the area to be cleaned is identified based on the image of the drain area, including: extracting edge features of the drain and the area to be cleaned through an edge detection algorithm; extracting the drain boundary from the edge features through a shape matching algorithm; and performing image segmentation through the drain boundary to extract the area to be cleaned.

[0017] In a further scheme of the present application, the automatic cleaning method also includes: after the oxygen gun is installed in place and the oxygen-rich gas is sprayed to the water outlet for a preset spraying time, controlling the robotic arm to disassemble the oxygen gun; determining whether the disassembly is in place based on the torque feedback of the mechanical sensor installed at the water outlet; and controlling the robotic arm to reset after the disassembly is in place.

[0018] After the oxygen lance completes the preset spraying time, the robot arm is controlled to disassemble the oxygen lance to ensure the integrity and continuity of the cleaning operation; through the torque feedback of the mechanical sensor, it can be quickly determined whether the oxygen lance is disassembled in place, thus avoiding unnecessary waiting and inspection time.

[0019] In a further scheme of the present application, the diameter of the water outlet is larger than the end nozzle of the oxygen gun, and the automatic cleaning method also includes: generating a cleaning path according to the area to be cleaned, the cleaning path includes an initial point and an end point; generating a movement path of the robot arm based on the oxygen gun rack coordinate information on the oxygen gun rack, the current position of the end of the robot arm, and the cleaning path.

[0020] In a further scheme of the present application, the robot arm is controlled to pick up the oxygen gun from a preset oxygen gun rack, and the oxygen gun is controlled to spray oxygen-rich gas to the water outlet based on the working coordinates, and the control parameters of the oxygen gun are adjusted according to the area to be cleaned, including: the robot arm moves according to the motion path, and only controls the oxygen gun to spray oxygen-rich gas to the water outlet during the cleaning path stage, and adjusts the control parameters of the oxygen gun according to the area to be cleaned.

[0021] In a further solution of the present application, a motion path of the robot arm is generated based on the oxygen lance rack coordinate information on the oxygen lance rack, the current position of the end of the robot arm, and the cleaning path, including: reading the current oxygen lance rack coordinate information from the control system of the oxygen lance rack, the oxygen lance rack coordinate information including the remaining oxygen lance coordinates and the remaining placement position coordinates; calculating a first path from the current position of the end of the robot arm to the remaining oxygen lance coordinates, a second path from the remaining oxygen lance coordinates to the initial point, a third path from the end point to the remaining placement position coordinates, and a fourth path from the placement position coordinates to the reset point; combining the first path, the second path, the third path, and the fourth path to generate a motion path;

[0022] In a further scheme of the present application, a cleaning path is generated according to the area to be cleaned, including: when the molten iron area is larger than a preset area threshold, marking the corresponding quadrant as the target quadrant; and calculating the shortest path in the preset cleaning trajectory according to the target quadrant as the cleaning path.

[0023] In a further scheme of the present application, the oxygen lance is controlled to spray oxygen-rich gas to the water outlet only in the cleaning path stage, and the control parameters of the oxygen lance are adjusted according to the area to be cleaned, including: dividing the boundary of the water outlet into multiple quadrants of equal angles according to 360 degrees; counting the area of ​​molten iron in the area to be cleaned in each quadrant; in the cleaning path stage, adjusting the control parameters of the oxygen lance in different quadrants according to the molten iron area.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] Based on a general inventive concept provided by the present application, oxygen-rich gas is automatically sprayed using an oxygen lance, and the gas contains a relatively high concentration of oxygen, which enables it to quickly induce an oxidation reaction when it comes into contact with fixed residues at the drain (such as molten iron, oxides, etc.); an oxidation reaction is a chemical reaction between a substance and oxygen, usually accompanied by the release of energy (combustion); in this process, elements in the residue (such as iron) combine with oxygen to form new oxides, which, in conjunction with the physical scouring of the spray, scour and erode the residues on the surface of the drain; by combining physical scouring with chemical oxidation reactions, the residues at the drain can be removed more efficiently, making it easier to fall off from the inner wall surface of the drain, and flow back into the ladle after the ladle is tipped over.

[0026] This application realizes the automated operation of drain cleaning and reduces manual intervention through automated technologies such as image recognition, robot arm control and oxygen gun injection. By using image recognition technology, the drain area can be accurately located in the on-site image and the area to be cleaned can be extracted to ensure the accuracy of the cleaning work. According to the analysis of the area to be cleaned, the control parameters of the oxygen gun can be flexibly adjusted to achieve the best cleaning effect, while ensuring the effective use of oxygen-rich gas and reducing energy waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0028] Figure 1 This is a general flow chart of the automatic cleaning method for the ladle outlet provided by an embodiment of the present invention;

[0029] Figure 2 It is a further flow chart of step S10 in the automatic cleaning method for the ladle outlet provided by an embodiment of the present invention;

[0030] Figure 3 is a further flow chart of step S20 in the automatic cleaning method for the ladle outlet provided by an embodiment of the present invention;

[0031] Figure 4 is a further flow chart of step S30 in the automatic cleaning method for the ladle outlet provided by an embodiment of the present invention;

[0032] Figure 5 is a further flow chart of step S50 in the automatic cleaning method for the ladle outlet provided by an embodiment of the present invention;

[0033] Figure 6is a further flow chart of step S60 in the automatic cleaning method for the ladle outlet provided by an embodiment of the present invention;

[0034] Figure 7 Schematic diagram of drain outlets of different specifications provided by an embodiment of the present invention;

[0035] Figure 8 A schematic diagram of a scene image provided by an embodiment of the present invention; and

[0036] Fig. 9 A schematic diagram of positioning working coordinates provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The specific implementation of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the present invention, and is not used to limit the present invention.

[0038] See also Figure 1 , Figure 8 as well as Fig. 9 , Figure 1 The present invention is a flowchart of the steps of the automatic cleaning method provided by the embodiment of the present invention. In order to solve the problems in the prior art, the present invention provides a method for disassembling and assembling an oil cylinder for hot repair of a ladle. The automatic cleaning method comprises the following steps:

[0039] Step S10, after controlling the ladle to tip over, collecting the on-site image at the ladle outlet;

[0040] Step S20, comparing pixels of the preset reference template with the on-site image in sequence to extract the drain area image and locate the working coordinates of the drain area image;

[0041] Step S30, extracting the drain port boundary in the drain port area image by edge detection and segmenting the area to be cleaned;

[0042] Step S40, control the robot arm to pick up the oxygen gun from the preset oxygen gun rack, and control the oxygen gun to spray oxygen-rich gas to the water outlet based on the working coordinates; wherein, during the oxygen blowing process, the control parameters of the oxygen gun are adjusted based on the area of ​​the area to be cleaned, and the control parameters include at least air pressure and oxygen blowing distance.

[0043] First, the ladle is tilted to a suitable angle, such as 95°, through the control system; the main purpose is to make the end of the drain away from the ladle relatively higher than the end close to the ladle, that is, when the ladle is tilted to a suitable angle, a certain slope is formed between the high end and the low end of the drain, so that the molten steel accumulated at the drain can flow back, and then trigger the use of image acquisition equipment (such as a high-definition camera) to capture the scene where the ladle drain is located to obtain the scene image (such as Figure 8); Based on image recognition technology, according to the characteristics of the ladle outlet (such as shape, color, position mark, etc.), the precise position of the outlet is located in the captured on-site image, and the image of the area is extracted.

[0044] Specifically, a reference template of the sewer area is pre-set and stored, which contains the typical features of the sewer area; the on-site image is compared with the reference template at the pixel level to find the area matching the template, that is, the sewer area image; through image processing and recognition algorithms, the position of the sewer area image in the overall image is determined, and its working coordinates are calculated.

[0045] At the same time, based on the drain area image, the reference coordinate system for subsequent cleaning work is further determined, that is, the working coordinates of the drain area image required for the robotic arm to locate. This coordinate system is usually associated with the control system of the robotic arm or other automated equipment to ensure the accuracy of the cleaning operation.

[0046] In the extracted image of the drain area, edge detection algorithms, such as the Laplace operator and the Canny edge detector, are applied to identify the boundary of the drain area. According to the identified boundary, the area to be cleaned is segmented, which may include molten iron residue, oxide accumulation, etc.

[0047] By controlling the robot arm to move to the preset oxygen gun rack and grab the oxygen gun, based on the previously determined working coordinates (such as Fig. 9 ), the robot arm positions the oxygen gun at the drain outlet, adjusts the control parameters of the oxygen gun according to the specific conditions of the area to be cleaned, starts the oxygen gun to spray oxygen-rich gas (usually containing a higher concentration of oxygen to enhance the combustion effect), cleans the drain outlet, and the oxygen-rich gas reacts with the residue at the drain outlet to generate high temperature, thereby removing the residue.

[0048] Based on a general inventive concept provided by the present application, an oxygen-rich gas is automatically sprayed with an oxygen gun, and the gas contains a high concentration of oxygen, which enables it to quickly induce an oxidation reaction when it contacts fixed residues (such as molten iron, oxides, etc.) at the drain. An oxidation reaction is a chemical reaction between a substance and oxygen, usually accompanied by the release of energy. In this process, elements (such as iron) in the residue combine with oxygen to form new oxides, which, under the action of physical scouring by the jet, scour and erode the residues on the surface of the drain. This physical scouring action combined with the chemical oxidation reaction can more efficiently remove the residues at the drain and make it easier to fall off from the inner wall surface of the drain.

[0049] This application realizes the automated operation of drain cleaning and reduces manual intervention through automated technologies such as image recognition, robot arm control and oxygen gun injection. By using image recognition technology, the drain area can be accurately located in the on-site image and the area to be cleaned can be extracted to ensure the accuracy of the cleaning work. According to the analysis of the area to be cleaned, the control parameters of the oxygen gun can be flexibly adjusted to achieve the best cleaning effect, while ensuring the effective use of oxygen-rich gas and reducing energy waste.

[0050] See also Figure 2 Based on an implementation of the above step S10, a live image of the ladle drain is collected, and the drain area image is located and extracted in the live image according to the features of the ladle drain and the working coordinates are located, including:

[0051] Step S110, preprocessing the scene image, and performing pixel-by-pixel comparison between each region of the preprocessed scene image and a preset basic template to calculate the region similarity;

[0052] Step S120, sorting the region similarities, and taking the region corresponding to the highest region similarity as the reference region;

[0053] Step S130: extracting the drain area image and locating the working coordinates of the drain in the on-site graphics according to the positional relationship between the reference area and the drain.

[0054] The collected on-site images are preprocessed; preprocessing may include denoising, contrast enhancement, color balance adjustment and other operations to improve the quality and readability of the image. Next, a preset basic template (this template is usually made based on the known features or shape of the manhole) is used to perform a pixel-by-pixel comparison with each area of ​​the preprocessed on-site image. By calculating the similarity between each area and the template (possibly using methods such as mean square error, structural similarity index, histogram comparison or feature matching), the degree of match between these areas and the template can be evaluated; the calculated similarities of each area are sorted to find the area with the highest similarity. This area is considered to be the reference area, that is, the area most likely to contain the manhole, and the reference area is determined as the most credible match.

[0055] An example is: preset multiple basic templates (which may be the shape, color, texture and other features of the drain), and then match the template with each area of ​​the scene image. The matching process may involve algorithms such as pixel-by-pixel comparison, feature extraction and matching. In the scene image, identify multiple feature areas such as a, b, c, including edges, corners, bottoms, etc.; compare each feature area with the preset template at the pixel level, and calculate the similarity between them; sort the similarities between each feature area and the template, and find the area with the highest similarity, that is, the reference area, which can be considered as the most accurate identification position at present. According to the original position relationship between the reference area and the drain (for example, the drain is located at the center, edge or specific direction of the reference area, etc.), the center position of the drain is inferred.

[0056] Through pixel-by-pixel comparison, feature extraction and matching algorithms, areas in the scene image that are similar to the preset template can be accurately identified; since multiple feature areas are used for matching and the similarity is calculated for sorting, the possibility of misidentification can be reduced.

[0057] In an optional manner of the present application, the scene image is preprocessed, and a pixel-by-pixel comparison is performed between each region of the preprocessed scene image and a preset basic template, and the regional similarity is calculated including:

[0058] Step S121, preprocessing the scene image by Gaussian filtering;

[0059] Use a two-dimensional Gaussian kernel to convolve with the image. The weight of the Gaussian kernel is determined by the Gaussian function, and its shape is a centrally symmetrical bell-shaped curve. The standard deviation σ determines the width of the curve, that is, the smoothness of the filter. By adjusting the size of σ, the degree of smoothing can be controlled, and the Sobel operator is used to calculate the horizontal (Gx) and vertical (Gy) gradients of the image. The Sobel operator is a discrete differential operator that combines Gaussian smoothing and differential derivation, which can effectively suppress noise and extract edges. Through the above steps, the Canny algorithm can effectively extract the edge of the drain in the image and perform denoising; Gaussian filtering is used to smooth the image and reduce noise; gradient calculation and non-maximum suppression are used to determine the position and strength of the edge; in this way, the final edge image will clearly show the position and shape of the drain, providing an accurate basis for subsequent positioning and analysis.

[0060] Step S122: input the basic template and the scene graph, and calculate the regional similarity using the following formula:

[0061]

[0062] Where T(i, j) is the pixel value of the base template, I(x+i, y+j) is the pixel value corresponding to each area of ​​the image at position (x, y), and are the average values ​​of the base template and the image area, respectively.

[0063] See also Figure 3 Based on an implementation of the above step S20, the area to be cleaned is identified according to the drain area image, including:

[0064] Step S210, extracting edge features of the drain and the area to be cleaned by using an edge detection algorithm;

[0065] Using advanced edge detection algorithms such as Canny edge detection can accurately identify the edges in the image, thus helping us distinguish between the drain, the area to be cleaned, and the background. For example, after edge detection, a binary image is obtained, in which the edge features are displayed in the form of white lines and the background is black.

[0066] Step S220, extracting the drain outlet boundary from the edge features by using a shape matching algorithm;

[0067] A shape matching algorithm (such as Hausdorff distance, shape context, etc.) is used to compare the template with the edge features, so as to find the edge area that best matches the template, that is, the boundary of the drain.

[0068] In an alternative solution of step S220, a circular fitting may also be performed on the edge features to find the corresponding drain port boundary;

[0069] Step S230: performing image segmentation based on the drain outlet boundary to extract the area to be cleaned.

[0070] A region-based image segmentation algorithm (such as region growing, split merging, etc.) is used; the image can be divided into different regions according to the drain outlet boundary, and the area to be cleaned that is closely connected to the drain outlet can be identified; through the above steps, the area to be cleaned can be effectively identified based on the drain outlet area image, the edge features in the image can be clarified, and the shape matching algorithm can accurately extract the drain outlet boundary.

[0071] In an embodiment of the present application, the automatic cleaning method further includes:

[0072] Step S410, obtaining the area of ​​the area to be cleaned, and when the area of ​​the area to be cleaned is greater than a preset area threshold, triggering the control robot arm to pick up the oxygen gun from the oxygen gun rack;

[0073] After identifying the area to be cleaned (such as through the previous step S230), the area of ​​the area is calculated; the calculated area is compared with a preset area threshold; if the area of ​​the area to be cleaned is greater than the area threshold, an instruction is issued to trigger the control robot arm to move to the position of the oxygen gun rack and pick up the oxygen gun.

[0074] and / or

[0075] Step S420: At each preset time interval, the robot arm is triggered to pick up the oxygen lance from the oxygen lance rack.

[0076] Set a preset time interval, such as every 2 hours; use a timer or similar timing device to track the time interval; when the timer reaches the preset time interval, issue a command to trigger the control robot arm to move to the position of the oxygen gun rack and pick up the oxygen gun.

[0077] That is, the present application can adopt two triggering mechanisms. One is to accurately calculate the area of ​​the area to be cleaned and promptly trigger the robot arm to pick up the oxygen gun when the area exceeds the threshold, so as to ensure that the cleaning work is carried out when it is most needed, thereby improving the timeliness and efficiency of cleaning; the other is a triggering mechanism with a preset time interval to ensure the regular execution of the cleaning work. Regardless of the actual size of the area to be cleaned, the continuity and regularity of the cleaning can be maintained, which helps to maintain the long-term stable operation of the equipment. The two methods can be carried out simultaneously, and only one of them needs to be met; or only one triggering method can be selected.

[0078] Since the sewer has Figure 7 The various models shown take into account that there are various models of water outlets, and the diameter of some water outlets matches the size of the nozzle at the end of the oxygen lance, so that direct sealed installation of the oxygen lance can be achieved; for the matching water outlets, the robot arm is controlled to install the oxygen lance directly and tightly to the water outlet to ensure a good seal between the two; the sealed installation helps prevent the oxygen-rich gas from leaking during the injection process, thereby improving the cleaning effect; after the oxygen lance is installed in place, the gas injection system is started to spray oxygen-rich gas to the water outlet to meet the preset injection time.

[0079] See also Figure 4 Correspondingly, the above-mentioned control robot arm picks up the oxygen gun from the preset oxygen gun rack, and controls the oxygen gun to spray oxygen-rich gas to the water outlet based on the working coordinate, including:

[0080] Step S310, controlling the robot arm to pick up the oxygen lance from the oxygen lance rack;

[0081] Step S320, controlling the oxygen gun to be connected to the water outlet according to the working coordinates, and determining whether it is installed in place according to the torque feedback of the mechanical sensor installed at the water outlet;

[0082] Step S330: When the device is installed in place, trigger the oxygen gun to spray oxygen-rich gas to the water outlet.

[0083] After identifying the area to be cleaned and triggering the robot arm to pick up the oxygen gun (as in the previous step), the robot arm is controlled to accurately install the oxygen gun at the drain outlet; during the installation process, it can be ensured that the connection between the oxygen gun and the drain outlet is tight and stable to ensure the effect of subsequent injection of oxygen-rich gas.

[0084] Once the oxygen gun is installed, the gas injection system is immediately started to spray oxygen-rich gas to the downspout; the injection time should meet the preset requirements to ensure that the impurities in the downspout are fully oxidized and cleaned. During the injection process, parameters such as gas flow, pressure and temperature can be monitored in real time to ensure the injection effect and safety.

[0085] During the oxygen blowing process, adjusting the control parameters of the oxygen gun based on the area of ​​the area to be cleaned includes: mapping the corresponding gas pressure according to the area of ​​the area to be cleaned.

[0086] According to the mapping rules, corresponding air pressure values ​​are set for different areas to be cleaned; these air pressure values ​​should be able to ensure that during the cleaning process, neither incomplete cleaning will occur due to too low air pressure, nor energy will be wasted due to too high air pressure, resulting in too large sparks and environmental pollution.

[0087] After the oxygen-rich gas is sprayed for a preset time, the robot arm is controlled to start disassembling the oxygen gun. During the disassembly process, the disassembly torque is monitored by a mechanical sensor installed at the water outlet. When the torque feedback from the mechanical sensor reaches the preset disassembly threshold, it can be determined that the oxygen gun has been disassembled. This step helps to ensure the accuracy and safety of disassembly and avoid incomplete disassembly or damage to the equipment.

[0088] After confirming that the oxygen gun is disassembled in place, control the robot arm to reset to the preset reset point.

[0089] See also Figure 5 , the automatic cleanup method also includes:

[0090] Step S510: After the oxygen lance is installed in place and the oxygen-rich gas is sprayed to the water outlet for a preset spraying time, the robot arm is controlled to disassemble the oxygen lance;

[0091] Step S520, determining whether the disassembly is complete according to the torque feedback of the mechanical sensor installed at the drain outlet;

[0092] Step S530: After the disassembly is completed, the robot arm is controlled to reset to the reset point.

[0093] After the oxygen lance has been correctly installed at the drain and has completed the injection of oxygen-rich gas according to the preset injection time, the oxygen lance will be removed by controlling the robot arm; ensuring that the oxygen lance can be safely and accurately removed from the drain after completing its cleaning task for subsequent cleaning or maintenance work.

[0094] Similarly, using the mechanical sensor installed at the drain, the system will continuously monitor the torque changes during the disassembly process and judge whether the oxygen lance has been completely disassembled. Through real-time monitoring and feedback, the accuracy and safety of the disassembly process can be ensured to prevent equipment damage or personal injury caused by incomplete or excessive disassembly. Once it is confirmed that the oxygen lance has been disassembled, the system will control the robot arm to reset to the initial reset point according to the preset path and speed. Ensure that the robot arm can quickly and accurately return to the starting position after completing the cleaning task until the next cleaning task is triggered.

[0095] See also Figure 6 , correspondingly, the automatic cleaning method also includes:

[0096] Step S610: generating a cleaning path according to the area to be cleaned, where the cleaning path includes an initial point and an end point;

[0097] Step S620, generating a motion path of the robot arm based on the oxygen lance rack coordinate information on the oxygen lance rack, the current position of the end of the robot arm, and the cleaning path;

[0098] Based on the specifications of the oxygen lance and non-standard cost considerations, this application uses a consistent oxygen lance; when the diameter of the downspout is larger than the end nozzle of the oxygen lance, a mechanism needs to be designed to ensure that the oxygen lance moves effectively in the downspout to achieve comprehensive cleaning.

[0099] It is understandable that when the diameter of the drain is larger than the end nozzle of the oxygen lance, the oxygen lance can move in the entire drain instead of being directly installed and plugged into the drain. Use image recognition to determine the specific area in the drain that needs to be cleaned; based on the characteristics of the area to be cleaned, generate one or more cleaning paths. These paths should include the starting point (the position where the oxygen lance starts to move) and the end point (the position where the oxygen lance stops moving).

[0100] The motion path of the robot arm is generated by combining the cleaning path, the oxygen lance coordinate information and the current position of the end of the robot arm. This path should ensure that the robot arm can smoothly and accurately control the oxygen lance to move along the cleaning path in the drain.

[0101] Based on the above, in step S30, the robot arm is controlled to pick up the oxygen lance from the preset oxygen lance rack, and the oxygen lance is controlled to spray oxygen-rich gas to the water outlet based on the working coordinates, and the control parameters of the oxygen lance are adjusted according to the area to be cleaned, including:

[0102] Step S310', the robot arm moves according to the motion path, and controls the oxygen lance to spray oxygen-rich gas to the water outlet only in the cleaning path stage, and adjusts the control parameters of the oxygen lance according to the area to be cleaned.

[0103] That is, during the entire cleaning process, the spraying of the oxygen gun is precisely controlled and is carried out only in the area and time that needs to be cleaned; before or after the robot arm moves to the cleaning path, that is, when the robot arm moves outside the drain or prepares / ends the cleaning task, the oxygen gun usually does not spray oxygen-rich gas; while the robot arm moves along the cleaning path, the control system controls the spraying of the oxygen gun according to preset parameters (such as spraying pressure, distance, etc.).

[0104] In an implementation manner corresponding to step S610, generating a cleaning path according to the area to be cleaned includes:

[0105] Step S611, dividing the boundary of the drain into a plurality of quadrants of equal angles according to 360 degrees;

[0106] This step is achieved using geometric methods or image processing techniques. The drain is a circular area. From a two-dimensional perspective, the circle can be divided into 8, 16 or more sectors (quadrants) with equal angles, such as quadrant 1, quadrant 2, quadrant 3, etc. The angle of each quadrant can be calculated by dividing 360 degrees by the number of divisions; the area within the drain boundary is divided into multiple parts to facilitate separate analysis and management of each part. It helps to locate the area where molten iron accumulates more accurately and formulate corresponding cleaning strategies.

[0107] Step S612, counting the area of ​​molten iron in the area to be cleaned in each quadrant;

[0108] Each quadrant is cropped to obtain a separate quadrant image so that the pixel count of the molten iron area in each quadrant can be performed separately; since the image resolution and the actual size of the water outlet area are known, the number of pixels can be converted into actual area units through simple mathematical calculations. For example, the area of ​​the molten iron area can be obtained by calculating the actual area represented by each pixel and then multiplying the number of pixels by this area.

[0109] Step S613: when the molten iron area is greater than a preset area threshold, the corresponding quadrant is marked as a target quadrant;

[0110] The molten iron area in each quadrant is compared with the set threshold. When the molten iron area is greater than the threshold, the corresponding quadrant is marked as the target quadrant.

[0111] Step S614: Calculate the shortest path on the preset cleaning trajectory according to the target quadrant as the cleaning path.

[0112] According to the above-mentioned molten iron area statistics, after determining which quadrants are the target quadrants, one or more cleaning trajectories are preset according to the capabilities of the robot arm and the characteristics of the experimental area, such as a circular trajectory.

[0113] Based on the location of the target quadrants and the preset cleaning trajectories, all possible cleaning paths are generated; these paths should cover all target quadrants and minimize repeated and redundant cleaning work.

[0114] Use an appropriate distance calculation method (such as Euclidean distance, Manhattan distance, etc.) to calculate the total length of each possible path, compare the lengths of all possible paths, and select the shortest one as the optimal cleaning path.

[0115] For example, suppose there are eight quadrants, numbered 1 to 8, where the 1st, 5th, and 6th quadrants are the target quadrants. The cleaning trajectory preset by the robot arm is a circular motion. Then the possible cleaning paths generated are:

[0116] Path 1: Starting from the center of quadrant 1, passing through quadrants 2, 3, and 4, reaching the center of quadrant 5, and then passing through quadrant 6; this path forms an incomplete circle but covers all target quadrants.

[0117] Path 2: Starting from the center of the 6th quadrant, passing through the 7th and 8th quadrants, reaching the center of the 1st quadrant, and then passing through the 2nd, 3rd, and 4th quadrants to reach the center of the 5th quadrant; this path also covers all target quadrants, but in a different order.

[0118] Using the distance calculation method, calculate the total length of path 1 and path 2 separately. This usually involves calculating the distance the robot moves between adjacent quadrants and adding these distances to get the total length; compare the lengths of all possible paths and select the shortest one as the optimal cleaning path

[0119] In one possible implementation of step S620, the motion path of the robot arm is generated based on the oxygen lance rack coordinate information on the oxygen lance rack, the current position of the end of the robot arm, and the cleaning path, including:

[0120] Step S621, reading the current oxygen lance rack coordinate information from the oxygen lance rack control system, the oxygen lance rack coordinate information including the remaining oxygen lance coordinates and the remaining placement position coordinates;

[0121] Step S622, calculating a first path from the current position of the end of the machine arm to the remaining oxygen lance coordinates, a second path from the remaining oxygen lance coordinates to the initial point, a third path from the end point to the remaining placement position coordinates, and a fourth path from the placement position coordinates to the reset point;

[0122] Step S623: generate a motion path by combining the first path, the second path, the third path and the fourth path.

[0123] The remaining oxygen lance coordinates represent the specific positions of the oxygen lances that have not been used yet; the remaining placement position coordinates represent the free positions where the oxygen lances can be placed after use; in the above, the first path represents the specific coordinates from the current position of the robot arm to the remaining oxygen lances; the second path represents the position of the oxygen lance to the start of the task or the initial point of the operation; the third path represents the return from the end point of the operation to the remaining placement positions of the oxygen lance after the task is completed; the fourth path represents moving from the placement position to the reset point of the robot arm to prepare for the next task, where the initial point and the end point are the results obtained in step S610.

[0124] Integrate the path data of the first path, the second path, the third path and the fourth path, that is, integrate the starting point, the end point and the key points in the middle of each path into a path data structure; this data structure can be a list, an array or other data structure suitable for storing multiple point coordinates.

[0125] If there are inflection points or sudden changes in direction between paths, these points may need to be smoothed to ensure that the robot can smoothly transition from one path to another. This can be achieved through interpolation algorithms (such as linear interpolation, polynomial interpolation, etc.).

[0126] For example, there are two known points P1 (x1, y1, z1) and P2 (x2, y2, z2), which represent two adjacent positions on the robot path; the interpolation parameter t is usually a floating point number in the range [0, 1]. When t = 0, the interpolation result is equal to P1; when t = 1, the interpolation result is equal to P2; when t is between (0, 1), the interpolation result is between P1 and P2.

[0127] The coordinates of the interpolation point are calculated using the linear interpolation formula: Pinterp(x,y,z)=(1-t)·P1(x1,y1,z1)+t·P2(x2,y2,z2).

[0128] Select a series of t values ​​(e.g., equally spaced points from 0 to 1) and apply the above formula to each t value to generate a series of interpolated points. These points will constitute smooth transitions on the robot's path. Use the generated sequence of interpolated points as key points in the robot's path planning to ensure that the robot can move smoothly from one point to the next.

[0129] Based on the above, in step S30, the oxygen lance is controlled to spray oxygen-rich gas to the water outlet only in the cleaning path stage, and the control parameters of the oxygen lance are adjusted according to the area to be cleaned, including:

[0130] Step S331, dividing the boundary of the drain into a plurality of quadrants of equal angles according to 360 degrees;

[0131] Step S332, counting the area of ​​molten iron in the area to be cleaned in each quadrant;

[0132] Step S333: In the cleaning path stage, the control parameters of the oxygen lance in different quadrants are adjusted according to the molten iron area.

[0133] As above, determine the center point of the drain and establish a polar coordinate system with it as the center; divide 360 ​​degrees into multiple quadrants, such as 8 (one quadrant every 45 degrees) or 16 (one quadrant every 22.5 degrees), the specific number depends on the required accuracy, and the division method is consistent with the above path planning; each quadrant is defined by a starting angle and an ending angle, for example, the first quadrant is 0 degrees to 45 degrees (or -22.5 degrees to 22.5 degrees, depending on how the position of 0 degrees is defined).

[0134] The molten iron region is identified by image processing algorithms (such as threshold segmentation, edge detection, etc.); the area of ​​the molten iron region in each quadrant is calculated, and pixel counting or other area measurement methods can be used.

[0135] Specifically, in each quadrant, the control parameters of the oxygen lance are adjusted according to the size and distribution of the molten iron area; for example, in a quadrant with a larger molten iron area, the air pressure may need to be increased to speed up the blowing rate; while in a quadrant with a smaller molten iron area, the air pressure parameters need to be appropriately reduced, thereby achieving dynamic adjustment throughout the cleaning path stage.

[0136] In a specific scheme, in the cleaning path stage, the control parameters of the oxygen lance are adjusted according to the molten iron area, which can be achieved by establishing a weighted correlation between the area correlation and the air pressure and the oxygen blowing distance. This correlation can dynamically adjust the air pressure (P) and oxygen blowing distance (D) of the oxygen lance according to the size of the molten iron area to achieve the best cleaning effect.

[0137] First, define a molten iron area ratio (Area_Ratio), which represents the ratio of the molten iron area in each quadrant to the total area of ​​the quadrant; this ratio value is between 0 and 1: Area_Ratioi=Total_Area i / Iron_Areai.

[0138] Among them, Iron_Area i is the molten iron area in the i-th quadrant, and Total_Area i is the total area of ​​the i-th quadrant; then the gas pressure (P) and oxygen blowing distance (D) of the oxygen lance are calculated by the following correlation formula:

[0139] P i=Pbase+kP·Area_Rat ioi(1)

[0140] D i=Dbase-kD·Area_Rat ioi(2)

[0141] Among them, Pi and Di are the air pressure and oxygen blowing distance corresponding to the i-th quadrant, respectively; Pbase and Dbase are the basic air pressure and basic oxygen blowing distance, respectively, which are the default values ​​when the molten iron area ratio is 0 (that is, when there is no molten iron to be cleaned); kP and kD are the weighted coefficients of air pressure and oxygen blowing distance, respectively, which determine the influence of the molten iron area ratio on the air pressure and oxygen blowing distance.

[0142] In order to solve the problems in the prior art, the present invention also provides a computer-readable storage medium, which stores one or more programs. The one or more programs can be executed by one or more processors to implement the above-mentioned automatic cleaning method for the ladle drain.

[0143] In the present invention, unless otherwise clearly specified and limited, the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. The first feature being "up" or "down" the second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, the first feature being "above", "above" and "above" the second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. The first feature being "below", "below" and "below" the second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature, and therefore cannot be understood as limiting the present invention.

[0144] The preferred embodiments of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited thereto. Within the technical concept of the present invention, the technical solution of the present invention can be subjected to a variety of simple modifications, including the combination of various specific technical features in any appropriate manner. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations. However, these simple modifications and combinations should also be regarded as the contents disclosed by the present invention and belong to the protection scope of the present invention.

[0145] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. An automatic cleaning method for a ladle outlet, characterized in that: include: After controlling the ladle to tip over, collect the on-site image of the ladle outlet; By comparing pixels of a preset reference template with the on-site image in sequence, the image of the drain area is extracted and the working coordinates of the image of the drain area are located; Extracting the drain port boundary in the drain port area image and segmenting the area to be cleaned by edge detection; The robot arm is controlled to pick up the oxygen gun from the preset oxygen gun rack, and the oxygen gun is controlled to spray oxygen-rich gas to the water outlet based on the working coordinates; wherein, during the oxygen blowing process, the control parameters of the oxygen gun are adjusted based on the area of ​​the area to be cleaned, and the control parameters include at least air pressure and oxygen blowing distance.

2. The automatic cleaning method for the ladle outlet according to claim 1, characterized in that: The pixel comparison between the preset reference template and the on-site image in sequence to extract the drain area image and locate the working coordinates of the drain area image includes: Preprocessing the scene image, and performing pixel-by-pixel comparison with each region of the preprocessed scene image through a preset basic template to calculate the region similarity; Sort the regional similarities, and take the region corresponding to the highest regional similarity as the benchmark region; According to the positional relationship between the reference area and the drain, the drain area image is extracted from the on-site graphics and the working coordinates of the drain are located.

3. The automatic cleaning method for the ladle outlet according to claim 2, characterized in that: The preprocessing of the scene image and comparing each region of the preprocessed scene image pixel by pixel using a preset basic template to calculate the region similarity comprises: Preprocessing the scene image by Gaussian filtering; Input the basic template and the site graph, and use the following formula to calculate the regional similarity: Where T(i, j) is the pixel value of the base template, I(x+i, y+j) is the pixel value corresponding to each area of ​​the image at position (x, y), and T- and I-(x, y) are the average values ​​of the base template and the image area, respectively.

4. The automatic cleaning method for a ladle outlet according to claim 1, characterized in that: The automatic cleaning method also includes: Acquire the area of ​​the area to be cleaned, and when the area of ​​the area to be cleaned is greater than a preset area threshold, trigger the control robot arm to pick up the oxygen gun from the oxygen gun rack; and / or At every preset time interval, the robot arm is triggered to pick up the oxygen gun from the oxygen gun rack.

5. The automatic cleaning method for the ladle outlet according to claim 4, characterized in that: The diameter of the water outlet is consistent with the end nozzle of the oxygen lance, and the control robot arm picks up the oxygen lance from the oxygen lance rack, and controls the oxygen lance to spray oxygen-rich gas to the water outlet at the water outlet, including: Controlling the robot arm to pick up the oxygen lance from the oxygen lance rack; Controlling the oxygen gun to be connected to the water outlet according to the working coordinates, and determining whether it is installed in place according to the torque feedback of the mechanical sensor installed at the water outlet; When installed in place, the oxygen gun is triggered to spray oxygen-rich gas to the water outlet; In the oxygen blowing process, the control parameters of the oxygen gun are adjusted based on the area of ​​the area to be cleaned, including: Map the corresponding air pressure according to the area to be cleaned.

6. The automatic cleaning method for a ladle drain port according to any one of claims 1 to 5, characterized in that: The method of extracting the drain port boundary in the drain port area image by edge detection and segmenting the area to be cleaned includes: The edge features of the drain and the area to be cleaned are extracted by edge detection algorithm; The drain boundary is extracted from the edge features by shape matching algorithm; Image segmentation is performed through the drain port boundary to extract the area to be cleaned.

7. The automatic cleaning method for the ladle outlet according to claim 4, characterized in that: The automatic cleaning method also includes: After the oxygen lance is installed in place and the oxygen-rich gas is sprayed to the water outlet for a preset spraying time, the robot arm is controlled to disassemble the oxygen lance; Determine whether the disassembly is in place based on the torque feedback of the mechanical sensor installed at the drain port; After being disassembled into place, the control robot arm is reset.

8. The automatic cleaning method for a ladle drain port according to any one of claims 1 to 5, characterized in that: The diameter of the water outlet is larger than the end nozzle of the oxygen lance, and the automatic cleaning method further comprises: Generate a cleaning path according to the area to be cleaned, wherein the cleaning path includes an initial point and an end point; Generate a motion path of the robot arm based on the oxygen gun rack coordinate information on the oxygen gun rack, the current position of the end of the robot arm, and the cleaning path; The control robot arm picks up the oxygen gun from the preset oxygen gun rack, and controls the oxygen gun to spray oxygen-rich gas to the water outlet based on the working coordinates, and adjusts the control parameters of the oxygen gun according to the area to be cleaned, including: The robot arm moves according to the motion path, and controls the oxygen lance to spray oxygen-rich gas to the water outlet only in the cleaning path stage, and adjusts the control parameters of the oxygen lance according to the area to be cleaned.

9. The automatic cleaning method for the ladle drain port according to claim 8, characterized in that: The motion path of the robot arm is generated based on the oxygen gun rack coordinate information on the oxygen gun rack, the current position of the end of the robot arm, and the cleaning path, including: Reading the current oxygen lance rack coordinate information from the oxygen lance rack control system, wherein the oxygen lance rack coordinate information includes the remaining oxygen lance coordinates and the remaining placement position coordinates; Calculate a first path from the current position of the end of the machine arm to the remaining oxygen lance coordinates, a second path from the remaining oxygen lance coordinates to the initial point, a third path from the end point to the remaining placement position coordinates, and a fourth path from the placement position coordinates to the reset point; generating a motion path by combining the first path, the second path, the third path and the fourth path; The step of generating a cleaning path according to the area to be cleaned includes: When the molten iron area is greater than a preset area threshold, marking the corresponding quadrant as a target quadrant; The shortest path is calculated based on the preset cleaning trajectory of the target quadrant as the cleaning path.

10. The automatic cleaning method for the ladle drain port according to claim 9, characterized in that: The method of controlling the oxygen lance to spray oxygen-rich gas to the water outlet only in the cleaning path stage and adjusting the control parameters of the oxygen lance according to the area to be cleaned includes: Divide the boundary of the drain into a plurality of quadrants of equal angles according to 360 degrees; Count the area of ​​molten iron in the area to be cleaned in each quadrant; In the cleaning path stage, the control parameters of the oxygen lance in different quadrants are adjusted according to the molten iron area.

Images