Intelligent machining method and system based on stainless steel welded steel pipe seamlessness

Abstract

According to the intelligent machining method and system based on seamless stainless steel welded steel pipes, a preset stainless steel plate is heated to the forming temperature and sequentially passes through a plate pre-flanging module, a U-shaped forming module and an O-shaped forming module, and a pre-flanging plate, a U-shaped plate and an O-shaped seamed pipe are gradually formed. Thermal imaging data of a pre-bent edge section, a U-shaped section, an O-shaped section and a joint are obtained in the machining process, the thermal imaging data collected in all forming stages are analyzed through a pre-trained welding seam quality estimation model, and the welding seam quality score of the O-shaped seamed pipe is predicted before actual welding operation. Through predictive quality evaluation, the rejection rate and the rework cost are effectively reduced; the production efficiency and the stability of the product quality are improved; and a reliable technical guarantee is provided for intelligent production of the stainless steel welded steel pipe.

Description

Technical Field

[0001] The present application relates to the field of metal forming, and in particular to an intelligent processing method and system based on seamless stainless steel welded pipes. Background Art

[0002] Stainless steel welded pipes, as an indispensable basic material in modern industry, are widely used in high-demand fields such as petrochemicals, food processing, medical equipment, and nuclear energy. Their processing and manufacturing technology directly affects the quality of the product and subsequent application performance.

[0003] At present, the main processing methods for stainless steel welded pipes include argon arc welding, plasma welding, laser welding and resistance welding. As a traditional process, argon arc welding has the advantages of good weld formation and relatively simple operation, but it has serious defects such as slow welding speed, wide heat-affected zone, and easy generation of pores and slag inclusions. Although plasma welding can achieve a higher welding speed, the equipment is complex, the cost is high, and the operating environment requirements are extremely strict. The slightest carelessness will lead to unstable weld quality. Although laser welding technology has high precision and a small heat-affected zone, the equipment investment is huge, the maintenance cost is high, and the surface cleanliness of the workpiece is stringent. Once there is an oxide layer or contaminants on the surface, welding defects are very likely to occur. The resistance welding process is simple and efficient, but the weld strength is relatively low, and it is difficult to achieve welding of complex cross-sections.

[0004] Common drawbacks of these traditional processing methods include uneven microstructure, concentrated residual stress, and reduced corrosion resistance in the weld area. In particular, grain coarsening at the weld seam severely impacts the overall performance of the material, significantly reducing the reliability of welded steel pipes in high-temperature, high-pressure, or corrosive environments. Furthermore, defects such as spatter, lack of fusion, and undercutting during welding not only impact the appearance but also severely limit the product's service life.

[0005] Weld quality is crucial for subsequent seamless fabrication. Seamless fabrication requires a good metallurgical bond between the weld and the base material, a uniform and dense microstructure, and no significant strength differences. If defects such as porosity, slag inclusions, and cracks are present in the original weld, these defects will further expand under the high temperature and high pressure conditions of seamless fabrication, leading to fabrication failure or even product rejection. The weld's chemical composition uniformity, microstructural continuity, and residual stress distribution directly determine the success rate of seamless fabrication and the performance of the final product. Therefore, optimizing the welding process and improving weld quality are key technical prerequisites for achieving high-quality seamless fabrication. Summary of the Invention

[0006] The present application provides an intelligent processing method for seamless stainless steel welded pipes, comprising the following steps: A1, heating the preset stainless steel plate to the preset forming temperature; A2, passing the heated stainless steel sheet through a preset sheet pre-bending module to form a pre-bent sheet; A3, obtaining thermal imaging data of the pre-bent edge cross section of the pre-bent edge plate; A4, passing the pre-bent sheet through a preset U-forming module to form a U-shaped sheet; A5, obtains the thermal imaging data of the U-shaped cross section of the U-shaped plate; A6, passes the U-shaped sheet through the preset O-shaped forming module to form an O-shaped seamed pipe; A7, obtains the O-shaped cross-section thermal imaging data and the seam thermal imaging data of the O-shaped seamed pipe; A8, based on the pre-bent edge section thermal imaging data, U-section thermal imaging data, O-section thermal imaging data, and seam thermal imaging data of stainless steel plates, uses a pre-trained weld quality prediction model to generate an estimated weld quality score for O-shaped seamed pipes.

[0007] By adopting the above technical solution, the intelligent processing method based on seamless stainless steel welded steel pipes can preheat the plate and collect thermal imaging data at each node of the forming process, and combine it with a pre-trained weld quality prediction model to predict the weld quality before welding, realizing the transition from passive quality detection to active quality control, effectively reducing the scrap rate and rework costs, and improving production efficiency and product quality stability.

[0008] Optionally, the intelligent processing method based on seamless stainless steel welded steel pipe further comprises the following steps: A9: If the estimated weld quality score corresponding to the O-shaped seam pipe is greater than or equal to the preset weld score qualification threshold, the seam of the O-shaped seam pipe is welded by the preset welding module to form a stainless steel welded pipe; A10, if the estimated weld quality score corresponding to the O-seam pipe is less than the weld score passing threshold, a worker manually evaluates and corrects the seam of the O-seam pipe. After the correction, the seam of the O-seam pipe is welded using the welding module to form a stainless steel welded pipe. A11, obtaining weld image data of stainless steel welded pipe; A12, obtaining ultrasonic testing data of the weld seam of the stainless steel welded pipe through a preset ultrasonic testing module; A13, evaluates the welds of stainless steel welded pipes based on weld image data and weld ultrasonic testing data and generates corresponding actual weld quality scores; A14: If the actual weld quality score is greater than or equal to the weld score qualification threshold, the corresponding stainless steel welded pipe is defined as a qualified stainless steel welded pipe.

[0009] By adopting the above technical solution, the intelligent processing method based on seamless stainless steel welded steel pipes can realize intelligent prediction before welding and comparison verification after welding through a dual quality control system of estimated scoring and actual scoring. When the estimated quality does not meet the standard, manual intervention and correction can be carried out in time to avoid unqualified products from entering the welding process. At the same time, combined with internal and external evaluations of weld image data and ultrasonic detection data, the reliability and consistency of the final product quality are ensured, thereby improving production efficiency and product qualification rate.

[0010] Optionally, the intelligent processing method based on seamless stainless steel welded steel pipe further comprises the following steps: B1, obtaining outer wall image data and inner wall image data of a qualified stainless steel welded pipe; B2, determining the corresponding weld pipe to-be-cleaned area data using a preset cleaning area recognition algorithm based on the weld seam image data, the outer wall image data, and the inner wall image data; B3, clean the inner wall, outer wall and weld of the qualified stainless steel welded pipe in a preset welded pipe cleaning method according to the data of the welded pipe area to be cleaned; B4, obtaining weld seam image data, outer wall image data, and inner wall image data of the qualified stainless steel welded pipe after cleaning and defining them as weld seam cleaning image data, outer wall cleaning image data, and inner wall cleaning image data, respectively; B5, calculating the corresponding weld pipe cleanliness based on the weld cleaning image data, the outer wall cleaning image data, and the inner wall cleaning image data; B6, if the cleanliness of the welded pipe is greater than or equal to a preset cleanliness threshold, the qualified stainless steel welded pipe after the cleaning operation is expanded according to a preset expansion size through a preset pipe expansion module to form a stainless steel seamless welded pipe of a predetermined size.

[0011] By adopting the above technical solution, the intelligent processing method based on seamless stainless steel welded steel pipes can locate the area to be cleaned of the welded pipe through the recognition algorithm based on the image of the qualified stainless steel welded pipe, thereby realizing targeted cleaning operations and avoiding the waste of resources of traditional comprehensive cleaning. At the same time, the image data is used to quantitatively evaluate the cleanliness to ensure that the pipe is expanded only after the cleaning effect meets the standards, and finally a seamless welded pipe of a predetermined size is formed, which improves production efficiency while reducing the cost of manual quality inspection and ensures the surface quality and dimensional accuracy of the final product.

[0012] Optionally, step A8 includes the following steps: A801, generating corresponding pre-bend section line feature data based on the pre-bend section thermal imaging data using a preset section image feature extraction algorithm; A802, generating corresponding U-shaped cross-section line feature data based on the U-shaped cross-section thermal imaging data by using a cross-section image feature extraction algorithm; A803 generates corresponding O-section line feature data based on the O-section thermal imaging data using a cross-section image feature extraction algorithm; A804, generating corresponding seam line feature data based on the seam thermal imaging data using a preset seam image feature extraction algorithm; A805 generates a corresponding estimated weld quality score through a weld quality estimation model based on the pre-bent edge section line feature data, U-section line feature data, O-section line feature data and seam line feature data.

[0013] By adopting the above technical solution, the intelligent processing method based on seamless stainless steel welded steel pipe can convert thermal imaging data into corresponding line feature data through cross-sectional image feature extraction algorithm and seam image feature extraction algorithm, thereby realizing accurate quantification and standardized processing of key geometric parameters in the forming process, so that the weld quality prediction model can make predictions based on more accurate and consistent input data, which can improve the prediction accuracy and the generalization ability of the model, and provide effective data support for achieving high-precision weld quality pre-control.

[0014] Optionally, the cross-sectional image feature extraction algorithm includes the following steps: C1, determine the effective range of grayscale value according to the molding temperature; C2, generating corresponding plate cross-section thermal imaging grayscale image data according to the preset plate cross-section thermal imaging image data; C3, traverse each pixel point in the grayscale image data of the thermal imaging of the plate cross section, and if the grayscale value of the pixel point is not within the valid grayscale value range, assign the grayscale value of the pixel point to a preset default value; C4, after traversing each pixel point in the grayscale image data of the thermal imaging of the plate cross section, generates the plate cross section grayscale image processing data; C5, intercepting the grayscale image data of the corresponding effective area of ​​the plate cross section according to the preset default value and the plate cross section grayscale image processing data; C6, generating corresponding plate cross-section grayscale image edge feature data according to the plate cross-section effective area grayscale image data through a preset edge feature extraction algorithm and defining it as corresponding plate cross-section line feature data.

[0015] By adopting the above technical solution, the intelligent processing method based on seamless stainless steel welded steel pipes can effectively filter out noise and environmental information in the thermal imaging image by screening the effective range of grayscale values ​​based on the forming temperature, and then accurately locate the key feature areas of the plate cross section through default value processing and effective area interception. Finally, the edge feature extraction algorithm is used to generate standardized line feature data, realizing the effective conversion from raw thermal imaging data to structured feature data, and providing more accurate, stable and reliable input data for subsequent weld quality estimation.

[0016] Optionally, the seam image feature extraction algorithm includes the following steps: D1, determine the effective range of grayscale value according to the molding temperature; D2, generating corresponding seam thermal imaging grayscale image data according to the seam thermal imaging image data; D3, traverse each pixel point in the seam thermal imaging grayscale image data, and if the grayscale value of the pixel point is not within the valid grayscale value range, assign the grayscale value of the pixel point to a preset default value; D4, after traversing each pixel point in the seam thermal imaging grayscale image data, generating seam grayscale image processing data; D5, determining corresponding seam grayscale image valid area positioning data according to the preset default value and the seam grayscale image processing data; D6, obtaining an image of a corresponding area in the seam thermal imaging grayscale image data according to the seam grayscale image effective area positioning data and defining the image as the seam effective area grayscale image data; D7, traverse each pixel point in the grayscale image data of the effective seam area, and if the grayscale value of the pixel point is within the effective grayscale value range, assign the grayscale value of the pixel point to a preset default value; D8, after traversing each pixel point in the grayscale image data of the effective seam area, generating grayscale image processing data of the effective seam area; D9, generating corresponding edge feature data of the grayscale image of the effective seam area through an edge feature extraction algorithm according to the grayscale image processing data of the effective seam area and defining the edge feature data as the corresponding seam line feature data.

[0017] By adopting the above technical solution, the intelligent processing method based on seamless stainless steel welded steel pipe can use a dual grayscale value processing mechanism to first filter out invalid environmental pixels and locate the effective area of ​​the seam, and then reversely process the pixels in the effective area to highlight the seam boundary features, thereby realizing the recognition and feature enhancement of the seam area, and improving the contrast of the seam area in the thermal imaging image. The seam line feature data finally generated can better reflect the geometric characteristics of the seam, provide higher-quality input data for the weld quality prediction model, and improve the accuracy and reliability of welding quality prediction.

[0018] Optionally, the cleaning area identification algorithm includes the following steps: E1, generating corresponding weld grayscale image data, outer wall grayscale image data and inner wall grayscale image data respectively according to the weld image data, outer wall image data and inner wall image data; E2, respectively calculating the corresponding average grayscale value of the weld grayscale image, the average grayscale value of the outer wall grayscale image, and the average grayscale value of the inner wall grayscale image according to the grayscale value of each pixel point in the weld grayscale image data, the outer wall grayscale image data, and the inner wall grayscale image data; E3, segmenting the weld grayscale image data, the outer wall grayscale image data, and the inner wall grayscale image data according to a preset segmentation grid to obtain corresponding weld grid grayscale image data, outer wall grid grayscale image data, and inner wall grid grayscale image data; E4, calculating the standard deviation of the grayscale values ​​of all pixels in each weld mesh grayscale image data according to the average grayscale value of the weld grayscale image and defining it as the standard deviation of the weld mesh grayscale image; E5, if the standard deviation of the weld mesh grayscale image is greater than the preset mesh grayscale standard deviation threshold, then define the area of ​​the corresponding weld mesh grayscale image data in the segmented mesh as the area to be cleaned; E6, calculating the standard deviation of the grayscale values ​​of all pixels in each outer wall grid grayscale image data according to the average grayscale value of the outer wall grayscale image and defining it as the standard deviation of the outer wall grid grayscale image; E7, if the standard deviation of the outer wall grid grayscale image is greater than the grid grayscale standard deviation threshold, then define the area of ​​the corresponding outer wall grid grayscale image data in the segmented grid as the area to be cleaned; E8, calculating the standard deviation of the grayscale values ​​of all pixels in each inner wall grid grayscale image data according to the average grayscale value of the inner wall grayscale image and defining it as the standard deviation of the inner wall grid grayscale image; E9, if the standard deviation of the inner wall grid grayscale image is greater than the grid grayscale standard deviation threshold, then define the area of ​​the corresponding inner wall grid grayscale image data in the segmented grid as the area to be cleaned; E10, combines all areas to be cleaned to generate the data of the welded pipe area to be cleaned.

[0019] By adopting the above technical solution, the intelligent processing method based on seamless stainless steel welded steel pipes can identify abnormal areas on the welds, outer walls and inner wall surfaces through grid segmentation and standard deviation statistical analysis, and automatically locate the specific locations of pollution, oxide scale, spatter, etc. that need to be cleaned based on the change characteristics of the gray value standard deviation, thereby achieving precise positioning and cleaning. This not only reduces the time and cost of cleaning operations, but also avoids damage to the pipe surface caused by excessive cleaning, improves cleaning efficiency and consistency of cleaning quality, and provides guarantees for subsequent pipe expansion and final product quality.

[0020] The present application also provides an intelligent processing system based on seamless stainless steel welded pipes, comprising: Forming module; Welding module; Clean up the module; Pipe expansion module; Processing control module; Wherein, the forming module, the welding module, the cleaning module and the pipe expansion module are respectively data-connected to the processing control module; The forming module includes a preheating module, a plate pre-bending module, a U-forming module, an O-shaped forming module and a thermal imaging module, and the preheating module, the plate pre-bending module, the U-forming module, the O-shaped forming module and the thermal imaging module are respectively data-connected to the processing control module; The intelligent processing system based on seamless stainless steel welded steel pipe further includes a quality control strategy, including the following steps: F1, heating the preset stainless steel plate to a preset forming temperature through the preheating module; F2, passing the heated stainless steel sheet through the sheet pre-bending module to form a pre-bent sheet; F3, obtaining thermal imaging data of a pre-bent edge cross section of the pre-bent edge plate through the thermal imaging module; F4, passing the pre-bent sheet through the U-forming module to form a U-shaped sheet; F5, obtaining thermal imaging data of the U-shaped cross section of the U-shaped plate through the thermal imaging module; F6, passing the U-shaped sheet through the O-shaped forming module to form an O-shaped seamed pipe; F7, obtaining thermal imaging data of the O-shaped cross section and the seam of the O-shaped seamed pipe by the thermal imaging module; F8, through the processing control module, generates an estimated weld quality score corresponding to the O-shaped seam pipe according to the pre-bent edge section thermal imaging data, U-shaped section thermal imaging data, O-shaped section thermal imaging data and seam thermal imaging data of the stainless steel plate using a preset weld quality estimation model.

[0021] By adopting the above technical solution, the intelligent processing system based on seamless stainless steel welded steel pipes can preheat the plates and collect thermal imaging data at each node of the forming process. In combination with the pre-trained weld quality prediction model, it can predict the weld quality before welding, realizing the transition from passive quality detection to active quality pre-control, effectively reducing the scrap rate and rework costs, and improving production efficiency and product quality stability.

[0022] In summary, this application includes at least one of the following beneficial technical effects: 1. By preheating the sheet metal and collecting thermal imaging data at various points in the forming process, combined with a pre-trained weld quality prediction model, weld quality can be predicted before welding, achieving a shift from passive quality detection to active quality control. This effectively reduces scrap rates and rework costs, and improves production efficiency and product quality stability.

[0023] 2. Through the dual quality control system of estimated scoring and actual scoring, intelligent prediction before welding and comparison verification after welding can be achieved. When the estimated quality does not meet the standard, manual intervention and correction can be carried out in time to prevent unqualified products from entering the welding process. At the same time, combined with internal and external evaluation of weld image data and ultrasonic detection data, the reliability and consistency of the final product quality are ensured, and production efficiency and product qualification rate are improved.

[0024] 3. The area to be cleaned of the welded pipe can be located through the recognition algorithm based on the image of the qualified stainless steel welded pipe, realizing targeted cleaning operations and avoiding the resource waste of traditional comprehensive cleaning. At the same time, the image data is used to quantitatively evaluate the cleanliness to ensure that the pipe is expanded only after the cleaning effect meets the standards, and finally a seamless welded pipe of predetermined size is formed, which improves production efficiency while reducing the cost of manual quality inspection and ensures the surface quality and dimensional accuracy of the final product. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a process schematic diagram of an intelligent processing method based on seamless stainless steel welded steel pipes according to the present invention.

[0026] Figure 2 It is a schematic diagram of the principle of an intelligent processing system based on seamless stainless steel welded steel pipes of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0028] The embodiments of the present application are described in further detail below with reference to the accompanying drawings.

[0029] refer to Figure 1 The present invention provides an intelligent processing method based on seamless stainless steel welded pipes, which is used to process stainless steel seamless welded pipes, including the following steps: A1, heating the preset stainless steel plate to the preset forming temperature; The forming temperature is a pre-set heating temperature used to heat the stainless steel plate to be formed to the corresponding temperature to improve its plasticity. The forming temperature can be set according to factors such as the specific processing technology, plate material, plate thickness, plate size, etc., or a fixed universal forming temperature can be set, for example, 200°C, 400°C, or a temperature between 200°C and 400°C.

[0030] A2, passing the heated stainless steel sheet through a preset sheet pre-bending module to form a pre-bent sheet; The plate pre-bending module is an equipment module for bending the edge of the plate, which can bend both sides of the plate at the same time or in sequence; Pre-bent plates are stainless steel plates with bent edges.

[0031] A3, obtaining thermal imaging data of the pre-bent edge cross section of the pre-bent edge plate; The pre-bent edge cross-section thermal imaging data is the thermal imaging data of the cross-section of the pre-bent edge plate, which can be obtained by photographing the cross-section of the plate through a thermal imaging device. Since the stainless steel plate is preheated before forming, the image of its cross-section can appear more clearly in the thermal imaging image.

[0032] A4, passing the pre-bent sheet through a preset U-forming module to form a U-shaped sheet; The U-forming module is an equipment module used to extrude pre-bent sheet materials into a U shape; The U-shaped plate is a stainless steel plate formed by the U-forming module after the pre-bent plate is formed.

[0033] A5, obtains the thermal imaging data of the U-shaped cross section of the U-shaped plate; The U-section thermal imaging data is the thermal imaging data of the U-section plate section.

[0034] A6, passes the U-shaped sheet through the preset O-shaped forming module to form an O-shaped seamed pipe; The O-shaped forming module is an equipment module used to extrude pre-bent sheet materials into an O-shape; The O-shaped seamed pipe is a stainless steel plate formed by the U-forming module. At this time, the stainless steel plate has been formed into a tube shape with a gap left.

[0035] A7, obtains the O-shaped cross-section thermal imaging data and the seam thermal imaging data of the O-shaped seamed pipe; The O-type cross-section thermal imaging data is the thermal imaging data of the O-type seamed pipe cross-section; The seam thermal imaging data is a top-view thermal imaging data of an O-type seamed pipe, including an image of the contact area.

[0036] A8, based on the pre-bent edge section thermal imaging data, U-section section thermal imaging data, O-section section thermal imaging data, and seam thermal imaging data of the stainless steel plate, generates an estimated weld quality score for the O-shaped seam pipe using a pre-trained weld quality prediction model; The weld quality prediction model is a recognition model pre-trained based on historical data collection, and is used to evaluate the quality of O-type seam pipes after welding. The thermal imaging data of the pre-bent edge section, U-section section, and O-section section can provide the deformation characteristics of the stainless steel plate during forming, and further indirectly provide the stress characteristics of the forming process. Combined with the thermal imaging data of the joint, the weld quality prediction model can be used to predict the weld quality after welding.

[0037] Through the above steps, the intelligent processing method based on seamless stainless steel welded steel pipes can preheat the plate and collect thermal imaging data at each node of the forming process. In combination with the pre-trained weld quality prediction model, the weld quality can be predicted before welding, realizing the transition from passive quality detection to active quality pre-control, effectively reducing the scrap rate and rework cost, and improving production efficiency and product quality stability.

[0038] Furthermore, the intelligent processing method based on seamless stainless steel welded steel pipe further includes the following steps: A9: If the estimated weld quality score corresponding to the O-shaped seam pipe is greater than or equal to the preset weld score qualification threshold, the seam of the O-shaped seam pipe is welded by the preset welding module to form a stainless steel welded pipe; The weld score acceptance threshold is a pre-set reference value used to determine the size of the estimated weld quality score of O-type seam pipes; Stainless steel welded pipe is an O-shaped seamed pipe after the joints are welded inside and outside; When the estimated weld quality score of the O-type seam pipe is greater than or equal to the weld score qualification threshold, it can be determined that its forming process is qualified and the next step of welding can be carried out.

[0039] A10, if the estimated weld quality score corresponding to the O-seam pipe is less than the weld score passing threshold, a worker manually evaluates and corrects the seam of the O-seam pipe. After the correction, the seam of the O-seam pipe is welded using the welding module to form a stainless steel welded pipe. When the estimated weld quality score of the O-type seam pipe is less than the weld score qualification threshold, it can be determined that its forming process is unqualified and needs to be corrected or repaired before the next welding step can be carried out.

[0040] A11, obtaining weld image data of stainless steel welded pipe; The weld image data is image data of the joint after welding, including the weld image of the inner wall of the welded pipe and the weld image of the outer wall of the welded pipe.

[0041] A12, obtaining ultrasonic testing data of the weld seam of the stainless steel welded pipe through a preset ultrasonic testing module; The ultrasonic testing module is an ultrasonic device used to detect welds and evaluate weld quality based on ultrasonic data; The weld ultrasonic testing data is the ultrasonic reflection data at the weld of the stainless steel welded pipe.

[0042] A13, evaluates the welds of stainless steel welded pipes based on weld image data and weld ultrasonic testing data and generates corresponding actual weld quality scores; The actual weld quality score is the quality score of the weld obtained after evaluating the weld image data and the weld ultrasonic testing data. It can be calculated through a certain data processing algorithm or generated through a specific pre-trained recognition model.

[0043] A14: If the actual weld quality score is greater than or equal to the weld score qualification threshold, the corresponding stainless steel welded pipe is defined as a qualified stainless steel welded pipe; Qualified stainless steel welded pipes are stainless steel welded pipes whose actual weld quality score is greater than or equal to the weld score qualified threshold.

[0044] Through the above steps, the intelligent processing method based on seamless stainless steel welded steel pipes can realize intelligent prediction before welding and comparison verification after welding through a dual quality control system of estimated scoring and actual scoring. When the estimated quality does not meet the standard, manual intervention and correction can be carried out in time to avoid unqualified products from entering the welding process. At the same time, combined with internal and external evaluations of weld image data and ultrasonic detection data, the reliability and consistency of the final product quality are ensured, thereby improving production efficiency and product qualification rate.

[0045] Furthermore, the intelligent processing method based on seamless stainless steel welded steel pipe further includes the following steps: B1, obtaining outer wall image data and inner wall image data of a qualified stainless steel welded pipe; The outer wall image data is image data of the outer wall of a qualified stainless steel welded pipe; The inner wall image data is image data of the inner wall of a qualified stainless steel welded pipe.

[0046] B2, determining the corresponding weld pipe to-be-cleaned area data using a preset cleaning area recognition algorithm based on the weld seam image data, the outer wall image data, and the inner wall image data; The cleaning area recognition algorithm is a pre-set algorithm used to identify the areas that need to be cleaned on the inner wall and outer surface of qualified stainless steel welded pipes based on images; The welded pipe area to be cleaned data is the relative positioning data of the areas that need to be cleaned on the inner wall and the outer surface of the qualified stainless steel welded pipe.

[0047] B3, clean the inner wall, outer wall and weld of the qualified stainless steel welded pipe in a preset welded pipe cleaning method according to the data of the welded pipe area to be cleaned; Weld pipe cleaning methods are pre-set cleaning methods, which can be various cleaning methods for cleaning weld pipes and welds. For example, mechanical cleaning can be done by grinding the welds and the inner and outer pipe surfaces with tools such as grinders and wire brushes; chemical cleaning can be done by using pickling fluid to remove oxide scale and welding slag; and gas cutting cleaning may be required for thick welding slag. According to the data of the welded pipe area to be cleaned, the cleaning can be carried out more accurately, avoiding full coverage cleaning.

[0048] B4, obtaining weld seam image data, outer wall image data, and inner wall image data of the qualified stainless steel welded pipe after cleaning and defining them as weld seam cleaning image data, outer wall cleaning image data, and inner wall cleaning image data, respectively; The weld cleaning image data is weld image data after cleaning operation; The outer wall cleaning image data is the outer wall image data after the cleaning operation; The inner wall cleaning image data is inner wall image data after the cleaning operation.

[0049] B5, calculating the corresponding weld pipe cleanliness based on the weld cleaning image data, the outer wall cleaning image data, and the inner wall cleaning image data; The cleanliness of the welded pipe is the quantitative value of the cleanliness of the inner and outer walls of the cleaned stainless steel welded pipe; The cleanliness of welded pipes can be calculated using specific image processing algorithms. For example, image data of standard parts with no dirt, clean inner and outer walls, and clean welds can be collected and image comparison can be performed to calculate the cleanliness of welded pipes.

[0050] B6, if the cleanliness of the welded pipe is greater than or equal to a preset cleanliness threshold, then the qualified stainless steel welded pipe after the cleaning operation is expanded according to the preset expansion size through the preset pipe expansion module to form a stainless steel seamless welded pipe of a predetermined size; The cleanliness threshold is a pre-set reference value used to judge the cleanliness of the welded pipe; The expansion size is a pre-set processing parameter used to determine the final expansion size of the stainless steel welded pipe; The pipe expansion module is used to expand qualified stainless steel welded pipes; The predetermined size stainless steel seamless welded pipe is the stainless steel seamless welded pipe under the expanded size; After the expansion operation, the corresponding seamless welded pipe can be obtained through existing seamless processes, such as hot tension reducing process, cold drawing process, cold rolling process and global or local heat treatment of welds.

[0051] Through the above steps, the intelligent processing method based on seamless stainless steel welded steel pipes can locate the area to be cleaned of the welded pipe through the recognition algorithm according to the image of the qualified stainless steel welded pipe, thereby realizing targeted cleaning operations and avoiding the waste of resources of traditional comprehensive cleaning. At the same time, the cleanliness is quantitatively evaluated by image data to ensure that the cleaning effect meets the standards before the pipe is expanded, and finally a seamless welded pipe of a predetermined size is formed, which improves production efficiency while reducing the cost of manual quality inspection and ensures the surface quality and dimensional accuracy of the final product.

[0052] Furthermore, the step A8 includes the following steps: A801, generating corresponding pre-bend section line feature data based on the pre-bend section thermal imaging data using a preset section image feature extraction algorithm; The cross-sectional image feature extraction algorithm is a pre-set image feature extraction algorithm; The pre-bent edge section line feature data is image feature data of the plate section corresponding to the pre-bent edge section thermal imaging data, reflecting the linear features of the pre-bent edge section.

[0053] A802, generating corresponding U-shaped cross-section line feature data based on the U-shaped cross-section thermal imaging data by using a cross-section image feature extraction algorithm; The U-shaped cross-section line feature data is feature data of the plate cross-section image corresponding to the U-shaped cross-section thermal imaging data, and reflects the linear features of the U-shaped cross-section.

[0054] A803 generates corresponding O-section line feature data based on the O-section thermal imaging data using a cross-section image feature extraction algorithm; The O-section line feature data is image feature data of the plate cross section corresponding to the O-section thermal imaging data, and reflects the linear features of the O-section.

[0055] A804, generating corresponding seam line feature data based on the seam thermal imaging data using a preset seam image feature extraction algorithm; The seam image feature extraction algorithm is a pre-set image feature extraction algorithm; The seam line feature data is the seam image feature data corresponding to the seam thermal imaging data, and reflects the linear features of the seam.

[0056] A805 generates a corresponding estimated weld quality score through a weld quality estimation model based on the pre-bent edge section line feature data, U-section section line feature data, O-section section line feature data, and seam line feature data; The pre-bent edge section line feature data, U-section line feature data, O-section line feature data and seam line feature data are combined and input into the weld quality prediction model to generate a corresponding estimated weld quality score.

[0057] Through the above steps, the intelligent processing method based on seamless stainless steel welded steel pipes can convert thermal imaging data into corresponding line feature data through cross-sectional image feature extraction algorithms and seam image feature extraction algorithms, thereby realizing accurate quantification and standardized processing of key geometric parameters in the forming process, so that the weld quality prediction model can make predictions based on more accurate and consistent input data, thereby improving the prediction accuracy and the generalization ability of the model, and providing effective data support for achieving high-precision weld quality pre-control.

[0058] Furthermore, the cross-sectional image feature extraction algorithm includes the following steps: C1, determine the effective range of grayscale value according to the molding temperature; The effective range of grayscale values ​​is the range of grayscale values ​​corresponding to the molding temperature in the thermal image; For example, the standard part can be heated to the molding temperature, and a 16-bit thermal imaging grayscale image of the standard part at the molding temperature can be obtained by photographing it. The value range of the pixel grayscale value corresponding to the standard part image in the thermal imaging grayscale image can be determined based on the thermal imaging grayscale image.

[0059] C2, generating corresponding plate cross-section thermal imaging grayscale image data according to the preset plate cross-section thermal imaging image data; The plate cross-section thermal imaging data is the plate cross-section that requires cross-section image feature extraction, such as the above-mentioned pre-bent edge cross-section thermal imaging data, U-shaped cross-section thermal imaging data, and O-shaped cross-section thermal imaging data; The plate cross-section thermal imaging grayscale image data is a grayscale image corresponding to the plate cross-section thermal imaging image data, such as a 16-bit grayscale image, which can be generated by a specific grayscale image conversion algorithm.

[0060] C3, traverse each pixel point in the grayscale image data of the thermal imaging of the plate cross section, and if the grayscale value of the pixel point is not within the valid grayscale value range, assign the grayscale value of the pixel point to a preset default value; The preset default value is a preset pixel grayscale value, for example, it can be 0; If the grayscale value of a pixel is not within the valid grayscale value range, it can be determined to be a pixel corresponding to the surrounding environment, and it can be assigned a preset default value to eliminate the image of the surrounding environment after traversal, reducing interference in subsequent image processing.

[0061] C4, after traversing each pixel point in the grayscale image data of the thermal imaging of the plate cross section, generates the plate cross section grayscale image processing data; The plate cross-section grayscale image processing data is the plate cross-section thermal imaging grayscale image data after traversing and processing each pixel point.

[0062] C5, intercepting the grayscale image data of the corresponding effective area of ​​the plate cross section according to the preset default value and the plate cross section grayscale image processing data; The plate cross-section effective area grayscale image data is the minimum area containing the plate cross-section grayscale image in the plate cross-section grayscale image processing data; According to the preset default value, the image area that does not contain the plate cross section in the plate cross section grayscale image processing data can be cropped to obtain the minimum area that contains the plate cross section grayscale image.

[0063] C6, generating corresponding plate cross-section grayscale image edge feature data based on the plate cross-section effective area grayscale image data through a preset edge feature extraction algorithm and defining it as corresponding plate cross-section line feature data; The edge feature extraction algorithm is a pre-set image algorithm used to extract edge features in an image; The plate cross-section grayscale image edge feature data is the image edge feature of the plate cross-section corresponding to the plate cross-section effective area grayscale image data; The plate cross-section line feature data is the plate cross-section grayscale image edge feature data, which reflects the shape characteristics of the plate cross-section through the edge features.

[0064] Through the above steps, the intelligent processing method based on seamless stainless steel welded steel pipes can effectively filter out noise and environmental information in the thermal imaging image by screening the effective range of grayscale values ​​based on the forming temperature, and then accurately locate the key feature areas of the plate cross section through default value processing and effective area interception. Finally, the edge feature extraction algorithm is used to generate standardized line feature data, realizing the effective conversion from raw thermal imaging data to structured feature data, providing more accurate, stable and reliable input data for subsequent weld quality estimation.

[0065] Furthermore, the seam image feature extraction algorithm includes the following steps: D1, determine the effective range of grayscale value according to the molding temperature; The effective range of grayscale values ​​is the range of grayscale values ​​corresponding to the molding temperature in the thermal imaging image.

[0066] D2, generating corresponding seam thermal imaging grayscale image data according to the seam thermal imaging image data; The seam thermal imaging grayscale image data is a grayscale image corresponding to the seam thermal imaging image data, such as a 16-bit grayscale image, which can be generated by a specific grayscale image conversion algorithm.

[0067] D3, traverse each pixel point in the seam thermal imaging grayscale image data, and if the grayscale value of the pixel point is not within the valid grayscale value range, assign the grayscale value of the pixel point to a preset default value; If the grayscale value of a pixel is not within the valid grayscale value range, it can be determined to be a pixel corresponding to the surrounding environment, and it can be assigned a preset default value to eliminate the image of the surrounding environment after traversal, reducing interference in subsequent image processing.

[0068] D4, after traversing each pixel point in the seam thermal imaging grayscale image data, generating seam grayscale image processing data; The seam grayscale image processing data is the grayscale image data of the seam thermal imaging grayscale image data after traversal processing.

[0069] D5, determining corresponding seam grayscale image valid area positioning data according to the preset default value and the seam grayscale image processing data; The effective area positioning data of the seam grayscale image is the minimum area in the seam grayscale image processing data that contains the top view grayscale image of the O-type pipe. Usually, the seam image is located in the middle of the effective area positioning data of the seam grayscale image. According to the preset default value, the image area that does not contain the O-type pipe in the seam grayscale image processing data can be cropped to obtain the minimum area containing the O-type pipe grayscale image.

[0070] D6, obtaining an image of a corresponding area in the seam thermal imaging grayscale image data according to the seam grayscale image effective area positioning data and defining the image as the seam effective area grayscale image data; The seam effective area grayscale image data is a grayscale image corresponding to the seam grayscale image effective area positioning data in the initial seam thermal imaging grayscale image data.

[0071] D7, traverse each pixel point in the grayscale image data of the effective seam area, and if the grayscale value of the pixel point is within the effective grayscale value range, assign the grayscale value of the pixel point to a preset default value; By traversing and assigning preset default values ​​to the pixels whose grayscale values ​​are within the effective grayscale value range, the pipe itself can be removed from the grayscale image data of the effective seam area while retaining the corresponding image of the seam.

[0072] D8, after traversing each pixel point in the grayscale image data of the effective seam area, generating grayscale image processing data of the effective seam area; The seam effective area grayscale image processing data is an image corresponding to the seam of the O-type pipe in the seam effective area grayscale image data.

[0073] D9, generating corresponding edge feature data of the grayscale image of the effective seam area through an edge feature extraction algorithm based on the grayscale image processing data of the effective seam area and defining the edge feature data as the corresponding seam line feature data; The edge feature data of the grayscale image of the effective seam area is the edge feature data corresponding to the grayscale image processing data of the effective seam area; The seam line feature data is the edge feature data of the grayscale image of the effective seam area, and the shape characteristics of the seam are reflected through the edge features.

[0074] Through the above steps, the intelligent processing method based on seamless stainless steel welded steel pipe can use a dual grayscale value processing mechanism to first filter out invalid environmental pixels and locate the effective area of ​​the seam, and then reversely process the pixels in the effective area to highlight the seam boundary features, thereby realizing the recognition and feature enhancement of the seam area, and improving the contrast of the seam area in the thermal imaging image. The seam line feature data finally generated can better reflect the geometric characteristics of the seam, provide higher-quality input data for the weld quality prediction model, and improve the accuracy and reliability of welding quality prediction.

[0075] Furthermore, the cleaning area identification algorithm includes the following steps: E1, generating corresponding weld grayscale image data, outer wall grayscale image data and inner wall grayscale image data respectively according to the weld image data, outer wall image data and inner wall image data; The weld grayscale image data is a grayscale image corresponding to the weld image data; The outer wall grayscale image data is a grayscale image corresponding to the outer wall image data; The inner wall grayscale image data is a grayscale image corresponding to the inner wall image data.

[0076] E2, respectively calculating the corresponding average grayscale value of the weld grayscale image, the average grayscale value of the outer wall grayscale image, and the average grayscale value of the inner wall grayscale image according to the grayscale value of each pixel point in the weld grayscale image data, the outer wall grayscale image data, and the inner wall grayscale image data; The average grayscale value of the weld grayscale image is the average grayscale value of all pixels in the weld grayscale image data; The average grayscale value of the outer wall grayscale image is the average grayscale value of all pixels in the outer wall grayscale image data; The average grayscale value of the inner wall grayscale image is the average grayscale value of all pixels in the inner wall grayscale image data.

[0077] E3, segmenting the weld grayscale image data, the outer wall grayscale image data, and the inner wall grayscale image data according to a preset segmentation grid to obtain corresponding weld grid grayscale image data, outer wall grid grayscale image data, and inner wall grid grayscale image data; The segmentation grid is a pre-set grid used to divide the image into multiple rectangles of equal size; The weld seam mesh grayscale image data is image data of a single mesh formed by segmenting the weld seam grayscale image data; The outer wall grid grayscale image data is image data of a single grid formed by dividing the outer wall grayscale image data; The inner wall grid grayscale image data is image data in a single grid formed by dividing the inner wall grayscale image data.

[0078] E4, calculating the standard deviation of the grayscale values ​​of all pixels in each weld mesh grayscale image data according to the average grayscale value of the weld grayscale image and defining it as the standard deviation of the weld mesh grayscale image; The standard deviation of the weld mesh grayscale image is the standard deviation of the grayscale values ​​of all pixels in the weld mesh grayscale image data calculated with the average grayscale value of the weld grayscale image as the average value.

[0079] E5, if the standard deviation of the weld mesh grayscale image is greater than the preset mesh grayscale standard deviation threshold, then define the area of ​​the corresponding weld mesh grayscale image data in the segmented mesh as the area to be cleaned; The grid grayscale standard deviation threshold is a pre-set reference value used to determine the size of the standard deviation of the weld grid grayscale image; The area to be cleaned is the surface area corresponding to the inner wall and outer wall of the O-type pipe that needs to be cleaned later.

[0080] E6, calculating the standard deviation of the grayscale values ​​of all pixels in each outer wall grid grayscale image data according to the average grayscale value of the outer wall grayscale image and defining it as the standard deviation of the outer wall grid grayscale image; The standard deviation of the outer wall grid grayscale image is the standard deviation of the grayscale values ​​of all pixels in the outer wall grid grayscale image data calculated with the average grayscale value of the outer wall grayscale image as the average value.

[0081] E7, if the standard deviation of the outer wall grid grayscale image is greater than the grid grayscale standard deviation threshold, then define the area of ​​the corresponding outer wall grid grayscale image data in the segmented grid as the area to be cleaned; Compare and determine the areas on the outer wall surface that need to be cleaned.

[0082] E8, calculating the standard deviation of the grayscale values ​​of all pixels in each inner wall grid grayscale image data according to the average grayscale value of the inner wall grayscale image and defining it as the standard deviation of the inner wall grid grayscale image; The standard deviation of the inner wall grid grayscale image is the standard deviation of the grayscale values ​​of all pixels in the inner wall grid grayscale image data calculated with the average grayscale value of the inner wall grayscale image as the average value.

[0083] E9, if the standard deviation of the inner wall grid grayscale image is greater than the grid grayscale standard deviation threshold, then define the area of ​​the corresponding inner wall grid grayscale image data in the segmented grid as the area to be cleaned; Compare and determine the areas on the inner wall surface that need to be cleaned.

[0084] E10, combines all areas to be cleaned to generate the data of the welded pipe area to be cleaned.

[0085] All the areas to be cleaned are combined and deduplicated to obtain the data of the areas to be cleaned of the welded pipe.

[0086] Through the above steps, the intelligent processing method based on seamless stainless steel welded steel pipes can identify abnormal areas on the weld, outer wall and inner wall surfaces through grid segmentation and standard deviation statistical analysis, and automatically locate the specific locations of pollution, oxide scale, spatter, etc. that need to be cleaned based on the change characteristics of the gray value standard deviation, thereby achieving precise positioning and cleaning. This not only reduces the time and cost of cleaning operations, but also avoids damage to the pipe surface caused by excessive cleaning, improves cleaning efficiency and consistency of cleaning quality, and provides guarantees for subsequent pipe expansion and final product quality.

[0087] refer to Figure 2 The present invention also provides an intelligent processing system based on seamless stainless steel welded steel pipes, comprising: Forming module 10; Welding module 20; Clean up module 30; Pipe expansion module 40; Processing control module 50; The forming module 10, the welding module 20, the cleaning module 30 and the pipe expansion module 40 are respectively data-connected to the processing control module 50; The forming module 10 includes a preheating module 11, a plate pre-bending module 12, a U-forming module 13, an O-shaped forming module 14, and a thermal imaging module 15. The preheating module 11, the plate pre-bending module 12, the U-forming module 13, the O-shaped forming module 14, and the thermal imaging module 15 are respectively data-connected to the processing control module 50. The forming module 10 is mainly used to form the stainless steel plate into an O-shaped pipe for subsequent welding into a stainless steel welded pipe; The preheating module 11 is mainly used to heat the stainless steel plate to a preset temperature to improve its plasticity; The plate pre-bending module 12 is mainly used to bend the edge of the stainless steel plate to facilitate the subsequent forming steps; The U-forming module 13 is mainly used to extrude the pre-bent sheet into a U-shaped sheet to facilitate subsequent forming steps; The O-shaped forming module 14 is mainly used to extrude the pre-bent plate into an O-shaped pipe for welding into a welded pipe in the subsequent step; The thermal imaging module 15 is mainly used to obtain thermal imaging data of each step in the processing of the stainless steel pipe.

[0088] The welding module 20 is mainly used to weld O-shaped pipes into stainless steel welded pipes.

[0089] The cleaning module 30 is mainly used to clean the inner and outer surfaces and welds of the stainless steel welded pipe.

[0090] The pipe expansion module 40 is mainly used to expand the size of the cleaned stainless steel welded pipe.

[0091] The processing control module 50 is mainly used to receive data from various modules and blocks for processing and to control the various modules and blocks.

[0092] The intelligent processing system based on seamless stainless steel welded steel pipe further includes a quality control strategy, including the following steps: F1, heating the preset stainless steel plate to a preset forming temperature through the preheating module 11; F2, passing the heated stainless steel sheet through the sheet pre-bending module 12 to form a pre-bent sheet; F3, obtaining thermal imaging data of a pre-bent edge cross section of the pre-bent edge plate through the thermal imaging module 15; F4, passing the pre-bent sheet through the U-forming module 13 to form a U-shaped sheet; F5, obtaining thermal imaging data of the U-shaped cross section of the U-shaped plate through the thermal imaging module 15; F6, passing the U-shaped plate through the O-shaped forming module 14 to form an O-shaped seamed pipe; F7, obtaining thermal imaging data of the O-shaped cross section and the seam of the O-shaped seamed pipe through the thermal imaging module 15; F8, through the processing control module 50, based on the pre-bent edge section thermal imaging data, U-section thermal imaging data, O-section thermal imaging data and seam thermal imaging data of the stainless steel plate, an estimated weld quality score corresponding to the O-shaped seam pipe is generated using a preset weld quality estimation model.

[0093] Through the above technical solution, the intelligent processing system based on seamless stainless steel welded steel pipes can preheat the plates and collect thermal imaging data at each node of the forming process. Combined with the pre-trained weld quality prediction model, it can predict the weld quality before welding, realizing the transition from passive quality detection to active quality control, effectively reducing the scrap rate and rework costs, and improving production efficiency and product quality stability.

[0094] The above are all preferred embodiments of the present application and are not intended to limit the scope of protection of this application. Unless otherwise specified, any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features. In other words, unless otherwise specified, each feature is merely an example of a series of equivalent or similar features.

Claims

1. An intelligent processing method based on seamless stainless steel welded steel pipe, characterized in that: The following steps are involved: A1, heating the preset stainless steel plate to the preset forming temperature; A2, passing the heated stainless steel sheet through a preset sheet pre-bending module to form a pre-bent sheet; A3, obtaining thermal imaging data of the pre-bent edge cross section of the pre-bent edge plate; A4, passing the pre-bent sheet through a preset U-forming module to form a U-shaped sheet; A5, obtains the thermal imaging data of the U-shaped cross section of the U-shaped plate; A6, passes the U-shaped sheet through the preset O-shaped forming module to form an O-shaped seamed pipe; A7, obtains the O-shaped cross-section thermal imaging data and the seam thermal imaging data of the O-shaped seamed pipe; A8, based on the pre-bent edge section thermal imaging data, U-section thermal imaging data, O-section thermal imaging data, and seam thermal imaging data of stainless steel plates, uses a pre-trained weld quality prediction model to generate an estimated weld quality score for O-shaped seamed pipes.

2. The intelligent processing method based on seamless stainless steel welded steel pipe according to claim 1 is characterized in that: Further comprising the steps of: A9: If the estimated weld quality score corresponding to the O-shaped seam pipe is greater than or equal to the preset weld score qualification threshold, the seam of the O-shaped seam pipe is welded by the preset welding module to form a stainless steel welded pipe; A10, if the estimated weld quality score corresponding to the O-seam pipe is less than the weld score passing threshold, a worker manually evaluates and corrects the seam of the O-seam pipe. After the correction, the seam of the O-seam pipe is welded using the welding module to form a stainless steel welded pipe. A11, obtaining weld image data of stainless steel welded pipe; A12, obtaining ultrasonic testing data of the weld seam of the stainless steel welded pipe through a preset ultrasonic testing module; A13, evaluates the welds of stainless steel welded pipes based on weld image data and weld ultrasonic testing data and generates corresponding actual weld quality scores; A14: If the actual weld quality score is greater than or equal to the weld score qualification threshold, the corresponding stainless steel welded pipe is defined as a qualified stainless steel welded pipe.

3. The intelligent processing method based on seamless stainless steel welded steel pipe according to claim 2 is characterized in that: Further comprising the steps of: B1, obtaining outer wall image data and inner wall image data of a qualified stainless steel welded pipe; B2, determining the corresponding weld pipe to-be-cleaned area data using a preset cleaning area recognition algorithm based on the weld seam image data, the outer wall image data, and the inner wall image data; B3, clean the inner wall, outer wall and weld of the qualified stainless steel welded pipe in a preset welded pipe cleaning method according to the data of the welded pipe area to be cleaned; B4, obtaining weld seam image data, outer wall image data, and inner wall image data of the qualified stainless steel welded pipe after cleaning and defining them as weld seam cleaning image data, outer wall cleaning image data, and inner wall cleaning image data, respectively; B5, calculating the corresponding weld pipe cleanliness based on the weld cleaning image data, the outer wall cleaning image data, and the inner wall cleaning image data; B6, if the cleanliness of the welded pipe is greater than or equal to a preset cleanliness threshold, the qualified stainless steel welded pipe after the cleaning operation is expanded according to a preset expansion size through a preset pipe expansion module to form a stainless steel seamless welded pipe of a predetermined size.

4. The intelligent processing method based on seamless stainless steel welded steel pipe according to claim 3 is characterized in that: Step A8 includes the following steps: A801, generating corresponding pre-bend section line feature data based on the pre-bend section thermal imaging data using a preset section image feature extraction algorithm; A802, generating corresponding U-shaped cross-section line feature data based on the U-shaped cross-section thermal imaging data by using a cross-section image feature extraction algorithm; A803 generates corresponding O-section line feature data based on the O-section thermal imaging data using a cross-section image feature extraction algorithm; A804, generating corresponding seam line feature data based on the seam thermal imaging data using a preset seam image feature extraction algorithm; A805 generates a corresponding estimated weld quality score through a weld quality estimation model based on the pre-bent edge section line feature data, U-section line feature data, O-section line feature data and seam line feature data.

5. The intelligent processing method based on seamless stainless steel welded steel pipe according to claim 4 is characterized in that: The cross-sectional image feature extraction algorithm comprises the following steps: C1, determine the effective range of grayscale value according to the molding temperature; C2, generating corresponding plate cross-section thermal imaging grayscale image data according to the preset plate cross-section thermal imaging image data; C3, traverse each pixel point in the grayscale image data of the thermal imaging of the plate cross section, and if the grayscale value of the pixel point is not within the valid grayscale value range, assign the grayscale value of the pixel point to a preset default value; C4, after traversing each pixel point in the grayscale image data of the thermal imaging of the plate cross section, generates the plate cross section grayscale image processing data; C5, intercepting the grayscale image data of the corresponding effective area of ​​the plate cross section according to the preset default value and the plate cross section grayscale image processing data; C6, generating corresponding plate cross-section grayscale image edge feature data according to the plate cross-section effective area grayscale image data through a preset edge feature extraction algorithm and defining it as corresponding plate cross-section line feature data.

6. The intelligent processing method based on seamless stainless steel welded steel pipe according to claim 4 is characterized in that: The seam image feature extraction algorithm comprises the following steps: D1, determine the effective range of grayscale value according to the molding temperature; D2, generating corresponding seam thermal imaging grayscale image data according to the seam thermal imaging image data; D3, traverse each pixel point in the seam thermal imaging grayscale image data, and if the grayscale value of the pixel point is not within the valid grayscale value range, assign the grayscale value of the pixel point to a preset default value; D4, after traversing each pixel point in the seam thermal imaging grayscale image data, generating seam grayscale image processing data; D5, determining corresponding seam grayscale image valid area positioning data according to the preset default value and the seam grayscale image processing data; D6, obtaining an image of a corresponding area in the seam thermal imaging grayscale image data according to the seam grayscale image effective area positioning data and defining the image as the seam effective area grayscale image data; D7, traverse each pixel point in the grayscale image data of the effective seam area, and if the grayscale value of the pixel point is within the effective grayscale value range, assign the grayscale value of the pixel point to a preset default value; D8, after traversing each pixel point in the grayscale image data of the effective seam area, generating grayscale image processing data of the effective seam area; D9, generating corresponding edge feature data of the grayscale image of the effective seam area through an edge feature extraction algorithm according to the grayscale image processing data of the effective seam area and defining the edge feature data as the corresponding seam line feature data.

7. The intelligent processing method based on seamless stainless steel welded steel pipe according to claim 6, characterized in that: The cleaning area identification algorithm The following steps are involved: E1, generating corresponding weld grayscale image data, outer wall grayscale image data and inner wall grayscale image data respectively according to the weld image data, outer wall image data and inner wall image data; E2, respectively calculating the corresponding average grayscale value of the weld grayscale image, the average grayscale value of the outer wall grayscale image, and the average grayscale value of the inner wall grayscale image according to the grayscale value of each pixel point in the weld grayscale image data, the outer wall grayscale image data, and the inner wall grayscale image data; E3, segmenting the weld grayscale image data, the outer wall grayscale image data, and the inner wall grayscale image data according to a preset segmentation grid to obtain corresponding weld grid grayscale image data, outer wall grid grayscale image data, and inner wall grid grayscale image data; E4, calculating the standard deviation of the grayscale values ​​of all pixels in each weld mesh grayscale image data according to the average grayscale value of the weld grayscale image and defining it as the standard deviation of the weld mesh grayscale image; E5, if the standard deviation of the weld mesh grayscale image is greater than the preset mesh grayscale standard deviation threshold, then define the area of ​​the corresponding weld mesh grayscale image data in the segmented mesh as the area to be cleaned; E6, calculating the standard deviation of the grayscale values ​​of all pixels in each outer wall grid grayscale image data according to the average grayscale value of the outer wall grayscale image and defining it as the standard deviation of the outer wall grid grayscale image; E7, if the standard deviation of the outer wall grid grayscale image is greater than the grid grayscale standard deviation threshold, then define the area of ​​the corresponding outer wall grid grayscale image data in the segmented grid as the area to be cleaned; E8, calculating the standard deviation of the grayscale values ​​of all pixels in each inner wall grid grayscale image data according to the average grayscale value of the inner wall grayscale image and defining it as the standard deviation of the inner wall grid grayscale image; E9, if the standard deviation of the inner wall grid grayscale image is greater than the grid grayscale standard deviation threshold, then define the area of ​​the corresponding inner wall grid grayscale image data in the segmented grid as the area to be cleaned; E10, combines all areas to be cleaned to generate the data of the welded pipe area to be cleaned.

8. An intelligent processing system based on seamless stainless steel welded pipes, characterized in that: include: Forming module; Welding module; Clean up the module; Pipe expansion module; Processing control module; Wherein, the forming module, the welding module, the cleaning module and the pipe expansion module are respectively data-connected to the processing control module; The forming module includes a preheating module, a plate pre-bending module, a U-forming module, an O-shaped forming module and a thermal imaging module, and the preheating module, the plate pre-bending module, the U-forming module, the O-shaped forming module and the thermal imaging module are respectively data-connected to the processing control module; The intelligent processing system based on seamless stainless steel welded steel pipe further includes a quality control strategy, including the following steps: F1, heating the preset stainless steel plate to a preset forming temperature through the preheating module; F2, passing the heated stainless steel sheet through the sheet pre-bending module to form a pre-bent sheet; F3, obtaining thermal imaging data of a pre-bent edge cross section of the pre-bent edge plate through the thermal imaging module; F4, passing the pre-bent sheet through the U-forming module to form a U-shaped sheet; F5, obtaining thermal imaging data of the U-shaped cross section of the U-shaped plate through the thermal imaging module; F6, passing the U-shaped sheet through the O-shaped forming module to form an O-shaped seamed pipe; F7, obtaining thermal imaging data of the O-shaped cross section and the seam of the O-shaped seamed pipe by the thermal imaging module; F8, through the processing control module, generates an estimated weld quality score corresponding to the O-shaped seam pipe according to the pre-bent edge section thermal imaging data, U-shaped section thermal imaging data, O-shaped section thermal imaging data and seam thermal imaging data of the stainless steel plate using a preset weld quality estimation model.

Images