Intelligent full welding method and welding system for a weld of a workpiece

Abstract

The present application relates to the field of workpiece welding in manufacturing, and particularly to an intelligent full welding method for a workpiece welding seam in welding technology. The present application comprises a gantry truss, a vertical plate, a positioner driving motor, a positioner fixed footing block, a turnover frame, a Y-axis linear motion machine, a Y-axis moving slider, an X-Y linear motion mechanism fixed rib plate, an X-axis linear motion mechanism, an X-axis moving slider, a six-degree-of-freedom welding mechanical arm, a positioner turnover turntable, a mechanical arm front end gripper, a welding torch, a mechanical vision scanner, and an electromagnet block. The present application can recognize the workpiece placement position using electromagnetic induction, achieve accurate positioning and fixation of the workpiece, reduce power consumption, judge whether the welding state is qualified through multiple visual scanning, and perform repair welding. The welding mechanical arm moves according to the recognition point to complete the second face welding work, ensuring the welding seam precision and improving the workpiece welding quality.

Description

Technical Field

[0001] This invention relates to the field of workpiece welding in manufacturing, specifically to an intelligent full-welding method and welding system for workpiece welds. Background Technology

[0002] Welding is a crucial part of machining. Traditional welding requires different fixtures to hold different workpieces in place, often necessitating multiple fixtures for various workpiece models, increasing production costs. Furthermore, for large-volume production of workpieces with high weld quality requirements and complex weld point locations, full welding is extremely difficult, demanding higher levels of manual skill. Manual welding suffers from incomplete welds, poor weld quality, and inconsistent weld conditions, failing to meet customer requirements for contour accuracy and mounting flatness. Summary of the Invention

[0003] The main technical problem solved by this invention is to provide an intelligent full-welding method and welding system for workpiece welds, which solves problems such as incomplete welds, poor weld quality, and inconsistent weld conditions. By using visual scanning to determine whether the workpiece is fully welded and qualified, the invention enables welding on both sides of the product, ensuring the quality requirements of the welded products and comprehensively and effectively improving production efficiency.

[0004] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A smart full-welding method for workpiece welds includes the following steps: 1) Construct welding equipment, which includes a flipping frame with both ends connected to a flipping turntable, and a six-degree-of-freedom welding robot arm located above the flipping frame. The six-degree-of-freedom welding robot arm is mounted on the X-axis moving slider of the X-axis linear motion mechanism. Both ends of the X-axis linear motion mechanism are connected to the Y-axis moving slider through the fixed ribs of the XY-axis linear motion mechanism, thereby connecting to the Y-axis linear motion mechanism. The two Y-axis linear motion mechanisms are respectively set on both sides of the flipping frame through a gantry truss. A machine vision scanner and a welding torch are installed at the welding end of the six-degree-of-freedom welding robot arm. 2) On the flip frame, multiple electromagnet blocks with built-in electromagnetic sensors are evenly distributed in a matrix; 3) Place the workpiece on the flipping frame. After the electromagnetic sensor at the corresponding position senses the workpiece, control the power supply to charge the corresponding electromagnet block to attract the workpiece. 4) Control the X-axis linear motion mechanism and Y-axis linear motion mechanism to move the six-degree-of-freedom welding robot arm above the workpiece. The machine vision scanner scans the workpiece to obtain an image of the workpiece. Based on the extracted image and the information sent by the corresponding electromagnetic sensor, a three-dimensional model of the workpiece is established. The three-dimensional model is input into the radial basis neural network model to generate a set of weld seam vector segments and the corresponding welding spatial trajectory. Based on the set of weld seam vector segments and the corresponding welding spatial trajectory, the six-degree-of-freedom welding robot arm is controlled to move, and the welding of the workpiece seam is completed by the welding torch. 5) After welding, the weld is scanned again using a mechanical vision scanner to obtain the image of the weld. The cross-sectional shape of the weld is obtained according to the laser triangulation method. The cross-sectional shape of the weld is compared with the complete cross-sectional shape stored in the database. If the similarity between the actual cross-sectional shape and the stored cross-sectional shape is more than 80%, it is considered qualified. 6) If the actual cross-sectional shape is less than 80% similar to the stored cross-sectional shape, repeat steps 4 and 5 until the condition is met; if the condition is met, start the flipping mechanism, flip the flipping frame, repeat step 4, and complete the welding of the other side of the workpiece. After welding, repeat step 5. If the condition is met, complete the welding of the workpiece weld.

[0005] Furthermore, before inputting the 3D model of the workpiece into the welding neural network model to generate each weld vector segment group and the corresponding welding spatial trajectory, the process also includes training the welding neural network model.

[0006] Furthermore, after the weld seams on both sides of the workpiece are repaired, the control power supply stops supplying power to the electromagnet block.

[0007] Furthermore, after the weld seams on both sides of the workpiece are repaired, the X-axis linear motion mechanism and the Y-axis linear motion mechanism are reset.

[0008] Furthermore, the qualified weld data is saved and the database is updated.

[0009] Furthermore, after the repair welding, mechanical recognition is performed again to evaluate the weld quality of the workpiece, and the cross-sectional shape types that are qualified and need repair welding are recorded. Through deep learning, the accuracy of mechanical vision in judging the weld after welding is continuously improved through iterative optimization.

[0010] A welding system based on the intelligent full-weld method for workpiece welds as described above, comprising: The welding equipment includes a tilting frame, with both ends of the tilting frame connected to a tilting mechanism, and a six-degree-of-freedom welding robot arm located above the tilting frame. The six-degree-of-freedom welding robot arm is mounted on the X-axis moving slider of the X-axis linear motion mechanism. Both ends of the X-axis linear motion mechanism are connected to the Y-axis linear motion mechanism. The two Y-axis linear motion mechanisms are respectively set on both sides of the tilting frame through a gantry truss. The six-degree-of-freedom welding robot arm is equipped with a machine vision scanner and a welding torch. An electromagnet block fixed to the flipping frame is used to attract the workpiece; The electromagnetic sensor located inside the corresponding electromagnet block is used to sense the position of the workpiece and send the position coordinate signal of the corresponding sensor to the controller. The controller is used to build a three-dimensional model of the workpiece based on the extracted image and the coordinate position of the corresponding electromagnetic sensor. The three-dimensional model is input into the radial basis neural network model to generate a set of weld vector segments and the corresponding spatial trajectory of welding. Based on the weld, the controller sends movement commands to the X-axis linear motion mechanism and the Y-axis linear motion mechanism, and sends welding commands to the six-degree-of-freedom welding robot arm. After the weld is completed, the cross-sectional shape of the weld is obtained by laser triangulation based on the image of the weld. The cross-sectional shape of the weld is compared with the database. If the conditions are not met, a movement command is sent to the X-axis linear motion mechanism and the Y-axis linear motion mechanism, and a welding command is sent to the six-degree-of-freedom welding robot until the conditions are met. If the conditions are met, a flipping command is sent to the flipping mechanism. The X-axis linear motion mechanism is used to execute movement commands; The Y-axis linear motion mechanism is used to execute movement commands; A flipping mechanism is used to execute flipping commands; A six-degree-of-freedom welding robotic arm is used to execute welding commands.

[0011] Furthermore, the flipping frame includes a frame body, within which multiple horizontal beams and multiple vertical beams are arranged horizontally and vertically, dividing the frame body into multiple welded cavities; each electromagnet block is located in its corresponding welded cavity and is fixedly connected to its corresponding horizontal beam.

[0012] Furthermore, the flipping mechanism includes upright plates disposed on both sides of the flipping frame. Each upright plate is fixedly connected to a shaft column extending from the flipping frame via a flipping turntable. One flipping turntable is connected to the rotating shaft of a positioner drive motor mounted on the upright plate, and the other flipping turntable is connected to the upright plate via a bearing.

[0013] Furthermore, the gantry truss includes longitudinal beams arranged along the Y-axis on the left and right sides of the flipping frame, with supports at both ends of each longitudinal beam. Two Y-axis linear motion mechanisms are respectively installed on the longitudinal beams. It also includes a crossbeam connected to the two Y-axis linear motion mechanisms, and an X-axis linear motion mechanism is fixed on the crossbeam.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This invention presents an intelligent full-welding method for workpiece welds. The provided flipping frame uses electromagnet blocks to magnetically hold the workpiece, flexibly adapting to most workpieces, reducing the use and cost of fixtures. Electromagnetic induction is used to identify the workpiece's placement position, using a minimal number of electromagnet blocks to hold the workpiece, reducing power consumption. Visual scanning is employed for positioning and identification, creating a 3D model of the workpiece and calibrating the weld seam's position coordinates. Finally, the weld seam's welding status is assessed using the visual scan image, offering greater accuracy and intelligence compared to traditional infrared methods, thus improving workpiece welding quality. A 3D animation coordinate model of the workpiece, the six-degree-of-freedom welding robot arm base, and the robot arm's front gripper is established, allowing for full-welding observation at the terminal during equipment operation. The welding position transformation, and the resulting tripartite coordinate system, improved the precision of the industrial control system, enhancing weld positioning and welding accuracy. An intelligent operation process for full welding was established, from fixing the workpiece with an electromagnet to removing it. While the operation is more complex than traditional welding processes, it significantly improves the processing precision of full welding, increasing workpiece yield and welding quality. A digital twin model system of a positioner, a moving gantry truss, and a six-degree-of-freedom welding robot was also established. This industrial system can observe the entire welding process in real time, iteratively optimize the welding positioning program, improve the positioning accuracy of visual scanning, and simultaneously save qualified weld status data, enhancing the accuracy of visual scanning in identifying weld status. Attached Figure Description

[0015] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a three-dimensional schematic diagram of the mechanical structure of the intelligent full welding method for workpiece welds according to the present invention.

[0017] Figure 2 This is a system control flowchart for the intelligent full welding method for workpiece welds described in this invention.

[0018] Figure 3This is a schematic diagram of the left and right types of gantry trusses in the intelligent full welding method for workpiece welds described in this invention.

[0019] Figure 4 This is a schematic diagram of the positioner structure for the intelligent full welding method for workpiece welds described in this invention.

[0020] Figure 5 This is a schematic diagram of a six-degree-of-freedom welding robot arm for the intelligent full welding method of workpiece weld seams described in this invention.

[0021] Figure 6 This is a schematic diagram of the XY-axis linear motion mechanism structure of the intelligent full welding method for workpiece welds according to the present invention.

[0022] The attached figures are labeled as follows: 1. Gantry truss; 2. Vertical plate; 3. Positioner drive motor; 4. Positioner fixed foot block; 5. Tilting frame; 6. Y-axis linear motion mechanism; 7. Y-axis moving slider; 8. XY-axis linear motion mechanism fixed rib plate; 9. X-axis linear motion mechanism; 10. X-axis moving slider; 11. Six-degree-of-freedom welding robot arm; 12. Positioner tilting turntable; 13. Robot arm front gripper; 14. Welding torch; 15. Machine vision scanner; 16. Electromagnetic block. Detailed Implementation

[0023] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] The present invention will be further described below with reference to specific embodiments.

[0025] like Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 and Figure 6 As shown, an intelligent full-welding method for workpiece welds includes the following steps: 1) Construct welding equipment, which includes a flipping frame with both ends connected to a flipping turntable, and a six-degree-of-freedom welding robot arm located above the flipping frame. The six-degree-of-freedom welding robot arm is mounted on the X-axis moving slider of the X-axis linear motion mechanism. Both ends of the X-axis linear motion mechanism are connected to the Y-axis moving slider through the fixed ribs of the XY-axis linear motion mechanism, thereby connecting to the Y-axis linear motion mechanism. The two Y-axis linear motion mechanisms are respectively set on both sides of the flipping frame through a gantry truss. A machine vision scanner and a welding torch are installed at the welding end of the six-degree-of-freedom welding robot arm. 2) On the flip frame, multiple electromagnet blocks with built-in electromagnetic sensors are evenly distributed in a matrix; 3) Place the workpiece on the flipping frame. After the electromagnetic sensor at the corresponding position senses the workpiece, control the power supply to charge the corresponding electromagnet block to attract the workpiece. 4) Control the X-axis linear motion mechanism and Y-axis linear motion mechanism to move the six-degree-of-freedom welding robot arm above the workpiece. The machine vision scanner scans the workpiece to obtain an image of the workpiece. Based on the extracted image and the information sent by the corresponding electromagnetic sensor, a three-dimensional model of the workpiece is established. The three-dimensional model is input into the radial basis neural network model to generate a set of weld seam vector segments and the corresponding welding spatial trajectory. Based on the set of weld seam vector segments and the corresponding welding spatial trajectory, the six-degree-of-freedom welding robot arm is controlled to move, and the welding of the workpiece seam is completed by the welding torch. 5) After welding, the weld is scanned again using a mechanical vision scanner to obtain the image of the weld. The cross-sectional shape of the weld is obtained according to the laser triangulation method. The cross-sectional shape of the weld is compared with the complete cross-sectional shape stored in the database. If the similarity between the actual cross-sectional shape and the stored cross-sectional shape is more than 80%, it is considered qualified. 6) If the actual cross-sectional shape is less than 80% similar to the stored cross-sectional shape, repeat steps 4 and 5 until the condition is met; if the condition is met, start the flipping mechanism, flip the flipping frame, repeat step 4, and complete the welding of the other side of the workpiece. After welding, repeat step 5. If the condition is met, complete the welding of the workpiece weld.

[0026] The present invention provides an intelligent full-welding method for workpiece welds, comprising: constructing a welding device, the welding device including a tilting frame with both ends connected to a tilting turntable; and a six-degree-of-freedom welding robot arm located above the tilting frame, the six-degree-of-freedom welding robot arm being mounted on the X-axis sliding block of an X-axis linear motion mechanism, the two ends of which are connected to a Y-axis linear motion mechanism, the two Y-axis linear motion mechanisms being respectively located on both sides of the tilting frame via a gantry truss; a machine vision scanner and a welding torch being installed at the welding end of the six-degree-of-freedom welding robot arm; placing the workpiece to be welded on the tilting frame containing an electromagnet block, and energizing and fixing the workpiece to be welded; the machine vision scanner scanning the workpiece placed on the tilting frame... The process involves scanning the workpiece to be welded on the flipping frame and transmitting the data to the controller; welding the workpiece on one side; scanning the welded side again; flipping the frame to turn it over; a machine vision scanner scanning the flipped workpiece; welding the other side of the workpiece; scanning the welded side again; and finally, power is cut off, the electromagnet loses its magnetism and stops attracting the workpiece, allowing it to be removed. This invention uses an electromagnet to magnetically hold the workpiece, which can flexibly adapt to most workpieces, reducing the use and cost of fixtures. It also uses electromagnetic induction to identify the workpiece's placement, achieving precise positioning and fixation while reducing power consumption. Furthermore, this invention uses visual scanning to determine the welding status and allows for re-welding. The moving welding robotic arm completes the second welding operation based on the identified points, ensuring weld accuracy and improving the overall welding quality.

[0027] In this embodiment, both the X-axis linear motion mechanism and the Y-axis linear motion mechanism use linear guides manufactured by Jinhengtai Precision Machinery Co., Ltd. Linear guides from other manufacturers can also be used, but details will not be provided here.

[0028] This embodiment includes training the welding neural network model before inputting the three-dimensional model of the workpiece into the welding neural network model to generate each weld vector segment group and the corresponding welding spatial trajectory.

[0029] The captured weld seam images and their feature output vectors are used as the training set for the neural network, while the normalized and standardized training set is used as the input layer. The process of obtaining the feature output vectors involves classifying the captured weld seam images according to their type and the cross-sectional features of that type. These types include common weld seam cross-sectional shapes such as I-type, V-type, Y-type, and U-type. Feature points are the horizontal and vertical coordinates of the weld seam cross-section, and point segments are vector segments formed by two feature points with a straight-line distance of 0.1 mm within the weld seam cross-section. A line segment symmetrical to the centerline and separated by a distance five times the weld groove is used as the feature output vector for a set of weld seam point segments. This classification of the captured weld seam images and the use of their feature output vectors as the training set for the neural network is then applied. The normalized and standardized training set is used as the input layer. A radial basis function neural network model is trained, and the output layer converts the weld seam features into vector combinations. The obtained weld seam feature combinations are used to establish a weld seam coordinate system, and the centerline of the weld is extracted to create a spatial trajectory array of the weld seam lines, enabling welding torch trajectory tracking for welding.

[0030] The present invention, through the above design, can establish a model in the controller and store the complete cross-sectional shape in the database, which is convenient for comparison with the workpiece welding cross-sectional shape generated by subsequent welding scanning.

[0031] After the weld seams on both sides of the workpiece are repaired in this embodiment, the control power supply stops supplying power to the electromagnet block.

[0032] Through the above design, the present invention can disconnect the adsorption of the flipping bracket on the workpiece after the workpiece welding is completed, thereby removing the workpiece.

[0033] After the weld seams on both sides of the workpiece are repaired in this embodiment, the X-axis linear motion mechanism and the Y-axis linear motion mechanism are reset.

[0034] The present invention can automatically reset the X-axis linear motion mechanism and the Y-axis linear motion mechanism through the above design, without the need for manual labor and facilitating the next welding.

[0035] This embodiment saves the welding data of qualified welds and updates the database.

[0036] Through the above design, the present invention can iteratively update qualified weld data and retain the data information closest to the standard in the database, thus providing a better standard for subsequent welded workpieces.

[0037] like Figure 3 As shown, in this embodiment, mechanical recognition is performed again after the repair welding to evaluate the weld quality of the workpiece. At the same time, the cross-sectional shape types that are qualified and those that need to be repaired are recorded. Through deep learning, the accuracy of the mechanical vision in judging the weld after welding is continuously improved through iterative optimization.

[0038] This invention, through the above design, can improve the accuracy of mechanical vision in judging weld seams after welding, ensure weld pass rate, and thus achieve the goal of rapid welding completion. A welding system based on an intelligent full-welding method for workpiece weld seams includes... The welding equipment includes a tilting frame, with both ends of the tilting frame connected to a tilting mechanism, and a six-degree-of-freedom welding robot arm located above the tilting frame. The six-degree-of-freedom welding robot arm is mounted on the X-axis moving slider of the X-axis linear motion mechanism. Both ends of the X-axis linear motion mechanism are connected to the Y-axis linear motion mechanism. The two Y-axis linear motion mechanisms are respectively set on both sides of the tilting frame through a gantry truss. The six-degree-of-freedom welding robot arm is equipped with a machine vision scanner and a welding torch. An electromagnet block fixed to the flipping frame is used to attract the workpiece; The electromagnetic sensor located inside the corresponding electromagnet block is used to sense the position of the workpiece and send the position coordinate signal of the corresponding sensor to the controller. The controller is used to build a three-dimensional model of the workpiece based on the extracted image and the coordinate position of the corresponding electromagnetic sensor. The three-dimensional model is input into the radial basis neural network model to generate a set of weld vector segments and the corresponding spatial trajectory of welding. Based on the weld, the controller sends movement commands to the X-axis linear motion mechanism and the Y-axis linear motion mechanism, and sends welding commands to the six-degree-of-freedom welding robot arm. After the weld is completed, the cross-sectional shape of the weld is obtained by laser triangulation based on the image of the weld. The cross-sectional shape of the weld is compared with the database. If the conditions are not met, a movement command is sent to the X-axis linear motion mechanism and the Y-axis linear motion mechanism, and a welding command is sent to the six-degree-of-freedom welding robot until the conditions are met. If the conditions are met, a flipping command is sent to the flipping mechanism. The X-axis linear motion mechanism is used to execute movement commands; The Y-axis linear motion mechanism is used to execute movement commands; A flipping mechanism is used to execute flipping commands; A six-degree-of-freedom welding robotic arm is used to execute welding commands; The present invention achieves double-sided welding of the workpiece to be welded through the above design, ensuring weld accuracy and improving the welding quality of the workpiece.

[0039] like Figure 4 As shown, the flipping frame in this embodiment includes a frame body, in which multiple horizontal beams and multiple vertical beams are arranged horizontally and vertically, and the multiple horizontal beams and multiple vertical beams divide the frame body into multiple welding cavities; each electromagnet block is located in the corresponding welding cavity and is fixedly connected to the corresponding horizontal beam.

[0040] The present invention provides a carrier for the electromagnet block through the above design, which is used to fix the workpiece to be welded. The welding cavity provides a welding space for the weld seam of the workpiece, which facilitates welding.

[0041] like Figure 4 As shown, the flipping mechanism in this embodiment includes upright plates disposed on both sides of the flipping frame. Each upright plate is fixedly connected to a shaft extending from the flipping frame via a flipping turntable. One flipping turntable is connected to the shaft of a positioner drive motor mounted on the upright plate, and the other flipping turntable is connected to the upright plate via a bearing. The present invention supports the flipping frame through the above design, while providing space for the flipping frame to flip and providing power for the flipping turntable. The flipping turntable drives the flipping frame to flip, thereby realizing the function of flipping the workpiece.

[0042] like Figure 1 As shown, the gantry truss includes longitudinal beams arranged along the Y-axis on the left and right sides of the flipping frame, with supports at both ends of each longitudinal beam. Two Y-axis linear motion mechanisms are respectively installed on the longitudinal beams. It also includes a crossbeam connected to the two Y-axis linear motion mechanisms, and an X-axis linear motion mechanism is fixed on the crossbeam.

[0043] The present invention provides a motion track for the movement of a six-degree-of-freedom robotic arm through the above design, and establishes an X and Y plane translation coordinate system, which can accurately move to each point for welding.

[0044] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent full welding of workpiece welds, characterized in that: Includes the following steps: 1) Construct welding equipment. The welding equipment includes a flip frame (5), with both ends of the flip frame (5) connected to a flip turntable (12). It also includes a six-degree-of-freedom welding robot arm (11) located above the flip frame (5). The six-degree-of-freedom welding robot arm (11) is installed on the X-axis moving slider (10) of the X-axis linear motion mechanism (9). Both ends of the X-axis linear motion mechanism (9) are connected to the Y-axis moving slider (7) through the XY-axis linear motion mechanism fixing rib (8), thereby connecting to the Y-axis linear motion mechanism (6). The two Y-axis linear motion mechanisms (6) are respectively set on both sides of the flip frame (5) through the gantry truss (1). A machine vision scanner (15) and a welding torch (14) are installed at the welding end of the six-degree-of-freedom welding robot arm (11). 2) On the flip frame (5), multiple electromagnet blocks (16) with built-in electromagnetic sensors are evenly distributed in a matrix; the electromagnetic sensors located in the corresponding electromagnet blocks (16) are used to sense the placement position of the workpiece and send the position coordinate signal of the corresponding sensor to the controller. 3) Place the workpiece on the flipping frame (5). After the electromagnetic sensor at the corresponding position senses the workpiece, control the power supply to charge the corresponding electromagnet block (16) to attract the workpiece. 4) Control the X-axis linear motion mechanism (9) and Y-axis linear motion mechanism (6) to move the six-degree-of-freedom welding robot (11) above the workpiece. The machine vision scanner (15) scans the workpiece to obtain the image of the workpiece. Based on the extracted image and the information sent by the corresponding position electromagnetic sensor, a three-dimensional model of the workpiece is established. The three-dimensional model is input into the radial basis neural network model to generate the weld seam vector segment group and the corresponding welding spatial trajectory. Based on the weld seam vector segment group and the corresponding welding spatial trajectory, the six-degree-of-freedom welding robot (11) is controlled to move and the welding of the workpiece weld seam is completed by the welding gun. 5) After welding, the weld is scanned again using a mechanical vision scanner to obtain the image of the weld. The cross-sectional shape of the weld is obtained according to the laser triangulation method. The cross-sectional shape of the weld is compared with the complete cross-sectional shape stored in the database. If the similarity between the actual cross-sectional shape and the stored cross-sectional shape is more than 80%, it is considered qualified. 6) If the actual cross-sectional shape is less than 80% similar to the stored cross-sectional shape, repeat steps 4 and 5 until the condition is met; if the condition is met, start the flipping mechanism, flip the flipping frame (5), repeat step 4, and complete the welding of the other side of the workpiece. After welding, repeat step 5. If the condition is met, complete the welding of the workpiece weld. After the repair welding, mechanical recognition is performed again to evaluate the weld quality of the workpiece, and the cross-sectional shape types that are qualified and need to be repaired are recorded. Through deep learning, the accuracy of mechanical vision in judging the weld after welding is continuously improved through iterative optimization.

2. The intelligent full welding method for workpiece welds according to claim 1, characterized in that: Before inputting the 3D model of the workpiece into the welding neural network model to generate each weld vector segment group and the corresponding welding spatial trajectory, the process also includes training the welding neural network model.

3. The intelligent full-welding method for workpiece welds according to claim 1, characterized in that: After the weld seams on both sides of the workpiece are repaired, the control power supply stops supplying power to the electromagnet block (16).

4. The intelligent full-welding method for workpiece welds according to claim 3, characterized in that: After the weld seams on both sides of the workpiece are repaired, the X-axis linear motion mechanism (9) and the Y-axis linear motion mechanism (6) are reset.

5. The intelligent full-welding method for workpiece welds according to claim 1, characterized in that: Save the welding data of qualified welds and update the database.

6. A welding system based on an intelligent full-weld method for workpiece welds as described in any one of claims 1-5, characterized in that: include The welding equipment includes a flipping frame (5), which is connected to a flipping mechanism at both ends. It also includes a six-degree-of-freedom welding robot arm (11) located above the flipping frame (5). The six-degree-of-freedom welding robot arm (11) is mounted on the X-axis moving slider (10) of the X-axis linear motion mechanism (9). The X-axis linear motion mechanism (9) is connected to the Y-axis linear motion mechanism (6) at both ends. The two Y-axis linear motion mechanisms (6) are respectively set on both sides of the flipping frame (5) through the gantry truss (1). The six-degree-of-freedom welding robot arm (11) is equipped with a machine vision scanner (15) and a welding torch (14). An electromagnet block (16) fixed to the flipping frame (5) is used to attract the workpiece; The electromagnetic sensor located in the corresponding electromagnet block (16) is used to sense the position of the workpiece and send the position coordinate signal of the corresponding sensor to the controller. The controller is used to establish a three-dimensional model of the workpiece based on the extracted image and the coordinate position of the corresponding electromagnetic sensor. The three-dimensional model is input into the radial basis neural network model to generate a group of weld vector segments and the corresponding spatial trajectory of welding. Based on the weld, the controller sends movement commands to the X-axis linear motion mechanism (9) and the Y-axis linear motion mechanism (6) and sends welding commands to the six-degree-of-freedom welding robot arm (11). After the weld is welded, the cross-sectional shape of the weld is obtained by laser triangulation based on the image of the weld. The cross-sectional shape of the weld is compared with the database. If the conditions are not met, a movement command is sent to the X-axis linear motion mechanism (9) and the Y-axis linear motion mechanism (6), and a welding command is sent to the six-degree-of-freedom welding robot (11) until the conditions are met. If the conditions are met, a flipping command is sent to the flipping mechanism. X-axis linear motion mechanism (9) is used to execute movement commands; Y-axis linear motion mechanism (6) is used to execute movement commands; A flipping mechanism is used to execute flipping commands; A six-degree-of-freedom welding robot (11) is used to execute welding commands.

7. The welding system of the intelligent full welding method for workpiece welds according to claim 6, characterized in that: The flip frame (5) includes a frame body, in which multiple horizontal beams and multiple vertical beams are arranged horizontally and vertically, and the multiple horizontal beams and multiple vertical beams divide the frame body into multiple welded cavities; each electromagnet block (16) is located in the corresponding welded cavity and is fixedly connected to the corresponding horizontal beam.

8. The welding system of the intelligent full welding method for workpiece welds according to claim 6, characterized in that: The flipping mechanism includes upright plates (2) on both sides of the flipping frame (5), and each upright plate (2) is fixedly connected to the shaft column extending from the flipping frame (5) through a flipping turntable (12); one flipping turntable (12) is connected to the shaft of the positioner drive motor (3) installed on the upright plate (2), and the other flipping turntable (12) is connected to the upright plate (2) through a bearing.

9. The welding system of the intelligent full welding method for workpiece welds according to claim 8, characterized in that: The gantry truss (1) includes longitudinal beams arranged along the Y-axis on the left and right sides of the flip frame (5), with legs at both ends of each longitudinal beam. Two Y-axis linear motion mechanisms (6) are respectively installed on the longitudinal beams. It also includes a crossbeam connected to the two Y-axis linear motion mechanisms (6), and an X-axis linear motion mechanism (9) is fixed on the crossbeam.

Images