Intelligent welding method and system for curved surface of corrugated plate of container

Abstract

The invention relates to the technical field of intelligent welding, and discloses an intelligent welding method and system for a curved surface of a corrugated plate of a container, and the method comprises the following steps: carrying out full-coverage scanning on the corrugated plate of the container by utilizing three-dimensional scanning equipment to obtain point cloud data of the corrugated plate of the container, and preprocessing the point cloud data; performing curved surface reconstruction by using the preprocessed point cloud data, generating a container corrugated plate curved surface three-dimensional model, and performing curved surface smoothing processing; an intelligent welding planning algorithm is adopted to generate an optimal welding path, and the container corrugated plate is welded through the optimal welding path; and performing welding defect identification on the welding seam state data by using a welding defect identification model, and feeding back and adjusting a welding mode based on a welding defect identification result. According to the method, the optimal welding track adapting to the geometric characteristics of the corrugated plate is generated in combination with intelligent path planning, the welding seam state is sensed in real time through the multi-mode sensor in the welding process, and the welding parameters and modes are dynamically adjusted.

Description

Technical Field

[0001] The present invention relates to the manufacture of welding equipment such as arc welding machines, and in particular to the field of intelligent welding, specifically to an intelligent welding method and system for the curved surface of container corrugated plates. Background Art

[0002] With the continuous improvement of automation and intelligence in the container manufacturing industry, the welding quality and production efficiency of container corrugated sheets have gradually become one of the key factors restricting the production capacity of the entire production line. Due to the special structural design with periodic peaks and troughs, container corrugated sheets can provide higher rigidity and strength within a limited thickness, but this structural characteristic also makes its welding process face higher technical difficulty. The surface of the corrugated sheet is a complex three-dimensional curved surface structure with frequent and large changes in local curvature. During the welding process, it is necessary not only to accurately control the three-dimensional motion trajectory of the welding gun, but also to ensure the degree of matching between the welding gun posture and the normal direction of the corrugated surface to avoid weld offset or uneven heat input. In addition, the control of the welding thermal cycle is particularly important. Excessive heat input may cause weld deformation and stress concentration, while too low heat input may cause quality defects such as incomplete welding.

[0003] In actual production, a slight deviation may lead to various welding defects such as undercut, incomplete penetration, weld bead, air holes, cracks, etc., which not only reduce the mechanical properties of the weld, but also affect the overall sealing and durability of the container, and even cause safety hazards.

[0004] Patent CN105945399B, a patent application in existing research, proposes a method for automatic weld tracking using visual recognition and positioning, and an intelligent welding robot. The method comprises a lifting device mounted on a mobile carriage, a manipulator mounted on the lifting device, and a welding gun mounted on the manipulator. The welding gun is equipped with an automatic weld tracker for visual recognition and positioning, comprising a visual sensor, a controller, and an actuator connected in sequence, with the actuator connected to the welding gun. A computer control system controls the mobile carriage, lifting device, manipulator, and welding gun. This solution enables visual recognition and automatic weld tracking for mobile robots used in arc welding in the heavy machinery industry, improving the level of automation.

[0005] However, this technology is mainly aimed at flat or small-curvature weld scenes, relying on a single visual sensor for positioning. It lacks the comprehensive perception capability of the curvature, spatial posture and welding thermal state of complex three-dimensional surfaces (such as container corrugated plates). It is difficult to cope with the rapid changes in weld paths, occlusions, and real-time identification and dynamic correction of various defects, resulting in limited welding accuracy and consistency.

[0006] To address this problem, the present invention proposes an intelligent welding method and system for the curved surface of container corrugated plates to reduce welding defects and ensure the stability of welding quality. Summary of the Invention

[0007] The present invention provides an intelligent welding method and system for the curved surface of container corrugated plates. Since the surface of the corrugated plates is prone to noise during the scanning process, directly using it for path planning will result in an unstable welding trajectory. The present invention adopts a curvature-sensitive surface smoothing algorithm to suppress noise while retaining the corrugation characteristics. In addition, traditional welding path planning is difficult to strike a balance between efficiency and accuracy, especially when dealing with surface fluctuations and welding gun motion constraints, which is prone to quality fluctuations. The present invention introduces curvature, adaptive speed and posture deviation constraints, which solves the technical problem of finding a balance between welding efficiency and forming quality, as well as the problem of delayed manual quality inspection and inability to correct deviations in real time, thereby realizing closed-loop quality control of the welding process.

[0008] To achieve the above-mentioned object, the present invention provides an intelligent welding method for the curved surface of a container corrugated plate, comprising the following steps: S1: Use 3D scanning equipment to fully scan the container corrugated plate to obtain point cloud data of the container corrugated plate, and pre-process the point cloud data; S2: Surface reconstruction is performed using the pre-processed point cloud data to generate a three-dimensional model of the container corrugated plate surface, and the three-dimensional model of the container corrugated plate surface is smoothed to obtain a smoothed three-dimensional model of the container corrugated plate surface; S3: Based on the smoothed 3D model of the container corrugated plate surface, an intelligent welding planning algorithm is used to generate the optimal welding path, and the container corrugated plate is welded using the optimal welding path; S4: Use multimodal sensors to collect real-time weld status data of container corrugated plates during the welding process, use welding defect recognition models to identify welding defects based on the weld status data, and adjust the welding method based on the feedback of welding defect recognition results.

[0009] As a further improvement method of the present invention: Optionally, a 3D scanning device is used to perform a full coverage scan of the container corrugated plate to obtain point cloud data of the container corrugated plate, including: The three-dimensional scanning equipment includes a high-precision three-dimensional laser scanner and a laser rangefinder; The point cloud data consists of three-dimensional position coordinates and signal strengths of the three-dimensional position coordinates.

[0010] Optionally, preprocessing the point cloud data includes: The preprocessing includes point cloud registration, denoising and downsampling processing in sequence; The point cloud registration includes coarse registration and fine registration, wherein coarse registration is to perform preliminary alignment on the point cloud data obtained by multiple and multi-view scans so that the point cloud data of each view roughly overlap in the same coordinate system; fine registration is to further optimize the point cloud after coarse registration, remove redundant point cloud data, and construct the retained point cloud data into a point cloud data set after point cloud registration; The denoising process is as follows: after point cloud registration, the number of neighborhood points of the point cloud data and the average distance between the point cloud data and the neighborhood points are calculated, the point cloud data with the number of neighborhood points lower than a preset threshold value is marked as noise points, the point cloud data with the average distance higher than the preset threshold value is marked as noise points, the noise points are removed, and the retained point cloud data is constructed into a denoised point cloud data set; The downsampling processing method is a voxel grid method, which uses the voxel grid method to perform downsampling processing on the denoised point cloud data set, and uses the point cloud data obtained by the downsampling processing as the preprocessed point cloud data.

[0011] Optionally, the pre-processed point cloud data is used to perform surface reconstruction to generate a three-dimensional model of the container corrugated plate surface, including: Calculating the normal vectors of the preprocessed point cloud data, and constructing the normal vectors of all the preprocessed point cloud data into a gradient field, wherein the normal vectors of the preprocessed point cloud data are sampling vectors in the gradient field; Perform spatial interpolation on the gradient field so that the preprocessed point cloud data forms a continuous gradient field on the discrete voxel grid; The space where the container corrugated plate is located is divided into a multi-scale octree grid, where different scale levels correspond to different resolutions, and a discrete Poisson equation is constructed on the multi-scale octree network. The Poisson equation is iteratively solved using the finite element method to generate global continuous scalar fields at different scale levels, and the global continuous scalar field at the highest resolution is selected as the implicit scalar field obtained by the solution; The Marching Cubes algorithm is used to extract the isosurface of the implicit scalar field to obtain a closed, continuous and smooth three-dimensional surface mesh. The mesh vertices and triangular facets are generated as the three-dimensional model of the container corrugated plate surface.

[0012] Optionally, performing surface smoothing processing on the three-dimensional model of the container corrugated plate surface to obtain a smoothed three-dimensional model of the container corrugated plate surface includes: Calculate the local curvature of mesh vertices in the 3D model of container corrugated plate surface; Based on the local curvature of the mesh vertices, the position coordinates of the mesh vertices in the three-dimensional model of the container corrugated plate surface are smoothed to obtain the smoothed three-dimensional model of the container corrugated plate surface.

[0013] Optionally, based on the smoothed three-dimensional model of the container corrugated plate surface, an intelligent welding planning algorithm is used to generate an optimal welding path, including: The process of the intelligent welding planning algorithm includes: Based on the local curvature of the mesh vertices in the smoothed 3D model of the container corrugated plate surface, mesh vertices are selected and included in the candidate trajectory point set; Selecting candidate trajectory points from the candidate trajectory point set as welding starting points to generate multiple groups of local welding paths; A path optimization function of the local welding path is constructed, in which the path optimization function uses the welding control parameters of the local welding path as optimization variables and the local curvature of the welding path coordinate points in the local welding path as control parameters to generate adaptive dynamic welding control parameters of the local welding path. During the welding process of the local welding path, the welding speed, welding acceleration and welding direction are adaptively controlled. The functional form of the constructed path optimization function is: ; ; ; ; ; ; ; in, represents the path optimization function, Represents the path optimization function value of the local welding path Load; Represents the local welding path Middle The velocity smoothing constraint term of the welding path coordinate point is: Represents the local welding path The welding speed at the nth welding path coordinate point, Represents the local welding path The welding speed at the n-1th welding path coordinate point, Represents the local welding path The local curvature at the nth welding path coordinate point in , Indicates the preset maximum speed value. represents the velocity term smoothing coefficient, Indicates the modulus of the calculation vector, N represents the local welding path The number of coordinate points in the welding path; is the weight coefficient, set They are 0.2, 0.2, 0.3, 0.3 respectively; Represents the local welding path The acceleration smoothing constraint term of the nth welding path coordinate point in , Represents the local welding path The welding acceleration at the nth welding path coordinate point in , Represents the local welding path The welding acceleration at the n-1th welding path coordinate point, Indicates the preset maximum acceleration value. represents the acceleration smoothing coefficient; Represents the local welding path The welding direction constraint item of the nth welding path coordinate point, Indicates the direction vector of the welding gun when welding the nth welding path coordinate point With mesh normal The angle deviation between Local welding path The normal vector of the grid where the nth welding path coordinate point is located, represents the deviation smoothing coefficient, represents the arccosine function; Represents the local welding path The welding efficiency loss term of the nth welding path coordinate point in is, Represents the local welding path The expected speed of the nth welding path coordinate point in, Indicates selection The maximum value of the welding efficiency loss term is used to penalize unnecessary speed limit; The path optimization function of the local welding path is optimized and solved to form the optimal welding parameter sequence of the local welding path. The optimal welding parameter sequence of all local welding paths is used as the optimal welding path. The optimal welding parameter sequence is composed of the position coordinates of the welding path coordinate points in the local welding path, the optimal welding speed, the optimal welding acceleration, and the optimal welding gun orientation direction vector, among which the optimal welding speed, the optimal welding acceleration, and the optimal welding gun orientation direction vector are the solution results of the path optimization function.

[0014] Optionally, selecting a candidate trajectory point from the candidate trajectory point set as a welding starting point to generate a local welding path includes: Select a candidate trajectory point close to the edge of the container corrugated plate from the candidate trajectory point set as the welding starting point, where the candidate trajectory point is a grid vertex in the candidate trajectory point set; Taking the welding starting point as the starting point, an iterative method is used to generate a sequence of welding path coordinate points with continuous positions to form a set of local welding paths.

[0015] Optionally, a multimodal sensor is used to collect weld state data of the container corrugated plate in real time during the welding process, and a welding defect recognition model is used to perform welding defect recognition on the weld state data, including: The multimodal sensor includes a laser vision sensor, an infrared thermal imaging sensor, and an acoustic sensor. The multimodal sensor is deployed in front of the welding gun to collect real-time weld status data of the container corrugated plate during the welding process; Extract features from weld state data to obtain multimodal weld features; A welding defect recognition model built based on a deep neural network is used to receive multimodal weld features and output welding defect categories as welding defect recognition results. The welding defect categories include normal welds, undercuts, weld bumps, incomplete penetration, pores, and cracks.

[0016] Optionally, adjusting the welding method based on feedback from the welding defect identification results includes: The adjustment strategy of the welding method is: When undercut is detected, reduce the welding speed and welding current; When a weld nub is identified, reduce the wire feed speed or current and increase the welding gun movement speed appropriately; When pores are identified, maintain a stable welding gun posture during welding, use dry welding wire or clean the surface of the base material; When cracks are identified, reduce the welding speed, weld in sections or preheat the base material to reduce thermal stress; When incomplete penetration is detected, increase the welding current or heat input and slow down the welding speed.

[0017] The present invention also proposes an intelligent welding system for the curved surface of a container corrugated plate, which includes a three-dimensional reconstruction device, a path planning module, and a feedback adjustment device: The three-dimensional reconstruction device is used to use a three-dimensional scanning device to perform a full-coverage scan of the container corrugated plate to obtain point cloud data of the container corrugated plate, and pre-process the point cloud data, use the pre-processed point cloud data to perform surface reconstruction to generate a three-dimensional model of the container corrugated plate surface, and perform surface smoothing on the three-dimensional model of the container corrugated plate surface to obtain a smoothed three-dimensional model of the container corrugated plate surface; The path planning module is used to generate an optimal welding path based on the smoothed three-dimensional model of the container corrugated plate surface using an intelligent welding planning algorithm, and weld the container corrugated plate using the optimal welding path; The feedback adjustment device is used to use a multimodal sensor to collect real-time weld state data of the container corrugated plate during the welding process, use a welding defect recognition model to identify welding defects in the weld state data, and adjust the welding method based on the feedback of the welding defect recognition results; This is to realize the intelligent welding method for the curved surface of the container corrugated plate as described above.

[0018] Compared with the existing technology, the present invention proposes an intelligent welding method and system for the curved surface of container corrugated plate, which has the following beneficial effects: First, the path optimization function extracted in this application comprehensively considers welding trajectory smoothness, posture deviation control, and welding efficiency. By introducing a welding efficiency penalty term, curvature-adaptive global speed optimization is achieved. Traditional cost functions only constrain adjacent point velocities, accelerations, and gun normal deviations. While this ensures trajectory smoothness and local quality, it fails to proactively improve welding efficiency in low-curvature areas. This results in lower welding speeds in flat or low-curvature sections, which in turn prolongs the overall welding time. By applying a quadratic penalty to welding speeds below the desired speed, the path optimization function maximizes the welding speed within the permitted quality range, significantly shortening the welding cycle. The desired speed automatically adjusts to the local curvature, while areas with high curvature or complex corrugations reduce the desired speed, ensuring uncompromised welding quality. Simultaneously, smoothness, acceleration, and posture constraints ensure trajectory execution and weld pool stability. The welding efficiency penalty term is continuously differentiable, facilitating stable convergence of the optimization algorithm and avoiding speed jitter caused by boundary switching. Furthermore, the proposed path optimization function allows for adaptive welding speed adjustment on different corrugated plate surfaces, eliminating the need for manual global speed benchmarking, thus improving the versatility and automation of the welding system.

[0019] At the same time, this application integrates multimodal sensors such as laser vision, infrared thermal imaging, and acoustic sensing in front of the welding torch to simultaneously collect weld images, thermal distribution, and high-frequency acoustic signals, fully reflecting the geometric, thermal, and acoustic characteristics of the welding process. Combining feature extraction with a deep neural network defect recognition model, it can accurately and real-timely determine the type of weld defect and automatically adjust parameters such as welding speed, current, wire feed speed, and posture based on the different defect types, achieving closed-loop adaptive control of the welding process. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A schematic flow chart of an intelligent welding method for a curved surface of a container corrugated plate provided in one embodiment of the present invention; The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0021] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0022] The present embodiment provides a smart welding method for the curved surface of container corrugated sheeting. This method can be executed by at least one of a server, a terminal, or other electronic device capable of executing the method provided by the present embodiment. In other words, this method can be executed by software or hardware installed on a terminal or server device. The software can be a blockchain platform. The server can include, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0023] Reference Figure 1 , embodiment 1 of the present invention is: An intelligent welding method for a container corrugated plate curved surface comprises the following steps: S1: Use 3D scanning equipment to perform full coverage scanning on the container corrugated plate to obtain point cloud data of the container corrugated plate, and pre-process the point cloud data.

[0024] Use 3D scanning equipment to fully scan the container corrugated plate to obtain point cloud data of the container corrugated plate, including: The three-dimensional scanning equipment includes a high-precision three-dimensional laser scanner and a laser rangefinder; The scanning process of the full coverage scan is as follows: Use a laser rangefinder to measure the length, width, and corrugation direction of the container corrugated sheet, identify the peak position and trough position distribution of the corrugation direction, and generate a segmentation line perpendicular to the corrugation direction based on the peak position to divide the container corrugated sheet into multiple strip scanning areas; Specifically, a laser line emitted by a laser rangefinder is scanned along a certain direction of the container corrugated plate. If the scanning curve shows obvious periodic fluctuations, the direction is perpendicular to the corrugation direction, and the corrugation direction is thus determined. Use a high-precision 3D laser scanner to scan the height of the container corrugated plate in the corrugation direction to obtain the height sequence of the container corrugated plate in the corrugation direction; The first-order derivative is used to process the board height sequence to identify the peak position and the trough position in the board height sequence, wherein the peak position is when the first-order derivative of the position changes from positive to negative and the position is a local maximum, and the trough position is when the first-order derivative of the position changes from negative to positive and the position is a local minimum; A high-precision 3D laser scanner performs serpentine scanning of a strip-shaped scanning area at different viewing angles (e.g., 0 degrees, -30 degrees, 30 degrees, -35 degrees, 45 degrees, etc.), collecting point cloud data of the container corrugated sheet. This point cloud data consists of 3D position coordinates and their signal intensities. The serpentine scanning method involves a back-and-forth scanning along the corrugation direction, covering the entire strip-shaped scanning area line by line. With each line change, the high-precision 3D laser scanner moves along a path perpendicular to the scanning direction, ensuring an inter-row overlap rate of 20% to 30%. As one embodiment of this application, the high-precision 3D laser scanner's posture (position, rotation angle) and timestamp are recorded in real time during the scanning process to ensure subsequent multi-viewpoint point cloud registration. As an embodiment of the present application, the peak area can be scanned quickly (small curvature, simple geometric information), and the trough area can be scanned at a reduced speed (large curvature, rapid geometric changes, and requiring a higher sampling density), thereby improving the overall scanning efficiency. The peak area is a container corrugated plate area with a width of 10 cm generated at the peak position along a direction perpendicular to the corrugation direction, and the trough area is a container corrugated plate area with a width of 10 cm generated at the trough position along a direction perpendicular to the corrugation direction. It should be noted that in the container corrugated plate structure, to ensure strength and easy forming, the peak area will have a certain flattening transition section rather than a sharp peak, so the local curvature radius is larger (the curvature value is smaller), while the trough area is designed as a deeper depression to enhance rigidity, which makes the bottom radius of the trough smaller (the curvature value is larger).

[0025] Specifically, this application first uses a laser rangefinder to accurately measure the length and width dimensions and corrugation direction of the container corrugated plate, and then divides the area into strips based on the spatial distribution of the peak and trough positions, effectively avoiding repeated collection and data redundancy caused by overall large-scale blind scanning; uses dividing lines to limit the scanning task to the strip scanning area, and adopts a multi-view serpentine scanning strategy, which not only ensures full coverage and high overlap rate of the scanning data, but also improves the ability to capture complex surface geometric information; quickly scans the peak area with small curvature and simple geometric information to reduce invalid sampling time; slows down the scanning of the trough area with large curvature and rapid geometric changes to improve sampling density and data accuracy, thereby significantly reducing the total scanning time and data processing burden while ensuring the quality of the point cloud.

[0026] Preprocessing the point cloud data includes: The preprocessing includes point cloud registration, denoising and downsampling processing in sequence; The point cloud registration includes coarse registration and fine registration, wherein coarse registration is to perform preliminary alignment on the point cloud data obtained by multiple, multi-view scans so that the point cloud data of each view roughly overlap in the same coordinate system; fine registration is to further optimize the point cloud after coarse registration to minimize the overall point cloud fusion error, and fuse the same point cloud data under each view to remove redundant point cloud data, and construct the retained point cloud data into a point cloud data set after point cloud registration; as an embodiment of the present application, the SAC-IA algorithm is a coarse registration method, and the ICP algorithm is a fine registration method; The denoising process is as follows: after point cloud registration, the number of neighborhood points of the point cloud data and the average distance between the point cloud data and the neighborhood points are calculated; point cloud data with a number of neighborhood points lower than a preset threshold value are marked as noise points; point cloud data with an average distance value higher than a preset threshold value are marked as noise points; the noise points are removed; and the retained point cloud data are constructed into a denoised point cloud data set. For any point cloud data c, the neighborhood points of point cloud data c are point cloud data in the point cloud data set after point cloud registration whose distance from point cloud data c is within a preset radius (e.g., 10 cm); wherein the distance between point cloud data is the Euclidean distance between three-dimensional position coordinates in the point cloud data; The downsampling processing method is a voxel grid method, which uses the voxel grid method to perform downsampling processing on the denoised point cloud data set, and uses the point cloud data obtained by the downsampling processing as the preprocessed point cloud data.

[0027] Specifically, the voxelized grid method divides the position space of the container corrugated plate into cubic grids of fixed size, and maps the point cloud data in the denoised point cloud data set to the cubic grids corresponding to the positions according to the three-dimensional position coordinates, calculates the three-dimensional position coordinate centroid and the mean signal strength of all the point cloud data in the cubic grid, and uses the three-dimensional position coordinate centroid and the mean signal strength as the point cloud data obtained by downsampling processing of all the point cloud data in the cubic grid.

[0028] S2: Surface reconstruction is performed using the preprocessed point cloud data to generate a three-dimensional model of the container corrugated plate surface. The three-dimensional model of the container corrugated plate surface is then smoothed to obtain a smoothed three-dimensional model of the container corrugated plate surface.

[0029] Surface reconstruction is performed using pre-processed point cloud data to generate a 3D model of the container corrugated plate surface, including: Calculating the normal vectors of the preprocessed point cloud data, and constructing the normal vectors of all the preprocessed point cloud data into a gradient field, wherein the normal vectors of the preprocessed point cloud data are sampling vectors in the gradient field; Perform spatial interpolation on the gradient field so that the preprocessed point cloud data forms a continuous gradient field on the discrete voxel grid; The space where the container corrugated plate is located is divided into a multi-scale octree grid, where different scale levels correspond to different resolutions, and a discrete Poisson equation is constructed on the multi-scale octree network. The Poisson equation is iteratively solved using the finite element method to generate global continuous scalar fields at different scale levels, and the global continuous scalar field at the highest resolution is selected as the implicit scalar field obtained by the solution; Specifically, a global continuous scalar field is generated at a low resolution first, and then the resolution is gradually increased to optimize local details, thereby improving computational efficiency and preserving smoothness. The Marching Cubes algorithm is used to extract the isosurface of the implicit scalar field to obtain a closed, continuous and smooth three-dimensional surface mesh. The mesh vertices and triangular facets are generated as the three-dimensional model of the container corrugated plate surface.

[0030] As an embodiment of the present application, the three-dimensional surface mesh is composed of scalar values ​​of the surface position in an implicit scalar field, and the three-dimensional surface mesh is composed of multiple voxel meshes, each voxel mesh has 8 vertices, and each vertex has its scalar value in the implicit scalar field. For each voxel mesh, if the scalar values ​​of the 8 vertices have a sign change (across 0), the 8 vertices of the voxel mesh are used as mesh vertices.

[0031] The three-dimensional model of the container corrugated plate surface is smoothed to obtain a smoothed three-dimensional model of the container corrugated plate surface, including: Calculate the local curvature of the mesh vertices in the three-dimensional model of the container corrugated plate surface; the calculation process of the local curvature is as follows: Obtain the position coordinates of the neighboring vertices of the mesh vertex node whose local curvature is to be calculated, wherein the neighboring vertices of the mesh vertex node are the mesh vertices directly connected to the mesh vertex node in the three-dimensional model of the container corrugated plate surface; Calculate the diagonal angle between the mesh vertex node and the two triangles where the neighboring vertex is located. The mesh vertex node is directly connected to the neighboring vertex and both are located on the two triangles. Estimate the local curvature of the mesh vertex node, where the estimation formula is: ; in, Represents 1 / 3 of the area of ​​the triangle connected to the mesh vertex node. Represents the domain vertex set of the grid vertex node, Represents a set of domain vertices Any vertex in the domain, is the diagonal angle between the two triangles where the grid vertex node and the neighboring vertex e are located, represents the cotangent function, Represents the position coordinates of the domain vertex e, Represents the position coordinates of the grid vertex node, Indicates the modulus of the calculated vector; It should be noted that this curvature estimation method is based on the discrete geometry principle of neighborhood vertices. By constructing local patches and calculating the geometric differences between vertices and their neighbors, it can provide high-precision curvature approximation for mesh surfaces of arbitrary topological structure without relying on continuous surface equations. The introduction of vertex-related local area weights during the calculation process effectively suppresses the interference of irregular mesh distribution on curvature estimation, improving the stability and accuracy of the calculation results. Compared with traditional curvature fitting methods, this method is computationally lightweight and suitable for the rapid processing of large-scale point clouds or meshes. Furthermore, by directly utilizing local geometric structure information, it can achieve both low-noise response in smooth areas and sensitive capture of areas of high curvature. Based on the local curvature of the mesh vertices, the position coordinates of the mesh vertices in the three-dimensional model of the container corrugated plate surface are smoothed to obtain the smoothed three-dimensional model of the container corrugated plate surface.

[0032] Specifically, the smoothness correction formula of the position coordinates of the mesh vertex node is: ; in, Indicates the smoothness correction result of the position coordinates of the mesh vertex node. Indicates the smoothing step size, set is 0.1, Indicates sensitive control parameters, set is 0.3, Represents the local curvature of the domain vertex e of the mesh vertex node, Represents the local curvature difference of the domain vertex e of the mesh vertex node, represents the smoothing weight of the domain vertex e; When the local curvature difference is large, the smoothing weight is reduced to reduce the pull across features, thereby effectively protecting key geometric features such as peaks and troughs; when the local curvature difference is small, the weight is close to 1, making the smoothing effect more uniform and helping to eliminate noise.

[0033] S3: Based on the smoothed 3D model of the container corrugated plate surface, an intelligent welding planning algorithm is used to generate the optimal welding path, and the container corrugated plate is welded using the optimal welding path.

[0034] Based on the smoothed 3D model of the container corrugated plate surface, an intelligent welding planning algorithm is used to generate the optimal welding path, including: The process of the intelligent welding planning algorithm includes: Based on the local curvature of the mesh vertices in the smoothed 3D model of the container corrugated plate surface, mesh vertices are selected and included in the candidate trajectory point set; Specifically, the grid vertices whose local curvature is higher than the preset maximum curvature threshold and located in the peak area and valley area are included in the candidate trajectory point set; Selecting candidate trajectory points from the candidate trajectory point set as welding starting points to generate multiple groups of local welding paths; A path optimization function of the local welding path is constructed, in which the path optimization function uses the welding control parameters of the local welding path as optimization variables and the local curvature of the welding path coordinate points in the local welding path as control parameters to generate adaptive dynamic welding control parameters of the local welding path. During the welding process of the local welding path, the welding speed, welding acceleration and welding direction are adaptively controlled. The functional form of the constructed path optimization function is: ; ; ; ; ; ; ; in, represents the path optimization function, Represents the path optimization function value of the local welding path Load; Represents the local welding path The velocity smoothing constraint term of the nth welding path coordinate point in , Represents the local welding path The welding speed at the nth welding path coordinate point, Represents the local welding path The welding speed at the n-1th welding path coordinate point, Represents the local welding path The local curvature at the nth welding path coordinate point in , Indicates the preset maximum speed value (for example, 1 meter per second), represents the speed term smoothing coefficient (e.g. 0.3), Indicates the modulus of the calculation vector, N represents the local welding path The number of coordinate points in the welding path; is the weight coefficient, set They are 0.2, 0.2, 0.3, 0.3 respectively; Represents the local welding path The acceleration smoothing constraint term of the nth welding path coordinate point in , Represents the local welding path The welding acceleration at the nth welding path coordinate point in , Represents the local welding path The welding acceleration at the n-1th welding path coordinate point, Indicates the preset maximum acceleration value (for example, 0.5 meters per second squared). Indicates the acceleration smoothing coefficient (e.g. 0.2); Represents the local welding path The welding direction constraint item of the nth welding path coordinate point, Indicates the direction vector of the welding gun when welding the nth welding path coordinate point With mesh normal The angle deviation between Local welding path The normal vector of the grid where the nth welding path coordinate point is located, represents the deviation smoothing coefficient (e.g. 0.2), represents the inverse cosine function; the normal vector of the grid is the mean of the normal vectors of each face of the grid; Represents the local welding path The welding efficiency loss term of the nth welding path coordinate point in is, Represents the local welding path The expected speed of the nth welding path coordinate point in, Indicates selection The maximum value of the welding efficiency loss term is used to penalize unnecessary speed limit; The path optimization function of the local welding path is optimized and solved to form the optimal welding parameter sequence of the local welding path. The optimal welding parameter sequence of all local welding paths is used as the optimal welding path. The optimal welding parameter sequence is composed of the position coordinates of the welding path coordinate points in the local welding path, the optimal welding speed, the optimal welding acceleration, and the optimal welding gun orientation direction vector, among which the optimal welding speed, the optimal welding acceleration, and the optimal welding gun orientation direction vector are the solution results of the path optimization function.

[0035] As an embodiment of the present application, the optimization solution method of the path optimization function is an improved genetic algorithm, wherein the improvement strategy of the genetic algorithm is: increase the mutation probability of the partial path with lower fitness in the local welding path, explore more welding parameters, increase the crossover probability of the partial path with higher fitness, and enhance the local optimization effect, wherein the low fitness segment may have speed, acceleration or direction deviation from the optimal, increasing the mutation can jump out of the local inferior solution, and maintaining stable crossover in the high fitness segment can speed up the convergence speed while avoiding disturbing the welding quality; perturb the welding parameters of any welding path coordinate point in the current optimal welding parameter sequence, re-evaluate the path optimization function value, and retain the improved sequence.

[0036] Select a candidate trajectory point from the candidate trajectory point set as the welding starting point to generate a local welding path, including: Select a candidate trajectory point close to the edge of the container corrugated plate from the candidate trajectory point set as the welding starting point, where the candidate trajectory point is a grid vertex in the candidate trajectory point set; Taking the welding starting point as the starting point, an iterative method is used to generate a sequence of welding path coordinate points with continuous positions to form a set of local welding paths. The iterative formula is: ; ; in, represents the mth welding path coordinate point obtained by iteration, represents the m+1th welding path coordinate point obtained by iteration; specifically, when m=0, is the selected welding starting point; Indicates the iterative step size for iterating the mth welding path coordinate point, A direction vector representing the direction of the ripples, represents the basic step length, Indicates the minimum threshold of local curvature (e.g. 0.05 ), Indicates the welding path coordinate point The local curvature of If the welding path coordinate point The local curvature change rate Above the expected minimum rate of change threshold (e.g. 0.1 ), the iteration is terminated, the current local welding path is output, and the welding starting point is reselected.

[0037] S4: Use multimodal sensors to collect real-time weld status data of container corrugated plates during the welding process, use welding defect recognition models to identify welding defects based on the weld status data, and adjust the welding method based on the feedback of welding defect recognition results.

[0038] Multimodal sensors are used to collect real-time weld status data of container corrugated plates during the welding process. Weld defect recognition models are used to identify weld defects based on the weld status data, including: The multimodal sensor includes a laser vision sensor, an infrared thermal imaging sensor, and an acoustic sensor. The multimodal sensor is deployed in front of the welding gun and collects real-time weld status data of the container corrugated plate during the welding process. The laser vision sensor is used to obtain weld images, the infrared thermal imaging sensor is used to obtain weld heat distribution, and the acoustic sensor is used to obtain high-frequency acoustic signals during the welding process. The weld images, weld heat distribution, and high-frequency acoustic signals are used as weld status data. Extract features from weld state data to obtain multimodal weld features; As an embodiment of the present application, the multimodal weld features include weld contour features, a heat distribution uniformity index, and a spectrum feature of a high-frequency acoustic signal. The weld contour features are extracted using convolution filtering denoising and a Canny edge detection algorithm, and the absolute value of the heat distribution difference on both sides of the welding gun is calculated as the heat distribution uniformity index. A welding defect recognition model built based on a deep neural network is used to receive multimodal weld features and output welding defect categories as welding defect recognition results. The welding defect categories include normal welds, undercuts, weld bumps, incomplete penetration, pores, and cracks.

[0039] Adjust the welding method based on the feedback of welding defect identification results, including: The adjustment strategy of the welding method is: When undercut is detected, reduce the welding speed and welding current; When a weld nub is identified, reduce the wire feed speed or current and increase the welding gun movement speed appropriately; When pores are identified, maintain a stable welding gun posture during welding, use dry welding wire or clean the surface of the base material; When cracks are identified, reduce the welding speed, weld in sections or preheat the base material to reduce thermal stress; When incomplete penetration is detected, increase the welding current or heat input and slow down the welding speed.

[0040] Example 2: An intelligent welding system for container corrugated plate curved surfaces includes a three-dimensional reconstruction device, a path planning module, and a feedback adjustment device: The three-dimensional reconstruction device is used to use a three-dimensional scanning device to perform a full-coverage scan of the container corrugated plate to obtain point cloud data of the container corrugated plate, and pre-process the point cloud data, use the pre-processed point cloud data to perform surface reconstruction to generate a three-dimensional model of the container corrugated plate surface, and perform surface smoothing on the three-dimensional model of the container corrugated plate surface to obtain a smoothed three-dimensional model of the container corrugated plate surface; The path planning module is used to generate an optimal welding path based on the smoothed three-dimensional model of the container corrugated plate surface using an intelligent welding planning algorithm, and weld the container corrugated plate using the optimal welding path; The feedback adjustment device is used to use a multimodal sensor to collect real-time weld state data of the container corrugated plate during the welding process, use a welding defect recognition model to identify welding defects in the weld state data, and adjust the welding method based on the feedback of the welding defect recognition results; This is to realize the intelligent welding method for the curved surface of the container corrugated plate as described above.

[0041] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0042] It should be noted that the serial numbers of the above-mentioned embodiments of the present invention are for descriptive purposes only and do not represent the advantages or disadvantages of the embodiments. In addition, the terms "including", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "including a ..." does not exclude the presence of other identical elements in the process, device, article or method comprising the element.

[0043] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0044] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An intelligent welding method for the curved surface of container corrugated plate, characterized in that: The method comprises: S1: Use 3D scanning equipment to fully scan the container corrugated plate to obtain point cloud data of the container corrugated plate, and pre-process the point cloud data; S2: Surface reconstruction is performed using the pre-processed point cloud data to generate a three-dimensional model of the container corrugated plate surface, and the three-dimensional model of the container corrugated plate surface is smoothed to obtain a smoothed three-dimensional model of the container corrugated plate surface; S3: Based on the smoothed 3D model of the container corrugated plate surface, an intelligent welding planning algorithm is used to generate the optimal welding path, and the container corrugated plate is welded using the optimal welding path; S4: Use multimodal sensors to collect real-time weld status data of container corrugated plates during the welding process, use welding defect recognition models to identify welding defects based on the weld status data, and adjust the welding method based on the feedback of welding defect recognition results.

2. The intelligent welding method for curved surface of container corrugated plate according to claim 1, characterized in that: Use 3D scanning equipment to fully scan the container corrugated plate to obtain point cloud data of the container corrugated plate, including: The three-dimensional scanning equipment includes a high-precision three-dimensional laser scanner and a laser rangefinder; The point cloud data consists of three-dimensional position coordinates and signal strengths of the three-dimensional position coordinates.

3. The intelligent welding method for curved surface of container corrugated plate according to claim 2, characterized in that: Preprocessing the point cloud data includes: The preprocessing includes point cloud registration, denoising and downsampling processing in sequence; The point cloud registration includes coarse registration and fine registration, wherein coarse registration is to perform preliminary alignment on the point cloud data obtained by multiple and multi-view scans so that the point cloud data of each view roughly overlap in the same coordinate system; fine registration is to further optimize the point cloud after coarse registration, remove redundant point cloud data, and construct the retained point cloud data into a point cloud data set after point cloud registration; The denoising process is as follows: after point cloud registration, the number of neighborhood points of the point cloud data and the average distance between the point cloud data and the neighborhood points are calculated, the point cloud data with the number of neighborhood points lower than a preset threshold value is marked as noise points, the point cloud data with the average distance higher than the preset threshold value is marked as noise points, the noise points are removed, and the retained point cloud data is constructed into a denoised point cloud data set; The downsampling processing method is a voxel grid method, which uses the voxel grid method to perform downsampling processing on the denoised point cloud data set, and uses the point cloud data obtained by the downsampling processing as the preprocessed point cloud data.

4. The intelligent welding method for curved surface of container corrugated plate according to claim 3, characterized in that: Surface reconstruction is performed using pre-processed point cloud data to generate a 3D model of the container corrugated plate surface, including: Calculating the normal vectors of the preprocessed point cloud data, and constructing the normal vectors of all the preprocessed point cloud data into a gradient field, wherein the normal vectors of the preprocessed point cloud data are sampling vectors in the gradient field; Perform spatial interpolation on the gradient field so that the preprocessed point cloud data forms a continuous gradient field on the discrete voxel grid; The space where the container corrugated plate is located is divided into a multi-scale octree grid, where different scale levels correspond to different resolutions, and a discrete Poisson equation is constructed on the multi-scale octree network. The Poisson equation is iteratively solved using the finite element method to generate global continuous scalar fields at different scale levels, and the global continuous scalar field at the highest resolution is selected as the implicit scalar field obtained by the solution; The Marching Cubes algorithm is used to extract the isosurface of the implicit scalar field to obtain a closed, continuous and smooth three-dimensional surface mesh. The mesh vertices and triangular facets are generated as the three-dimensional model of the container corrugated plate surface.

5. The intelligent welding method for curved surface of container corrugated plate according to claim 4, characterized in that: The three-dimensional model of the container corrugated plate surface is smoothed to obtain a smoothed three-dimensional model of the container corrugated plate surface, including: Calculate the local curvature of mesh vertices in the 3D model of container corrugated plate surface; Based on the local curvature of the mesh vertices, the position coordinates of the mesh vertices in the three-dimensional model of the container corrugated plate surface are smoothed to obtain the smoothed three-dimensional model of the container corrugated plate surface.

6. The intelligent welding method for curved surface of container corrugated plate according to claim 5, characterized in that: Based on the smoothed 3D model of the container corrugated plate surface, an intelligent welding planning algorithm is used to generate the optimal welding path, including: The process of the intelligent welding planning algorithm includes: Based on the local curvature of the mesh vertices in the smoothed 3D model of the container corrugated plate surface, mesh vertices are selected and included in the candidate trajectory point set; Selecting candidate trajectory points from the candidate trajectory point set as welding starting points to generate multiple groups of local welding paths; A path optimization function of the local welding path is constructed, in which the path optimization function uses the welding control parameters of the local welding path as the optimization variables and the local curvature of the welding path coordinate points in the local welding path as the control parameters to generate the adaptive dynamic welding control parameters of the local welding path. During the welding process of the local welding path, the welding speed, welding acceleration and welding direction are adaptively controlled. The functional form of the constructed path optimization function is: ； ； ； ； ； ； ； in, represents the path optimization function, Represents the path optimization function value of the local welding path Load; Represents the local welding path Middle The velocity smoothing constraint term of the welding path coordinate point is: Represents the local welding path The welding speed at the nth welding path coordinate point, Represents the local welding path The welding speed at the n-1th welding path coordinate point, Represents the local welding path The local curvature at the nth welding path coordinate point in , Indicates the preset maximum speed value. represents the velocity term smoothing coefficient, Indicates the modulus of the calculation vector, N represents the local welding path The number of coordinate points in the welding path; is the weight coefficient, set They are 0.2, 0.2, 0.3, 0.3 respectively; Represents the local welding path The acceleration smoothing constraint term of the nth welding path coordinate point in , Represents the local welding path The welding acceleration at the nth welding path coordinate point in , Represents the local welding path The welding acceleration at the n-1th welding path coordinate point, Indicates the preset maximum acceleration value. represents the acceleration smoothing coefficient; Represents the local welding path The welding direction constraint item of the nth welding path coordinate point, Indicates the direction vector of the welding gun when welding the nth welding path coordinate point With mesh normal The angle deviation between Local welding path The normal vector of the grid where the nth welding path coordinate point is located, represents the deviation smoothing coefficient, represents the arccosine function; Represents the local welding path The welding efficiency loss term of the nth welding path coordinate point in is, Represents the local welding path The expected speed of the nth welding path coordinate point in, Indicates selection The maximum value of the welding efficiency loss term is used to penalize unnecessary speed limit; The path optimization function of the local welding path is optimized and solved to form the optimal welding parameter sequence of the local welding path. The optimal welding parameter sequence of all local welding paths is used as the optimal welding path. The optimal welding parameter sequence is composed of the position coordinates of the welding path coordinate points in the local welding path, the optimal welding speed, the optimal welding acceleration, and the optimal welding gun orientation direction vector, among which the optimal welding speed, the optimal welding acceleration, and the optimal welding gun orientation direction vector are the solution results of the path optimization function.

7. The intelligent welding method for curved surface of container corrugated plate according to claim 6, characterized in that: Select a candidate trajectory point from the candidate trajectory point set as the welding starting point to generate a local welding path, including: Select a candidate trajectory point close to the edge of the container corrugated plate from the candidate trajectory point set as the welding starting point, where the candidate trajectory point is a grid vertex in the candidate trajectory point set; Taking the welding starting point as the starting point, an iterative method is used to generate a sequence of welding path coordinate points with continuous positions to form a set of local welding paths.

8. The intelligent welding method for curved surface of container corrugated plate according to claim 1, characterized in that: Multimodal sensors are used to collect real-time weld status data of container corrugated plates during the welding process. Weld defect recognition models are used to identify weld defects based on the weld status data, including: The multimodal sensor includes a laser vision sensor, an infrared thermal imaging sensor, and an acoustic sensor. The multimodal sensor is deployed in front of the welding gun to collect real-time weld status data of the container corrugated plate during the welding process; Extract features from weld state data to obtain multimodal weld features; A welding defect recognition model built based on a deep neural network is used to receive multimodal weld features and output welding defect categories as welding defect recognition results. The welding defect categories include normal welds, undercuts, weld bumps, incomplete penetration, pores, and cracks.

9. The intelligent welding method for curved surface of container corrugated plate according to claim 8, characterized in that: Adjust the welding method based on the feedback of welding defect identification results, including: The adjustment strategy of the welding method is: When undercut is detected, reduce the welding speed and welding current; When a weld nub is identified, reduce the wire feed speed or current and increase the welding gun movement speed appropriately; When pores are identified, maintain a stable welding gun posture during welding, use dry welding wire or clean the surface of the base material; When cracks are identified, reduce the welding speed, weld in sections or preheat the base material to reduce thermal stress; When incomplete penetration is detected, increase the welding current or heat input and slow down the welding speed.

10. An intelligent welding system for the curved surface of container corrugated plate, characterized in that: The intelligent welding system for the curved surface of container corrugated plate includes a three-dimensional reconstruction device, a path planning module, and a feedback adjustment device: The three-dimensional reconstruction device is used to use a three-dimensional scanning device to perform a full-coverage scan of the container corrugated plate to obtain point cloud data of the container corrugated plate, and pre-process the point cloud data, use the pre-processed point cloud data to perform surface reconstruction to generate a three-dimensional model of the container corrugated plate surface, and perform surface smoothing on the three-dimensional model of the container corrugated plate surface to obtain a smoothed three-dimensional model of the container corrugated plate surface; The path planning module is used to generate an optimal welding path based on the smoothed three-dimensional model of the container corrugated plate surface using an intelligent welding planning algorithm, and weld the container corrugated plate using the optimal welding path; The feedback adjustment device is used to use a multimodal sensor to collect real-time weld state data of the container corrugated plate during the welding process, use a welding defect recognition model to identify welding defects in the weld state data, and adjust the welding method based on the feedback of the welding defect recognition results; To realize the intelligent welding method for the curved surface of container corrugated plate as described in any one of claims 1-9.

Images