Automatic welding method and computer readable storage medium

Abstract

According to the automatic welding method and the computer readable storage medium, the space pose of a metal piece is obtained in real time through a three-dimensional vision system, and workpiece placement deviation is eliminated based on rigid body transformation registration; a welding path sequence with speed constraint is generated by calculating the geometrical relationship of the joint face, and a complex welding seam track is optimized through non-uniform rational spline interpolation; in the welding process, the pitch angle and yaw angle matching path curvature of a welding gun is dynamically adjusted, and the welding speed is controlled in combination with a heat input threshold value; arc spectrum sensing is used for avoiding an interference area, and welding gun pose deviation is corrected in real time synchronously through visual feedback of a molten pool; and finally, multi-process synchronous execution and closed-loop quality control are achieved through the industrial programmable logic controller, and the precision and reliability of precision component welding in the fields of aerospace, new energy automobiles and the like are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of welding, and more particularly to an automated welding method and a computer-readable storage medium. Background Technology

[0002] In industrial manufacturing, the welding quality of metal components directly affects the structural strength and service life of products. Traditional welding relies on manual experience to position workpieces, which suffers from insufficient positioning accuracy and reliance on pre-programmed templates for path planning. Especially for irregular curved surfaces or complex assemblies, manual teaching struggles to adapt to positional deviations, easily leading to defects such as weld misalignment and uneven penetration. While existing automated systems have incorporated basic vision positioning, they lack high-precision 3D reconstruction and dynamic path correction capabilities, and their stability is insufficient when facing reflective metal surfaces or environmental interference, necessitating urgent improvement. Summary of the Invention

[0003] This invention proposes an automated welding method, comprising: S1. Obtain the spatial pose of the first and second metal parts in the base coordinate system using a 3D vision system; S2. Based on the spatial pose, calculate the geometric relationship of the joint surfaces of the two metal parts and generate the welding path trajectory; S3. Control the welding execution unit to complete the welding operation along the welding path trajectory.

[0004] Furthermore, the spatial pose acquisition specifically includes: reconstructing a three-dimensional point cloud on the workpiece surface using a line laser scanner, performing rigid body transformation registration between the point cloud data and a preset model, and outputting pose parameters including translation vectors and rotation matrices.

[0005] Furthermore, the three-dimensional point cloud reconstruction specifically includes: projecting a structured laser mesh onto the workpiece surface, synchronously acquiring distorted fringe images using a binocular industrial camera, and calculating a surface point cloud model with an absolute phase field.

[0006] Furthermore, the solution of geometric relationships specifically includes: constructing a spatial pose transformation matrix of the contact surface between the first metal part and the second metal part, calculating the three-dimensional coordinates of the welding start point and the end point in the robot base coordinate system, and generating a spatiotemporal path sequence with velocity constraints.

[0007] Furthermore, the calculation of the welding start and end point coordinates specifically includes: extracting the curvature feature point set of the weld edge of the first metal part, mapping and matching the topological corresponding points of the joint surface of the second metal part, and generating a continuous spatial path vector through non-uniform rational spline interpolation.

[0008] Furthermore, the driving motion of the welding torch specifically includes: adjusting the pitch and yaw angles of the welding torch tool center point in real time according to the path curvature, matching the welding speed based on the material heat input threshold, and avoiding interference areas using arc spectrum sensing.

[0009] Furthermore, the specific implementation of the welding operation includes: driving the end of the welding torch to move along the path sequence and performing the welding operation, and dynamically matching the current and voltage pulse waveforms, and correcting the welding torch posture deviation in real time through visual feedback of the molten pool.

[0010] The present invention also proposes a computer-readable storage medium storing executable program code, which, when run by an industrial PLC, implements the automated welding method.

[0011] This invention can capture the spatial pose of metal parts in real time based on a 3D vision system and eliminate workpiece placement deviations through rigid body transformation registration; it can generate a welding path sequence with velocity constraints by combining the geometric relationship of the joint surface, and support non-uniform rational spline interpolation to optimize complex weld trajectories; it can dynamically adjust the pitch and yaw angles of the welding torch to match the path curvature, and control the welding speed through a heat input threshold; it can avoid interference areas in real time based on arc spectrum sensing, and correct pose deviations by combining visual feedback from the molten pool; it can synchronously execute multiple processes with the help of an industrial programmable logic controller to achieve high-precision welding and quality closed-loop control, which is especially suitable for precision manufacturing scenarios such as aerospace and new energy vehicles. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of an automated welding method proposed in this invention. Detailed Implementation

[0013] refer to Figure 1 This invention proposes an automated welding method, comprising: S1. Obtain the spatial pose of the first and second metal parts in the base coordinate system through a three-dimensional vision system.

[0014] S2. Based on the spatial pose, calculate the geometric relationship of the joint surfaces of the two metal parts and generate the welding path trajectory.

[0015] S3. Control the welding execution unit to complete the welding operation along the welding path trajectory.

[0016] Specifically, the 3D vision system is an optical sensing device used to capture the 3D shape of an object; the base coordinate system is the basic reference coordinate system for robot operation; the spatial pose is the position and orientation information of the object in space; the welding execution unit is the mechanical device that performs the welding operation; the welding path trajectory is the motion path planning of the welding tool; and the fusion operation is the process of fusing metals through a heat source.

[0017] Specifically, a 3D vision system scans the first and second metal parts placed on the worktable to capture their 3D position and orientation data. Based on this data, the geometric relationship between the contact surfaces of the two metal parts is calculated, including angle and distance parameters, to generate a continuous motion path for the weld. This path takes into account the spatial coordinates and orientation of the welding start and end points. Subsequently, the welding execution unit is precisely controlled to move the welding torch along this path while applying energy to complete the metal fusion. The entire process ensures weld quality and consistency, avoids human error, and visual feedback can be used to adjust path deviations in real time, improving welding accuracy and efficiency. This method is suitable for complex industrial applications such as automotive manufacturing or aerospace component assembly.

[0018] Furthermore, the spatial pose acquisition specifically includes: reconstructing a three-dimensional point cloud on the workpiece surface using a line laser scanner, performing rigid body transformation registration between the point cloud data and a preset model, and outputting pose parameters including translation vectors and rotation matrices.

[0019] Among them, the line laser scanner is a detection device that emits structured laser lines and receives reflected light; the three-dimensional point cloud is a dense set of three-dimensional coordinates of points on the surface of an object; the preset model is a pre-stored digital three-dimensional model of the workpiece; the rigid body transformation registration is the spatial alignment process between the point cloud and the model; the translation vector is a description of the direction and magnitude of the spatial displacement of the object; and the rotation matrix is ​​a mathematical representation of the rotational posture of the object.

[0020] Specifically, a line laser scanner projects a laser grid onto the surface of a metal part, and a sensor captures the reflected light to generate surface point cloud data. This point cloud contains the three-dimensional coordinate information of each point on the workpiece. It is then registered with a preset model, and the optimal transformation is calculated to make the point cloud coincide with the model. The output includes the pose result containing translation and rotation parameters. This process solves the problem of workpiece placement deviation and ensures the accuracy of subsequent path calculation. Registration can be achieved through an iterative nearest point algorithm, which optimizes the efficiency of pose parameter output and is suitable for welding positioning in high-precision industrial environments.

[0021] Furthermore, the three-dimensional point cloud reconstruction specifically includes: projecting a structured laser mesh onto the workpiece surface, synchronously acquiring distorted fringe images using a binocular industrial camera, and calculating a surface point cloud model with an absolute phase field.

[0022] Specifically, the structured laser grid is a projection pattern composed of regular laser lines; the binocular industrial camera is two industrial-grade camera devices used in pairs; the distorted fringe image is an image of the laser grid after deformation on the object surface; the absolute phase field is a three-dimensional shape parameter calculated based on the image phase; and the surface point cloud model is a reconstructed three-dimensional point set on the object surface.

[0023] Specifically, a laser grid pattern is projected onto the workpiece surface, and a binocular industrial camera captures images of deformed stripes from different angles. The absolute phase value is calculated using a phase extraction algorithm, which corresponds to the surface height information, thereby reconstructing a high-precision 3D point cloud model. The point cloud accuracy is affected by the camera resolution and algorithm optimization, ensuring the reliability of subsequent registration. This step reduces environmental interference errors, is suitable for metal surfaces with varying reflectivity, and improves the robustness of the overall welding system.

[0024] Furthermore, the solution of geometric relationships specifically includes: constructing a spatial pose transformation matrix of the contact surface between the first metal part and the second metal part, calculating the three-dimensional coordinates of the welding start point and the end point in the robot base coordinate system, and generating a spatiotemporal path sequence with velocity constraints.

[0025] Among them, the spatial pose transformation matrix is ​​a mathematical tool for describing the position and attitude transformation between two coordinate systems; the welding start point is the starting position of the weld; the welding end point is the ending position of the weld; the robot base coordinate system is the basic three-dimensional reference system for robot operation; and the spatiotemporal path sequence is a continuous motion trajectory sequence containing position and velocity information.

[0026] Specifically, based on the acquired spatial pose data, a transformation matrix of the metal part joint surface is constructed, and the three-dimensional coordinates of the weld start and end points are accurately calculated. These coordinates are defined in the base coordinate system. When generating the welding path sequence, a speed constraint is added to ensure that the welding torch moves at a uniform or variable speed to avoid welding defects. The sequence planning takes into account the tool center and path curvature, optimizes the trajectory smoothness, and makes the welding process efficient and stable, which is suitable for complex curved surface welding scenarios.

[0027] Furthermore, the calculation of the welding start and end point coordinates specifically includes: extracting the curvature feature point set of the weld edge of the first metal part, mapping and matching the topological corresponding points of the joint surface of the second metal part, and generating a continuous spatial path vector through non-uniform rational spline interpolation.

[0028] Specifically, the curvature feature point set is the coordinate set of the high curvature position at the weld edge; the topological corresponding point is the point of geometric matching between the surfaces of two metal parts; the non-uniform rational spline interpolation is the mathematical method for generating smooth curves; and the continuous spatial path vector is the description of the uninterrupted three-dimensional motion direction.

[0029] Specifically, high curvature feature points are identified from the weld edge of the first metal part. These points reflect key turning points and are mapped and aligned with corresponding points on the second metal part to ensure geometric consistency of the joint surface. Then, a non-uniform rational spline interpolation algorithm is applied to connect the point set to generate a smooth path vector. This path is continuous and without abrupt changes, adapting to irregular weld shapes, improving welding quality and aesthetics, and is suitable for automated assembly of irregularly shaped metal parts.

[0030] Furthermore, the driving motion of the welding torch specifically includes: adjusting the pitch and yaw angles of the welding torch tool center point in real time according to the path curvature, matching the welding speed based on the material heat input threshold, and avoiding interference areas using arc spectrum sensing.

[0031] Specifically, the tool center point is the reference point for the end-operation of the welding torch; the pitch angle is the tilt angle of the welding torch in the vertical plane; the yaw angle is the turning angle of the welding torch in the horizontal plane; the material heat input threshold is the maximum allowable thermal energy limit for the welding process; the arc spectral sensing is a sensing technology that detects the environment through arc light analysis; and the interference region is the obstacle area in the welding path.

[0032] Specifically, during the welding process, the path curvature change is monitored in real time, and the pitch and yaw angles of the welding torch are dynamically adjusted to maintain the optimal posture of the tool center point. At the same time, the welding speed is matched according to the material type to ensure that the heat input does not exceed the threshold to prevent deformation. The arc spectrum sensor analyzes the light signal of the molten pool to identify the interference zone and corrects the path in time to avoid obstacles. This closed-loop control improves welding safety and accuracy and adapts to the welding needs of multiple materials.

[0033] Furthermore, the specific implementation of the welding operation includes: driving the end of the welding torch to move along the path sequence and performing the welding operation, and dynamically matching the current and voltage pulse waveforms, and correcting the welding torch posture deviation in real time through visual feedback of the molten pool.

[0034] Specifically, the current and voltage pulse waveforms refer to the time control mode of welding energy output; the visual feedback of the molten pool refers to the image analysis system that monitors the molten metal pool through a camera; and the welding torch posture deviation refers to the offset between the actual position of the welding torch and the planned path.

[0035] Specifically, the welding torch moves along the generated path sequence and applies energy to weld the metal. At the same time, the current, voltage, and pulse parameters are dynamically adjusted according to the welding conditions to optimize heat input control. The weld pool vision system captures the weld pool image in real time, analyzes the shape and temperature distribution, detects positional deviations, and provides feedback to correct the welding torch position and angle, ensuring that the weld is uniform and defect-free, making it suitable for high-quality welding applications.

[0036] Furthermore, the present invention also proposes a computer-readable storage medium storing executable program code, which, when run by an industrial PLC, implements the automated welding method.

[0037] Specifically, computer-readable storage media refers to digital storage carriers that store program data; executable program code refers to a sequence of instructions that can be executed by a processor; and industrial programmable logic controllers refer to the core processing units for industrial automation control.

[0038] Specifically, after the code on the storage medium is loaded by the industrial programmable logic controller, it executes all steps of the automated welding method, including visually acquiring pose, solving geometric relationships, generating path trajectories, and controlling the welding operation. The controller coordinates the sensors and actuators to achieve closed-loop control, ensuring the method operates efficiently. It is suitable for large-scale production line deployment and improves the level of welding automation.

[0039] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An automated welding method, characterized in that, include: S1. Obtain the spatial pose of the first and second metal parts in the base coordinate system using a 3D vision system; S2. Based on the spatial pose, calculate the geometric relationship of the joint surfaces of the two metal parts and generate the welding path trajectory; S3. Control the welding execution unit to complete the welding operation along the welding path trajectory.

2. The automated welding method as described in claim 1, characterized in that, The spatial pose acquisition specifically includes: reconstructing a three-dimensional point cloud on the workpiece surface using a line laser scanner, performing rigid body transformation registration on the point cloud data and a preset model, and outputting pose parameters including translation vectors and rotation matrices.

3. The automated welding method as described in claim 2, characterized in that, The three-dimensional point cloud reconstruction specifically includes: projecting a structured laser mesh onto the workpiece surface, synchronously acquiring distorted fringe images using a binocular industrial camera, and calculating the surface point cloud model of the absolute phase field.

4. The automated welding method as described in claim 1, characterized in that, The solution of geometric relationships specifically includes: constructing a spatial pose transformation matrix of the contact surface between the first metal part and the second metal part, calculating the three-dimensional coordinates of the welding start point and end point in the robot base coordinate system, and generating a spatiotemporal path sequence with velocity constraints.

5. An automated welding method as described in claim 4, characterized in that, The calculation of the welding start and end point coordinates specifically includes: extracting the curvature feature point set of the weld edge of the first metal part, mapping and matching the corresponding points of the joint surface of the second metal part, and generating a continuous spatial path vector through non-uniform rational spline interpolation.

6. The automated welding method as described in claim 5, characterized in that, The specific methods for driving the welding torch motion include: adjusting the pitch and yaw angles of the welding torch tool center point in real time according to the path curvature, matching the welding speed based on the material heat input threshold, and avoiding interference areas using arc spectrum sensing.

7. An automated welding method as described in claim 1, characterized in that, The specific steps of performing the welding operation include: driving the end of the welding torch to move along the path sequence and performing the welding operation, dynamically matching the current and voltage pulse waveforms, and correcting the welding torch posture deviation in real time through visual feedback of the molten pool.

8. A computer-readable storage medium, characterized in that, It stores executable program code, which, when run by an industrial PLC, implements an automated welding method according to any one of claims 1-7.

Images