Intelligent welding method for tee

Abstract

The present invention relates to an intelligent welding method, which moves away from singular points by changing the diameter of the robot during first walking trajectory planning, effectively avoiding jitter when the multi-axis robot changes its posture due to being trapped in a wrist singular point during welding, and matching the current coordinates of the robot with the posture of trajectory planning to know the position of the currently scanned workpiece. According to the weld characteristics at this position, an adaptation algorithm is called to obtain the precise position of the tee workpiece, and the walking trajectory of the robot is calculated based on this position, so that each point on the weld is welded through the welding point of the welding gun, thereby improving the welding accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of welding methods, in particular to an intelligent welding method for a tee. Background Art

[0002] With the increasing prevalence of robots, manual welding has gradually replaced traditional methods in industry. However, with the increasing precision required for welding workpieces, traditional welding methods are no longer able to meet the welding needs of various workpiece types. Existing welding methods initially utilized a robotic arm to drive the welding gun, placing the workpiece on a positioner that coordinated with the robotic arm's movements to complete the welding process. However, this method, due to its complex structure and large variations in the weld pool, proved ineffective for certain high-precision welding applications and has been eliminated.

[0003] Chinese patent CN113523501B discloses an intelligent welding method in which the welding gun is fixed and the workpiece is fixed on a robotic arm. Computer control is used, and a visual sensor cooperates with a laser to identify the weld seam. The robotic arm moves along the identified welding trajectory and ensures that the workpiece corresponds to the posture of the welding gun in various postures during welding, thereby completing the welding.

[0004] However, the following disadvantages exist in this welding method in which the welding gun is fixed and the workpiece is fixed on the robot arm, and the robot arm is allowed to move in various postures to ensure that the workpiece corresponds to the posture of the welding gun during welding: First, when welding a tee workpiece, in order to reduce system complexity and facilitate the calculation of the relative coordinate system between the workpiece and the robot, when the tee workpiece is installed on the robot, the plane of the tee branch pipe is parallel to the plane of the robot flange by default. If the tee is of equal diameter, the maximum angle between the two pipes is 180 degrees, and a flat angle weld will be formed at the lowest position of the main pipe (such as Figure 13 As shown in the figure, the rest of the parts are normal curved welds. When the robot is installed in this way and drives the equal-diameter tee workpiece to move so that all points on the weld pass through the same point, when it moves to the flat angle, it will cause a multi-axis robot, such as a 6-axis robot, to have two axes (axis 4 and axis 6) on the same line (i.e., the angle of axis 5 is 0), and fall into a wrist singularity point. This will cause the robot to vibrate greatly when changing its posture, greatly affecting the welding accuracy. Summary of the Invention

[0005] In order to solve the above problems, the present invention provides a three-way intelligent welding method that can effectively avoid the jitter caused by the multi-axis robot falling into the wrist singularity point during welding, thereby improving the welding accuracy.

[0006] In order to achieve the above-mentioned object, the present invention designs a three-way intelligent welding method, which includes a three-way workpiece, a robot, a welding gun, a camera and a laser transmitter, and is characterized in that the welding is carried out according to the following steps:

[0007] S1. Calculate the robot's first trajectory based on the known parameters of the tee workpiece and the position of the tee workpiece relative to the robot. If the diameters of the main and branch pipes of the tee workpiece are known to be equal, modify the diameter parameters of the main and branch pipes of the tee workpiece when calculating the robot's first trajectory. Modify the diameter parameters of the main and branch pipes of the tee workpiece so that the diameter parameters of the main and branch pipes differ by N, and then calculate the robot's first trajectory.

[0008] S2. The welding gun, camera, and laser emitter are fixed to opposite sides of the robot, ensuring that all points on the weld seam of the tee are illuminated by the laser emitted by the laser emitter when the robot moves the tee along the first trajectory.

[0009] S3. When the robot moves the tee workpiece according to the first walking trajectory, the robot's current position matches the current scanning angle and calls the adaptation algorithm based on the weld characteristics of the tee workpiece at the scanning position;

[0010] S4. The camera captures an image of the intersection of the laser beam and the weld seam on the tee workpiece. The camera processes the captured image information to determine the precise weld seam location on the tee workpiece. Based on this location, the robot calculates a more accurate and smooth second trajectory.

[0011] S5. The robot moves along the second walking trajectory, driving the tee workpiece to move, so that all points on the weld seam of the tee workpiece pass through the welding points of the welding gun one by one to complete the welding.

[0012] A further solution is that in step S1, the tee workpiece is fixed to the end of the robot arm, and the tee workpiece branch plane is parallel to the robot flange plane; the first walking trajectory of the tee workpiece robot is obtained by the parametric equation of the intersection line or the calibration method.

[0013] A further solution is that in step S3, if the scanning position is a curved corner weld, the curved corner imaging algorithm is called; if it is an equal-diameter tee workpiece, the plane corner recognition algorithm is called. The plane corner recognition algorithm processes the image acquired by the camera according to the following steps:

[0014] a. The camera captures image A and creates a Gaussian pyramid twice on it. Image A is scaled down by a factor of 4 to obtain image B.

[0015] b. Normalize the grayscale of image B row by row, and then perform full Figure 2 The value is converted to obtain the C graph;

[0016] c. Perform morphological analysis on Image C, retaining only the binary regions with thin, straight features similar to laser lines, to obtain Image D.

[0017] d. Traverse image A row by row to obtain the coordinate point (Ax, Ay) in image A. Check the grayscale value of the corresponding coordinate point (Ax / 4, Ay / 4) on image D. If the grayscale value is 255, sum the grayscale values ​​in the coordinate interval (Ax±5, Ay±5) in image A to obtain S(Ax', Ay') at the current coordinate. Take the coordinate with the largest value in a row, Smax(Ax', Ay'), to obtain image E.

[0018] e. Find a laser line in Figure E and perform a straight line fit on the found laser line. Then, use the least squares method to fit the found laser line into a straight line L1 to obtain Figure F.

[0019] f. Dilate graph C, invert it and swap the X and Y axes to obtain graph C'. Then swap the X and Y axes of graph B to obtain B'. In graph B', perform row-by-row normalization using graph C' as the operator, normalizing only the non-zero parts of C' to obtain graph G.

[0020] g. Perform mean filtering on the G graph to obtain the H graph;

[0021] h. Set the binarization threshold Ta, assign grayscale values ​​greater than the set threshold Ta to 255, assign the remaining grayscale values ​​to 0, and swap the X-axis and Y-axis to obtain the I image;

[0022] i. Delete the pixels in graph I whose connected area is less than the threshold Tb, and obtain graph J.

[0023] j. Use the Hough variation line detection algorithm to detect the straight line features L2 and L3 in the J graph to obtain the K graph;

[0024] k. Calculate the intersection point P1 of the straight lines L2 and L3, then calculate the intersection point P2 of L1 and L2, and the intersection point P3 of L1 and L3. From the image features, we can know that the horizontal coordinates of the intersection points P1 and P3 are always greater than the corner intersection point P1. Multiply P2 by 4 to get the final point P, as shown in Figure L.

[0025] A further solution is that N>8 in step S1.

[0026] A further solution is that the threshold value Ta in step h is the grayscale value of the H image minus the grayscale value of the G image.

[0027] A further solution is that the threshold Tb in step i is 50.

[0028] The intelligent welding method designed in the present invention obtains the image of the intersection of the weld and the laser through the camera to obtain the specific positions of the curved weld and the flat weld on the tee workpiece in the robot posture trajectory. When the current coordinates of the robot coincide with the position of the flat angle weld in the posture trajectory, the diameter parameters of the main pipe and branch pipe of the tee workpiece in the posture trajectory are modified so that in the calculation of the robot posture trajectory, the angle between the main pipe and the branch pipe in the tee workpiece no longer has a 180-degree angle. This can effectively avoid the jitter of the multi-axis robot when it falls into the wrist singularity point during welding, thereby improving the welding accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 FIG. A of Example 1;

[0030] Figure 2 is Figure B of Example 1;

[0031] Figure 3 FIG. C of Example 1;

[0032] Figure 4 is Figure D of Example 1;

[0033] Figure 5 is Figure E of Example 1;

[0034] Figure 6 FIG. F of Example 1;

[0035] Figure 7 is the G diagram of Example 1;

[0036] Figure 8 is the H diagram of Example 1;

[0037] Figure 9 FIG1 is FIG1 of Example 1;

[0038] Figure 10 FIG1 is J of Example 1;

[0039] Figure 11 is the K diagram of Example 1;

[0040] Figure 12 is the L diagram of Example 1;

[0041] Figure 13 It is a flat angle weld formed between the main pipe and branch pipe of a medium diameter tee workpiece in the existing technology.

[0042] Figure 14 This is the final fitting effect diagram in Example 1. DETAILED DESCRIPTION

[0043] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0044] Example 1

[0045] The tee intelligent welding method described in this embodiment includes a tee workpiece, a robot, a welding gun, a camera, and a laser transmitter, and is characterized in that the welding is performed according to the following steps:

[0046] S1. Calculate the robot's first trajectory based on the known parameters of the tee workpiece and the position of the tee workpiece relative to the robot. If the diameters of the main and branch pipes of the tee workpiece are known to be equal, modify the diameter parameters of the main and branch pipes of the tee workpiece when calculating the robot's first trajectory. Modify the diameter parameters of the main and branch pipes of the tee workpiece so that the diameter parameters of the main and branch pipes differ by N, and then calculate the robot's first trajectory.

[0047] S2. The welding gun, camera, and laser emitter are fixed to opposite sides of the robot, ensuring that all points on the weld seam of the tee are illuminated by the laser emitted by the laser emitter when the robot moves the tee along the first trajectory.

[0048] S3. When the robot moves the tee workpiece according to the first walking trajectory, the robot's current posture matches the current scanning angle, and the adaptation algorithm is called based on the weld characteristics of the tee workpiece at the scanning position.

[0049] S4. The camera captures an image of the intersection of the laser light irradiated by the laser emitter on the weld of the tee workpiece and the weld. After processing the acquired image information, the precise weld position of the tee workpiece is obtained. Based on this position, the robot calculates a more accurate and smooth second walking trajectory.

[0050] S5. The robot moves along the second walking trajectory, driving the tee workpiece to move, so that all points on the weld seam of the tee workpiece pass through the welding points of the welding gun one by one to complete the welding.

[0051] A further solution is that in step S1, the tee workpiece is fixed to the end of the robot arm, and the tee workpiece branch plane is parallel to the robot flange plane; the first walking trajectory of the tee workpiece robot is obtained by the parametric equation of the intersection line or the calibration method.

[0052] In order to accurately and effectively obtain the position of the plane angle weld on the tee workpiece, in step S3, if the scanning position is a curved angle weld, the curved angle image algorithm is called. If it is an equal-diameter tee workpiece, the plane angle recognition algorithm is called. The plane angle recognition algorithm processes the image obtained by the camera according to the following steps:

[0053] a. The camera captures image A and performs two Gaussian pyramid operations on it. Image A is reduced by a factor of 4 to obtain image B.

[0054] b. Normalize the grayscale of image B row by row, and then perform full Figure 2 The value is converted to obtain the C graph;

[0055] c. Perform morphological analysis on Image C, retaining only the binary regions with thin, straight features similar to laser lines, to obtain Image D.

[0056] d. Traverse image A row by row to obtain the coordinate point (Ax, Ay) in image A. Check the grayscale value of the corresponding coordinate point (Ax / 4, Ay / 4) on image D. If the grayscale value is 255, sum the grayscale values ​​in the coordinate interval (Ax±5, Ay±5) in image A to obtain S(Ax', Ay') at the current coordinate. Take the coordinate with the largest value in a row, Smax(Ax', Ay'), to obtain image E.

[0057] e. Find a laser line in Figure E and perform a straight line fit on the found laser line. Then, use the least squares method to fit the found laser line into a straight line L1 to obtain Figure F.

[0058] f. Dilate graph C, invert it and swap the X and Y axes to obtain graph C'. Then swap the X and Y axes of graph B to obtain B'. In graph B', perform row-by-row normalization using graph C' as the operator, normalizing only the non-zero parts of C' to obtain graph G.

[0059] g. Perform mean filtering on the G graph to obtain the H graph;

[0060] h. Set the binarization threshold Ta, assign grayscale values ​​greater than the set threshold Ta to 255, assign the remaining grayscale values ​​to 0, and swap the X-axis and Y-axis to obtain the I image;

[0061] i. Delete the pixels in graph I whose connected area is less than the threshold Tb, and obtain graph J.

[0062] j. Use the Hough variation line detection algorithm to detect the straight line features L2 and L3 in the J graph to obtain the K graph;

[0063] k. Calculate the intersection point P1 of the straight lines L2 and L3, then calculate the intersection point P2 of L1 and L2, and the intersection point P3 of L1 and L3. From the image features, we can know that the horizontal coordinates of the intersection points P1 and P3 are always greater than the corner intersection point P1. Multiply P2 by 4 to get the final point P, as shown in Figure L.

[0064] A further solution is that N>8 in step S1.

[0065] A further solution is that the threshold value Ta in step h is the grayscale value of the H image minus the grayscale value of the G image.

[0066] A further solution is that the threshold value Tb in step i is set to 50. Of course, the threshold value Tb can also be set to other values ​​according to the relationship between the image brightness and the weld bead, and is not specifically limited here.

[0067] In this embodiment, when the image acquired by the camera is processed to obtain the robot posture trajectory, based on the filtering step of trajectory data processing, the points in the posture trajectory where the weld and the laser intersect with a changing speed that is too fast will be filtered out, and a fitting algorithm can be used to fill in these filtered out points. Specifically, when the tee workpiece is an unequal diameter tee, the points near the blank position are used as sample data, and the three-axis discrete least squares method is used, and the basis function uses a sine or cosine curve to fit the compensation point; if it is an equal diameter tee workpiece, within the range of plus or minus 45 degrees of the plane angle of the equal diameter tee workpiece, two parts of points with the angle as the boundary are divided, and a plane fitting is performed first to map these points to the plane, and then a two-dimensional straight line fitting is performed separately, and then the intersection point of the two straight lines is calculated, and the plane is inversely mapped back to the three-dimensional coordinate system to obtain the angle position, and then the blank points near the angle are interpolated according to the angle position and the sample points, so as to meet the welding requirements with the fitting accuracy, and effectively improve the posture of the robot when it moves according to the posture trajectory to be smoother and more accurate. Figure 14 The filtering steps of trajectory data processing mentioned in this embodiment are well known to those skilled in the art and will not be described in detail here.

[0068] The intelligent welding method provided by the embodiment of the present invention obtains the specific positions of the curved weld and the flat weld on the tee workpiece in the robot posture trajectory by acquiring the image of the intersection of the weld and the laser through the camera. When the current coordinates of the robot coincide with the position of the flat angle weld in the posture trajectory, the diameter parameters of the main pipe and branch pipe of the tee workpiece in the posture trajectory are modified so that in the calculation of the robot posture trajectory, the angle between the main pipe and the branch pipe in the tee workpiece no longer has a 180-degree angle. This can effectively avoid the jitter caused by the multi-axis robot falling into the wrist singularity during welding and causing the robot to change its posture, thereby improving the welding accuracy of the tee intelligent welding method.

[0069] In the description of the present invention, it should be noted that the terms "vertical", "up", "down", "horizontal", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, they cannot be understood as limiting the present invention.

[0070] In the description of the present invention, it should also be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; they may refer to mechanical connections or electrical connections; they may refer to direct connections or indirect connections through an intermediate medium; and they may refer to internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0071] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A three-way intelligent welding method, comprising a three-way workpiece, a robot, a welding gun, a camera and a laser transmitter, characterized in that: When welding, follow these steps: S1. Calculate the robot's first trajectory based on the known parameters of the tee workpiece and the position of the tee workpiece relative to the robot. If the diameters of the main and branch pipes of the tee workpiece are known to be equal, modify the diameter parameters of the main and branch pipes of the tee workpiece when calculating the robot's first trajectory. Modify the diameter parameters of the main and branch pipes of the tee workpiece so that the diameter parameters of the main and branch pipes differ by N, and then calculate the robot's first trajectory. S2. The welding gun, camera, and laser emitter are fixed to opposite sides of the robot, ensuring that all points on the weld seam of the tee are illuminated by the laser emitted by the laser emitter when the robot moves the tee along the first trajectory. S3. When the robot moves the tee workpiece according to the first walking trajectory, the robot's current position matches the current scanning angle and calls the adaptation algorithm based on the weld characteristics of the tee workpiece at the scanning position; S4. The camera captures an image of the intersection of the laser beam and the weld seam on the tee workpiece. The camera processes the captured image information to determine the precise weld seam location on the tee workpiece. Based on this location, the robot calculates a more accurate and smooth second trajectory. S5. The robot moves along the second walking trajectory, driving the tee workpiece to move, so that all points on the weld seam of the tee workpiece pass through the welding points of the welding gun one by one to complete the welding.

2. The three-way intelligent welding method according to claim 1 is characterized in that In step S1, the tee workpiece is fixed to the end of the robot arm, and the branch pipe plane of the tee workpiece is parallel to the robot flange plane; the first walking trajectory of the tee workpiece robot is obtained by the parametric equation of the intersection line or the calibration method.

3. The three-way intelligent welding method according to claim 2 is characterized in that In step S3, if the scanning position is a curved corner weld, the curved corner imaging algorithm is called. If it is an equal-diameter tee workpiece, the plane corner recognition algorithm is called. The plane corner recognition algorithm processes the image acquired by the camera according to the following steps: a. The camera captures image A and performs two Gaussian pyramid operations on it. Image A is reduced by a factor of 4 to obtain image B. b. Normalize the grayscale of image B row by row, and then binarize the entire image to obtain image C; c. Perform morphological analysis on Image C, retaining only the binary regions with thin, straight features similar to laser lines, to obtain Image D. d. Traverse image A row by row to obtain the coordinate point (Ax, Ay) in image A. Check the grayscale value of the corresponding coordinate point (Ax / 4, Ay / 4) on image D. If the grayscale value is 255, sum the grayscale values ​​in the coordinate interval (Ax±5, Ay±5) in image A to obtain S(Ax', Ay') at the current coordinate. Take the coordinate with the largest value in a row, Smax(Ax', Ay'), to obtain image E. e. Find a laser line in Figure E and perform a straight line fit on the found laser line. Then, use the least squares method to fit the found laser line into a straight line L1 to obtain Figure F. f. Dilate graph C, invert it and swap the X and Y axes to obtain graph C'. Then swap the X and Y axes of graph B to obtain B'. In graph B', perform row-by-row normalization using graph C' as the operator, normalizing only the non-zero parts of C' to obtain graph G. g. Perform mean filtering on the G graph to obtain the H graph; h. Set the binarization threshold Ta, assign grayscale values ​​greater than the set threshold Ta to 255, assign the remaining grayscale values ​​to 0, and swap the X-axis and Y-axis to obtain the I image; i. Delete the pixels in graph I whose connected area is less than the threshold Tb, and obtain graph J. j. Use the Hough variation line detection algorithm to detect the straight line features L2 and L3 in the J graph to obtain the K graph; k. Calculate the intersection point P1 of the straight lines L2 and L3, then calculate the intersection point P2 of L1 and L2, and the intersection point P3 of L1 and L3. From the image features, we can know that the horizontal coordinates of the intersection points P1 and P3 are always greater than the corner intersection point P1. Multiply P2 by 4 to get the final point P, as shown in Figure L.

4. The tee intelligent welding method according to claim 1 is characterized in that: N>8 described in step S1.

5. The tee intelligent welding method according to claim 3 is characterized in that: The threshold value Ta in step h is the grayscale value of the H image minus the grayscale value of the G image.

6. The tee intelligent welding method according to claim 3 is characterized in that: The threshold value Tb mentioned in step i=50.

Images