Welding seam-oriented ocean engineering steel structure welding method

Abstract

The invention relates to the technical field of robot welding, in particular to a welding seam-oriented ocean engineering steel structure welding method. The method comprises the following steps: collecting welding process data required by steel structure welding; recognizing and storing the boundary dimension and the welding seam distribution of the workpiece; a 3D camera is used for scanning an actual workpiece to obtain point cloud data, and the point cloud data and workpiece and weld joint data are matched and corrected by means of a coordinate registration algorithm; welding seams are grouped, and robot welding process planning data are generated; scanning the welding seam by adopting a laser scanning technology, extracting geometric characteristic parameters, and comparing the geometric characteristic parameters with welding seam distribution to calculate deviation; a self-adaptive algorithm is used for adjusting welding parameters to complete backing weld bead forming; and repeating the steps of scanning, feature extraction and parameter fine adjustment after each layer of welding is completed until filling welding is completed. The problems that a traditional welding system is complex in process and low in efficiency are solved by accurately obtaining the weld joint data, matching the optimal process and planning the welding sequence.

Description

Technical Field

[0001] This invention relates to the field of robotic welding technology, and more specifically, to a welding method for marine engineering steel structures oriented towards weld seams. Background Technology

[0002] Marine engineering steel structures typically operate in extreme environments, placing extremely high demands on welding. Currently, marine engineering steel structures primarily involve multi-layer, multi-pass welding of medium-thick plates. Traditional manual welding, influenced by welding experience, struggles to guarantee weld quality. Replacing traditional manual operations with welding robot systems is a crucial approach to achieving intelligent welding. However, traditional welding robot systems typically focus on the entire workpiece. For each type of steel structure, comprehensive path planning and welding process parameter design are required, leading to cumbersome and lengthy pre-weld preparation processes. This is especially problematic for medium-thick plate welds, where weld precision is affected by workpiece machining and assembly accuracy. Specific weld details and individual requirements are difficult to meet, resulting in inconsistent welding effects and quality. Therefore, there is an urgent need to design a weld-oriented welding method for marine engineering steel structures to address these issues.

[0003] This invention focuses on the weld seam, specifically its shape, size, and spatial distribution. It customizes and adjusts welding paths and process parameters, simplifying the cumbersome process of complex planning and parameter adjustments for each workpiece in traditional systems. Through weld seam identification and characteristic analysis, the system can quickly extract the weld seam's geometric characteristics and process requirements, enabling automated welding path planning and parameter setting. This approach offers greater flexibility, eliminating the need to design complete welding processes for each workpiece individually. Instead, it achieves rapid adaptation to various workpieces and welding requirements through real-time analysis and adaptive adjustment of weld seam characteristics. Summary of the Invention

[0004] The purpose of this invention is to provide a welding method for marine engineering steel structures oriented towards weld seams, in order to solve the problems of complex processes and low efficiency in traditional welding systems that take the workpiece as the core, as mentioned in the background art.

[0005] To achieve the above objectives, the present invention aims to provide a welding method for marine engineering steel structures oriented towards weld seams, comprising the following steps: S1. Collect welding process data required for steel structure welding and establish a welding process database; S2. Read the geometric shape and weld location data of the steel structure workpiece through the model file, identify and store the workpiece's external dimensions and weld distribution; S3. Scan the actual workpiece located in the welding area with a 3D camera to obtain the point cloud data of the workpiece; use a coordinate registration algorithm to match and correct the point cloud data with the workpiece shape and weld distribution, and calculate the precise position of the weld in the actual workpiece coordinate system. S4. Based on the precise location of the weld, match welding process data from the welding process database for each weld to generate an optimal template; group all welds and plan the welding sequence, and combine the optimal template to generate robot welding process planning data for controlling the welding robot; S5. Based on the results of S4, laser scanning technology is used to scan the weld, and the geometric feature parameters of the weld are extracted using artificial intelligence algorithms; the geometric feature parameters are compared with the weld distribution, and the deviation is calculated. S6. If an abnormality is detected in the geometric feature parameters, the welding path and welding process parameters are adjusted through an adaptive algorithm to complete the formation of the root pass. S7. Repeat the process from S5 to S6. That is, after each layer of welding is completed, the weld is re-scanned with laser and features are extracted. Based on the extracted feature parameters, the welding path and welding process parameters of the next layer are finely adjusted until the filling welding is completed.

[0006] As a further improvement to this technical solution, in S1, the welding process data includes at least material information, welding process parameters, and welding layer planning rules for medium and thick plates. The material information includes the steel plate material and thickness; Welding process parameters include welding current, welding voltage, welding speed, wire feed speed, welding torch oscillation amplitude, and welding torch posture; The planning rules for welding layers of medium and heavy plates include the number of welding passes for each layer, the welding sequence of each weld, the interpass temperature control range, and the welding path trajectory for each layer.

[0007] As a further improvement to this technical solution, the specific steps involved in identifying and storing the workpiece dimensions and weld distribution in step S2 are as follows: Read the model file and extract the workpiece's external dimensions in three-dimensional space, denoted as... ,in Represents the length of the workpiece. Represents the workpiece width. Represents the height of the workpiece; Locate all weld joints, and then extract the core geometric properties of each weld, including weld type and theoretical weld length. Theoretical weld width Theoretical bevel angle Theoretical root gap and theoretical bevel depth ; For straight welds, their spatial position and orientation in the workpiece coordinate system are defined by extracting the starting point coordinates and the ending point coordinates; For curved welds, a non-uniform rational B-spline curve is used for fitting, and its control point coordinate sequence and node vector are stored to describe its spatial position and orientation. The core geometric properties of the weld are associated with its spatial location and orientation, generating a unique identifier for each weld. ; Identifiers for all welds The core geometric attributes, spatial location, and orientation are used as the weld distribution, and the workpiece dimensions and weld distribution are stored in a structured manner in the system database.

[0008] As a further improvement to this technical solution, in step S3, the specific steps involved in matching and correcting the point cloud data with the workpiece's external dimensions and weld distribution using a coordinate registration algorithm to calculate the precise location of the weld are as follows: For raw point cloud data Preprocessing is performed using a statistical outlier removal algorithm to obtain denoised point cloud data. ; From the denoised point cloud data Extract the macroscopic geometric features of the workpiece, including the main outer contour planes, protruding edges and corners; A feature-based registration method is used to match macroscopic geometric features with the workpiece's external dimensions and core geometric properties to solve for an initial rigid body transformation matrix. This completes the initial alignment of the point cloud; Based on the initial alignment, the iterative nearest point algorithm is used for fine registration to solve for the optimal rigid body transformation matrix. ; Based on the optimal rigid body transformation matrix The spatial position and orientation in S2 are uniformly transformed to the actual workpiece coordinate system to obtain the precise position of each weld in actual space.

[0009] As a further improvement to this technical solution, in step S4, the specific steps involved in matching welding process data from the welding process database for each weld seam to generate the optimal template are as follows: Based on the precise location of the weld, the spatial orientation features of the current weld are extracted, including the spatial vector of the weld in the workpiece coordinate system. Angle of inclination of the weld to the horizontal plane The normal vector of the base material surface where the weld is located ; Determine the welding difficulty coefficient based on the spatial orientation characteristics of the weld. ; By combining the core geometric properties and spatial orientation features of the weld, a multi-dimensional matching query is constructed in the welding process database; Based on the bevel angle, welding difficulty coefficient, and root gap, the matching degree between the current weld and each process template in the welding process database is calculated. ; For tilt angle Greater than the horizontal welding threshold The welding position, based on the matching process template, and according to the welding difficulty coefficient. The welding current, welding voltage, and wire feed speed are corrected to obtain the corrected welding current, welding voltage, and wire feed speed; For each weld, select the matching degree. The minimum process template, combined with the modified welding current, welding voltage, and wire feed speed, yields the optimal template for all weld seams.

[0010] As a further improvement to this technical solution, the specific steps involved in S4, which involve grouping all weld seams and planning the welding sequence to generate robot welding process planning data, are as follows: The spatial distance between weld seams is less than a set distance threshold. Welds of the same type are grouped together and labeled with group numbers. ; Calculate the average of the start and end points of all welds in each group to obtain the coordinates of the geometric center point of that group. ; Using the starting point of the welding robot as a reference, according to the coordinates of the geometric center points of each group Sort the Euclidean distances from the starting point in ascending order to determine the welding sequence between groups. ; For each weld within each group, if its spatial vector Normal vector to the surface of the base material If the dot product is not less than 0, the welding direction is from the start point to the end point; otherwise, it is from the end point to the start point, thus obtaining the welding direction mark. ; For fillet welds and corner welds, set dwell time parameters at the start and end points of the welding path. At the same time, based on the welding difficulty coefficient Adjust the welding speed of the group containing this weld. ; Group number Inter-group welding sequence identifier Precise location of the weld and markings for the welding direction Welding speed Duration of stay The optimal template is integrated into robot welding process planning data and output.

[0011] As a further improvement to this technical solution, the specific steps involved in extracting the geometric feature parameters of the weld through artificial intelligence algorithms in step S5 are as follows: Laser scanning technology is used to scan the weld seam and obtain three-dimensional point cloud data of the weld seam surface. ; 3D point cloud data Data simplification was performed using a voxel grid downsampling method, and noise points were filtered out using a statistical outlier removal algorithm to obtain preprocessed point cloud data. ; Based on the precise weld location obtained from S3, in the preprocessed point cloud data The section is centered on the theoretical centerline of the weld, with a width equal to the section width. Region of interest; Within the region of interest, a feature point detection algorithm based on curvature change is used to calculate the curvature value of each point in the point cloud, and points with curvature greater than a curvature threshold are selected. The points are marked as feature points; The weld width is obtained by calculating the average Euclidean distance between the feature points on both sides of the bevel edge. ; The bevel depth is obtained by calculating the average vertical distance from the point cloud at the bottom of the bevel to the reference plane on the surface of the base material. ; The bevel angle is obtained by fitting the bevel planes on both sides and calculating the angle between the normal vectors of the two planes. ; The root gap is obtained by calculating the average distance between the nearest neighbors on both sides of the bevel root. ; Extracted weld width Bevel depth , bevel angle and root gap With weld identifier Associated and output as geometric feature parameters.

[0012] As a further improvement to this technical solution, in step S5, the geometric feature parameters are compared with the weld distribution, and the specific steps involved in calculating the deviation are as follows: The extracted geometric feature parameters are compared with the theoretical weld width in the weld distribution. Theoretical bevel angle Theoretical root gap and theoretical bevel depth Compare and calculate the width deviation, depth deviation, angle deviation, and gap deviation.

[0013] As a further improvement to this technical solution, the specific steps involved in adjusting the welding path and welding process parameters through the adaptive algorithm in step S6 are as follows: Set width deviation threshold Depth deviation threshold Angle deviation threshold and gap deviation threshold ; When the width deviation exceeds the width deviation threshold At the same time, the welding current is reduced proportionally according to the difference between the width deviation and the width deviation threshold, while the welding speed is increased proportionally to obtain the welding current adjustment amount and the welding speed adjustment amount. When the depth deviation exceeds the depth deviation threshold At that time, the wire feeding speed is increased proportionally based on the difference between the depth deviation and the depth deviation threshold to obtain the wire feeding speed adjustment amount; When the angle deviation exceeds the angle deviation threshold At the same time, the welding torch posture angle is adjusted proportionally according to the difference between the angle deviation and the angle deviation threshold, and the welding path trajectory is corrected proportionally to obtain the welding torch posture angle adjustment amount and the welding path trajectory correction amount. When the gap deviation exceeds the gap deviation threshold At that time, the welding torch oscillation amplitude is increased proportionally according to the difference between the gap deviation and the gap deviation threshold, so as to obtain the welding torch oscillation amplitude adjustment amount; Set an upper limit for the adjustment of welding process parameters to ensure that the adjustment range in a single instance does not exceed the positive and negative adjustment limits of the original welding process parameters. Among them, the adjustment amounts of welding process parameters include welding current adjustment, welding speed adjustment, wire feed speed adjustment, welding torch posture angle adjustment, welding path trajectory correction, and welding torch oscillation amplitude adjustment. The adjusted welding process parameters are superimposed onto the original welding process parameters to obtain the fine-tuned welding process parameters.

[0014] As a further improvement to this technical solution, the specific steps involved in S7, which involve fine-tuning the welding path and welding process parameters of the next layer based on the extracted feature parameters, are as follows: Obtain the actual geometric feature parameters and current welding process parameters obtained from the post-weld scan of the current layer; Calculate the deviation of each geometric feature parameter from its corresponding theoretical value; The adjustment amount of each welding process parameter is calculated based on the deviation, and the adjusted welding process parameters are obtained. Fine-tune the welding path based on the welding path trajectory correction amount in the adjustment amount of welding process parameters; The adjusted welding process parameters and welding path are then sent out to execute the next layer of welding.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. In this welding method for marine engineering steel structures oriented towards welds, the weld is the core. By identifying the distribution of welds and obtaining their precise locations, the optimal welding process template is matched for each weld and the welding sequence is planned. Compared with the traditional cumbersome process that takes the workpiece as the core and requires redesigning the path and parameters for each type of steel structure, there is no need to design a complete welding process for each one. It can quickly adapt to a variety of workpieces and welding requirements, greatly simplifying the pre-welding preparation process and solving the problems of complex process and low efficiency in traditional systems.

[0016] 2. In this method for welding marine engineering steel structures with weld seams, the geometric feature parameters of the weld seam are extracted and the deviation is calculated by artificial intelligence algorithm. Based on the deviation, the welding parameters (such as current, speed, welding torch trajectory, etc.) are adjusted in real time using an adaptive algorithm. Moreover, after each layer of welding is completed, the scanning and fine-tuning are repeated. Compared with the traditional manual welding, which is greatly affected by experience and welding robots are unable to cope with the problem of weld seam detail deviation caused by the precision of workpiece processing, this method can dynamically adapt to the actual state of the weld seam and effectively ensure the weld seam formation quality and welding quality stability of multi-layer and multi-pass welding of medium and thick plates. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the overall method of the present invention; Figure 2 This is a schematic diagram of the process of the present invention; Figure 3 This is a schematic diagram of weld grouping. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example: Please see Figure 1 As shown, this embodiment provides a welding method for marine engineering steel structures oriented towards welds, including the following steps: S1. Collect welding process data required for steel structure welding and establish a welding process database; In this embodiment, the welding process data includes at least material information, welding process parameters, and welding layer planning rules for medium and thick plates. The material information includes the steel plate material and thickness; Welding process parameters include welding current, welding voltage, welding speed, wire feed speed, welding torch oscillation amplitude, and welding torch posture; The planning rules for welding layers of medium and heavy plates include the number of welding passes for each layer, the welding sequence of each weld, the interpass temperature control range, and the welding path trajectory for each layer.

[0020] S2. Read the geometric shape and weld location data of the steel structure workpiece through the model file, identify and store the workpiece's external dimensions and weld distribution; In this embodiment, the model file is read, and the external dimensions of the workpiece in three-dimensional space are extracted, denoted as... ,in Represents the length of the workpiece. Represents the workpiece width. This represents the workpiece height, in millimeters. Locate all weld joints, and then for each weld, extract its core geometric properties, including weld type (flat, vertical, or overhead) and theoretical weld length. Theoretical weld width Theoretical bevel angle Theoretical root gap and theoretical bevel depth ; For straight welds, the starting point coordinates are extracted. and endpoint coordinates To define its spatial position and orientation in the workpiece coordinate system; The formula for calculating the direction of a straight weld is: in, The coordinates of the starting point of the theoretical weld location are given. This represents the coordinate value of the weld start point in the three-dimensional workpiece coordinate system. The coordinates of the endpoint of the theoretical weld position are given. This represents the coordinates of the weld endpoint in the three-dimensional workpiece coordinate system. It is a spatial vector indicating the direction and length of the weld; For curved welds, a non-uniform rational B-spline curve is used for fitting, and its control point coordinate sequence and node vector are stored to describe its spatial position and orientation. The core geometric properties of the weld are associated with its spatial location and orientation, generating a unique identifier for each weld. ; Identifiers for all welds The core geometric attributes, spatial location, and orientation are used as the weld distribution, and the workpiece dimensions and weld distribution are stored in a structured manner in the system database.

[0021] S3. Scan the actual workpiece located in the welding area with a 3D camera to obtain the point cloud data of the workpiece; use a coordinate registration algorithm to match and correct the point cloud data with the workpiece shape and weld distribution, and calculate the precise position of the weld in the actual workpiece coordinate system. In this embodiment, the original point cloud data Preprocessing is performed using a statistical outlier removal algorithm to obtain denoised point cloud data. ; The preprocessing step employs a statistical outlier removal algorithm on the raw point cloud data. For each point in the array, calculate its distance to... Average distance of (set to 50) nearest neighbors Assuming the distance follows a Gaussian distribution, add the distance that exceeds the mean. (Set to 1.5) times the standard deviation (i.e.) Points within a certain range are considered noise points and filtered out. From the denoised point cloud data Extract the macroscopic geometric features of the workpiece, including the main outer contour planes, protruding edges and corners; A feature-based registration method is used to match macroscopic geometric features with the workpiece's external dimensions and core geometric properties to solve for an initial rigid body transformation matrix. This completes the initial alignment of the point cloud; Based on the initial alignment, the iterative nearest point algorithm is used for fine registration to solve for the optimal rigid body transformation matrix. ; The iterative nearest-point algorithm iteratively calculates the optimal rigid body transformation matrix, ensuring that the theoretical model surface and the denoised point cloud data are aligned. Average distance error between Minimize, the calculation formula is: in, These are points in the denoised point cloud data. It is the distance on the surface of the theoretical model. The nearest point, The number of point pairs involved in the calculation; This represents the average distance error between the theoretical model surface and the denoised point cloud data. Based on the optimal rigid body transformation matrix The spatial position and orientation in S2 are uniformly transformed to the actual workpiece coordinate system to obtain the precise position of each weld in actual space; in, The coordinates of the starting point of the actual weld position. The coordinates of the end point of the actual weld location.

[0022] S4. Based on the precise location of the weld, match welding process data from the welding process database for each weld to generate an optimal template; group all welds and plan the welding sequence, and combine the optimal template to generate robot welding process planning data for controlling the welding robot; In this embodiment, based on the precise location of the weld, the spatial orientation features of the current weld are extracted, including the spatial vector of the weld in the workpiece coordinate system. Angle of inclination of the weld to the horizontal plane The normal vector of the base material surface where the weld is located ; Specifically, spatial vectors Based on the definition of weld direction, for straight welds, For curved welds, Determined by connecting the first and last control points; Angle of inclination of the weld to the horizontal plane Through spatial vectors The angle with the horizontal plane is calculated. Normal vector of the base material surface where the weld is located Based on the workpiece point cloud data after S3 fine registration, the point cloud set of adjacent regions on both sides of the weld is extracted, and least-squares plane fitting is performed on this point cloud set. The resulting plane unit normal vector is then obtained. Its direction points outward from the workpiece; Determine the welding difficulty coefficient based on the spatial orientation characteristics of the weld. ; in, The angle of inclination of the weld to the horizontal plane. Let be the normal vector of the base material surface where the weld is located. The vertical unit vector is (0,0,1); This is the welding difficulty coefficient; a larger value indicates a more complex welding posture. By combining the core geometric properties and spatial orientation features of the weld, a multi-dimensional matching query is constructed in the welding process database; Based on the bevel angle, welding difficulty coefficient, and root gap, the matching degree between the current weld and each process template in the welding process database is calculated. ; in, For the theoretical bevel angle, For the theoretical root gap, The bevel angle in the process template, This refers to the root gap in the process template. This represents the maximum allowable range of bevel angles. This represents the maximum permissible range of the root clearance. This is the weighting coefficient for the bevel angle (taken as 0.4). This is the weighting factor for the welding difficulty coefficient (taken as 0.3). The weighting factor for the root gap is 0.3. The value ranges from 0 to 1, representing the degree of matching between the weld and the process template. A smaller value indicates a higher degree of matching. The process templates in the welding process database are constructed based on historical welding process qualification data. Each template is associated with a specific set of joint geometry conditions and welding process parameters. The joint geometry conditions include theoretical bevel angles and theoretical root gaps, which are used to quickly assign appropriate welding process parameters to welds with similar geometric features. For tilt angle Greater than the horizontal welding threshold For welding positions (such as 45°), based on the matching process template and the welding difficulty coefficient... The welding current, welding voltage, and wire feed speed are corrected to obtain the corrected welding current, welding voltage, and wire feed speed; Specifically, for the tilt angle For vertical or overhead welding positions, based on the matching process template, adjust the welding current, welding voltage, and wire feed speed accordingly, with the adjustment range corresponding to the welding difficulty coefficient. They are positively correlated to overcome the effect of gravity on the molten pool; For each weld, select the matching degree. The minimum process template, combined with the modified welding current, welding voltage and wire feed speed, yields the optimal template for all welds; Furthermore, the spatial distance between welds is made smaller than a set distance threshold. Welds of the same type are grouped together and labeled with group numbers. ; in, The preset distance threshold is usually set to 200mm; grouping is to reduce the robot's movement path and avoid the superposition of thermal deformation. Calculate the average of the start and end points of all welds in each group to obtain the coordinates of the geometric center point of that group. ; in, For the precise coordinates of the weld start point, For the precise coordinates of the weld end point, This represents the number of welds in this group; The coordinates of the geometric center points of a set of welds; Using the starting point of the welding robot as a reference, according to the coordinates of the geometric center points of each group Sort the Euclidean distances from the starting point in ascending order to determine the welding sequence between groups. ; For each weld within each group, if its spatial vector Normal vector to the surface of the base material If the dot product is not less than 0, the welding direction is from the start point to the end point; otherwise, it is from the end point to the start point, thus obtaining the welding direction mark. ; For fillet welds and corner welds, set dwell time parameters at the start and end points of the welding path. (Typically a value of 0.5s to 1.0s), used to avoid undercut and ensure fusion; at the same time, depending on the welding difficulty coefficient Adjust the welding speed of the group containing this weld. ; The adjusted formula is as follows: in, As a reference welding speed, Difficulty Influence Coefficient (usually) ); Group number Inter-group welding sequence identifier Precise location of the weld and markings for the welding direction Welding speed Duration of stay The optimal template is integrated into robot welding process planning data and output.

[0023] S5. Based on the results of S4, laser scanning technology is used to scan the weld, and the geometric feature parameters of the weld are extracted using artificial intelligence algorithms; the geometric feature parameters are compared with the weld distribution, and the deviation is calculated. In this embodiment, laser scanning technology is used to scan the weld seam and obtain three-dimensional point cloud data of the weld seam surface. ; 3D point cloud data Data simplification was performed using a voxel grid downsampling method (voxel size set to 1mm×1mm×1mm), and noise points were filtered out using a statistical outlier removal algorithm to obtain preprocessed point cloud data. ; Based on the precise weld location obtained from S3, in the preprocessed point cloud data The section is centered on the theoretical centerline of the weld, with a width equal to the section width. Region of interest (cropping width) It is usually set to twice the theoretical weld width). Within the region of interest, a feature point detection algorithm based on curvature change is used to calculate the curvature value of each point in the point cloud, and points with curvature greater than a curvature threshold are selected. Points with values ​​ranging from 0.05 to 0.1 are marked as feature points; The weld width is obtained by calculating the average Euclidean distance between the feature points on both sides of the bevel edge. ; The bevel depth is obtained by calculating the average vertical distance from the point cloud at the bottom of the bevel to the reference plane on the surface of the base material. ; The bevel angle is obtained by fitting the bevel planes on both sides and calculating the angle between the normal vectors of the two planes. ; The root gap is obtained by calculating the average distance between the nearest neighbors on both sides of the bevel root. ; Extracted weld width Bevel depth , bevel angle and root gap With weld identifier Associated and output as geometric feature parameters; Furthermore, the extracted geometric feature parameters are compared with the theoretical weld width in the weld distribution. Theoretical bevel angle Theoretical root gap and theoretical bevel depth Compare and calculate the width deviation, depth deviation, angle deviation, and gap deviation; in, For width deviation, For depth deviation, For angular deviation, This refers to the gap deviation.

[0024] S6. If an abnormality is detected in the geometric feature parameters, the welding path and welding process parameters are adjusted through an adaptive algorithm to complete the formation of the root pass. In this embodiment, a width deviation threshold is set. (e.g., 0.15), depth deviation threshold (e.g., 0.10), angle deviation threshold (e.g., 5°) and gap deviation threshold (e.g., 0.20); When the width deviation exceeds the width deviation threshold At the same time, the welding current is reduced proportionally according to the difference between the width deviation and the width deviation threshold, while the welding speed is increased proportionally to obtain the welding current adjustment amount and the welding speed adjustment amount. in, This is the current adjustment coefficient. For speed adjustment coefficient, The original welding current, This represents the original welding speed; This is the welding current adjustment amount. Adjustment amount for welding speed; When the depth deviation exceeds the depth deviation threshold At that time, the wire feeding speed is increased proportionally based on the difference between the depth deviation and the depth deviation threshold to obtain the wire feeding speed adjustment amount; in, This is the wire feed speed adjustment coefficient. This is the original wire feeding speed; Adjustment amount for wire feeding speed; When the angle deviation exceeds the angle deviation threshold At the same time, the welding torch posture angle is adjusted proportionally according to the difference between the angle deviation and the angle deviation threshold, and the welding path trajectory is corrected proportionally to obtain the welding torch posture angle adjustment amount and the welding path trajectory correction amount. in, This is the attitude adjustment coefficient. This is the amount of adjustment for the welding torch's posture angle. For trajectory correction coefficients, For weld length, This is the correction amount for the welding path trajectory; When the gap deviation exceeds the gap deviation threshold At that time, the welding torch oscillation amplitude is increased proportionally according to the difference between the gap deviation and the gap deviation threshold, so as to obtain the welding torch oscillation amplitude adjustment amount; in, This is the adjustment coefficient for the swing amplitude of the gap. This represents the original swing amplitude. This is the adjustment amount for the welding torch oscillation amplitude; Set an upper limit for the adjustment of welding process parameters to ensure that the adjustment range in a single instance does not exceed the positive and negative adjustment limits of the original welding process parameters. (e.g., ±30%); among which, the adjustment amount of welding process parameters includes the adjustment amount of welding current, welding speed, wire feed speed, welding torch posture angle, welding path trajectory correction and welding torch oscillation amplitude. The adjusted welding process parameters are superimposed onto the original welding process parameters to obtain the fine-tuned welding process parameters.

[0025] S7. Repeat the process from S5 to S6. That is, after each layer of welding is completed, the weld is laser scanned and features are extracted again. Based on the extracted feature parameters, the welding path and welding process parameters of the next layer are fine-tuned until the filling welding is completed. In this embodiment, the actual geometric feature parameters and current welding process parameters obtained by scanning after welding the current layer are acquired. Calculate the deviation of each geometric feature parameter from its corresponding theoretical value; The adjustment amount of each welding process parameter is calculated based on the deviation, and the adjusted welding process parameters are obtained. Fine-tune the welding path based on the welding path trajectory correction amount in the adjustment amount of welding process parameters; The adjusted welding process parameters and welding path are then sent out to execute the next layer of welding.

[0026] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A welding method for marine engineering steel structures oriented towards weld seams, characterized in that: Includes the following steps: S1. Collect welding process data required for steel structure welding and establish a welding process database; S2. Read the geometric shape and weld location data of the steel structure workpiece through the model file, identify and store the workpiece's external dimensions and weld distribution; S3. Scan the actual workpiece located in the welding area with a 3D camera to obtain the point cloud data of the workpiece; use a coordinate registration algorithm to match and correct the point cloud data with the workpiece shape and weld distribution, and calculate the precise position of the weld in the actual workpiece coordinate system. S4. Based on the precise location of the weld, match welding process data from the welding process database for each weld to generate an optimal template; group all welds and plan the welding sequence, and combine the optimal template to generate robot welding process planning data for controlling the welding robot; S5. Based on the results of S4, laser scanning technology is used to scan the weld, and the geometric feature parameters of the weld are extracted using artificial intelligence algorithms; the geometric feature parameters are compared with the weld distribution, and the deviation is calculated. S6. If an abnormality is detected in the geometric feature parameters, the welding path and welding process parameters are adjusted through an adaptive algorithm to complete the formation of the root pass. S7. Repeat the process from S5 to S6. That is, after each layer of welding is completed, the weld is re-scanned with laser and features are extracted. Based on the extracted feature parameters, the welding path and welding process parameters of the next layer are finely adjusted until the filling welding is completed.

2. The welding method for marine engineering steel structures facing the weld seam according to claim 1, characterized in that: In S1, the welding process data includes at least material information, welding process parameters, and planning rules for welding layers of medium and thick plates. The material information includes the steel plate material and thickness; Welding process parameters include welding current, welding voltage, welding speed, wire feed speed, welding torch oscillation amplitude, and welding torch posture; The planning rules for welding layers of medium and heavy plates include the number of welding passes for each layer, the welding sequence of each weld, the interpass temperature control range, and the welding path trajectory for each layer.

3. The welding method for marine engineering steel structures facing the weld seam according to claim 1, characterized in that: In step S2, the specific steps involved in identifying and storing the workpiece dimensions and weld distribution are as follows: Read the model file and extract the workpiece's external dimensions in three-dimensional space, denoted as... ,in Represents the length of the workpiece. Represents the workpiece width. Represents the height of the workpiece; Locate all weld joints, and then extract the core geometric properties of each weld, including weld type and theoretical weld length. Theoretical weld width Theoretical bevel angle Theoretical root gap and theoretical bevel depth ; For straight welds, their spatial position and orientation in the workpiece coordinate system are defined by extracting the starting point coordinates and the ending point coordinates; For curved welds, a non-uniform rational B-spline curve is used for fitting, and its control point coordinate sequence and node vector are stored to describe its spatial position and orientation. The core geometric properties of the weld are associated with its spatial location and orientation, generating a unique identifier for each weld. ; Identifiers for all welds The core geometric attributes, spatial location, and orientation are used as the weld distribution, and the workpiece dimensions and weld distribution are stored in a structured manner in the system database.

4. The welding method for marine engineering steel structures facing the weld seam according to claim 1, characterized in that: In step S3, the specific steps involved in matching and correcting the point cloud data with the workpiece's external dimensions and weld distribution using a coordinate registration algorithm to calculate the precise location of the weld are as follows: For raw point cloud data Preprocessing is performed using a statistical outlier removal algorithm to obtain denoised point cloud data. ; From the denoised point cloud data Extract the macroscopic geometric features of the workpiece, including the main outer contour planes, protruding edges and corners; A feature-based registration method is used to match macroscopic geometric features with the workpiece's external dimensions and core geometric properties to solve for an initial rigid body transformation matrix. This completes the initial alignment of the point cloud; Based on the initial alignment, the iterative nearest point algorithm is used for fine registration to solve for the optimal rigid body transformation matrix. ; Based on the optimal rigid body transformation matrix The spatial position and orientation in S2 are uniformly transformed to the actual workpiece coordinate system to obtain the precise position of each weld in actual space.

5. The welding method for marine engineering steel structures facing the weld seam according to claim 1, characterized in that: In step S4, the specific steps involved in matching welding process data from the welding process database to generate the optimal template for each weld seam are as follows: Based on the precise location of the weld, the spatial orientation features of the current weld are extracted, including the spatial vector of the weld in the workpiece coordinate system. Angle of inclination of the weld to the horizontal plane The normal vector of the base material surface where the weld is located ; Determine the welding difficulty coefficient based on the spatial orientation characteristics of the weld. ; By combining the core geometric properties and spatial orientation features of the weld, a multi-dimensional matching query is constructed in the welding process database; Based on the bevel angle, welding difficulty coefficient, and root gap, the matching degree between the current weld and each process template in the welding process database is calculated. ; For tilt angle Greater than the horizontal welding threshold The welding position, based on the matching process template, and according to the welding difficulty coefficient. The welding current, welding voltage, and wire feed speed are corrected to obtain the corrected welding current, welding voltage, and wire feed speed; For each weld, select the matching degree. The minimum process template, combined with the modified welding current, welding voltage, and wire feed speed, yields the optimal template for all weld seams.

6. The welding method for marine engineering steel structures facing the weld seam according to claim 5, characterized in that: In step S4, the specific steps involved in grouping all weld seams and planning the welding sequence to generate robot welding process planning data are as follows: The spatial distance between weld seams is less than a set distance threshold. Welds of the same type are grouped together and labeled with group numbers. ; Calculate the average of the start and end points of all welds in each group to obtain the coordinates of the geometric center point of that group. ; Using the starting point of the welding robot as a reference, according to the coordinates of the geometric center points of each group Sort the Euclidean distances from the starting point in ascending order to determine the welding sequence between groups. ; For each weld within each group, if its spatial vector Normal vector to the surface of the base material If the dot product is not less than 0, the welding direction is from the start point to the end point; otherwise, it is from the end point to the start point, thus obtaining the welding direction mark. ; For fillet welds and corner welds, set dwell time parameters at the start and end points of the welding path. At the same time, based on the welding difficulty coefficient Adjust the welding speed of the group containing this weld. ; Group number Inter-group welding sequence identifier Precise location of the weld and markings for the welding direction Welding speed Duration of stay The optimal template is integrated into robot welding process planning data and output.

7. The welding method for marine engineering steel structures facing the weld seam according to claim 1, characterized in that: In step S5, the specific steps involved in extracting the geometric feature parameters of the weld using artificial intelligence algorithms are as follows: Laser scanning technology is used to scan the weld seam and obtain three-dimensional point cloud data of the weld seam surface. ; 3D point cloud data Data simplification was performed using a voxel grid downsampling method, and noise points were filtered out using a statistical outlier removal algorithm to obtain preprocessed point cloud data. ; Based on the precise weld location obtained from S3, in the preprocessed point cloud data The section is centered on the theoretical centerline of the weld, with a width equal to the section width. Region of interest; Within the region of interest, a feature point detection algorithm based on curvature change is used to calculate the curvature value of each point in the point cloud, and points with curvature greater than a curvature threshold are selected. The points are marked as feature points; The weld width is obtained by calculating the average Euclidean distance between the feature points on both sides of the bevel edge. ; The bevel depth is obtained by calculating the average vertical distance from the point cloud at the bottom of the bevel to the reference plane on the surface of the base material. ; The bevel angle is obtained by fitting the bevel planes on both sides and calculating the angle between the normal vectors of the two planes. ; The root gap is obtained by calculating the average distance between the nearest neighbors on both sides of the bevel root. ; Extracted weld width Bevel depth , bevel angle and root gap With weld identifier Associated and output as geometric feature parameters.

8. A welding method for marine engineering steel structures facing the weld seam according to claim 7, characterized in that: In step S5, the geometric feature parameters are compared with the weld distribution, and the specific steps involved in calculating the deviation are as follows: The extracted geometric feature parameters are compared with the theoretical weld width in the weld distribution. Theoretical bevel angle Theoretical root gap and theoretical bevel depth Compare and calculate the width deviation, depth deviation, angle deviation, and gap deviation.

9. A welding method for marine engineering steel structures facing the weld seam according to claim 1, characterized in that: In step S6, the specific steps involved in adjusting the welding path and welding process parameters using the adaptive algorithm are as follows: Set width deviation threshold Depth deviation threshold Angle deviation threshold and gap deviation threshold ; When the width deviation exceeds the width deviation threshold At the same time, the welding current is reduced proportionally according to the difference between the width deviation and the width deviation threshold, while the welding speed is increased proportionally to obtain the welding current adjustment amount and the welding speed adjustment amount. When the depth deviation exceeds the depth deviation threshold At that time, the wire feeding speed is increased proportionally based on the difference between the depth deviation and the depth deviation threshold to obtain the wire feeding speed adjustment amount; When the angle deviation exceeds the angle deviation threshold At the same time, the welding torch posture angle is adjusted proportionally according to the difference between the angle deviation and the angle deviation threshold, and the welding path trajectory is corrected proportionally to obtain the welding torch posture angle adjustment amount and the welding path trajectory correction amount. When the gap deviation exceeds the gap deviation threshold At that time, the welding torch oscillation amplitude is increased proportionally according to the difference between the gap deviation and the gap deviation threshold, so as to obtain the welding torch oscillation amplitude adjustment amount; Set an upper limit for the adjustment of welding process parameters to ensure that the adjustment range in a single instance does not exceed the positive and negative adjustment limits of the original welding process parameters. Among them, the adjustment amounts of welding process parameters include welding current adjustment, welding speed adjustment, wire feed speed adjustment, welding torch posture angle adjustment, welding path trajectory correction, and welding torch oscillation amplitude adjustment. The adjusted welding process parameters are superimposed onto the original welding process parameters to obtain the fine-tuned welding process parameters.

10. A welding method for marine engineering steel structures facing the weld seam according to claim 1, characterized in that: In step S7, the specific steps involved in fine-tuning the welding path and welding process parameters of the next layer based on the extracted feature parameters are as follows: Obtain the actual geometric feature parameters and current welding process parameters obtained from the post-weld scan of the current layer; Calculate the deviation of each geometric feature parameter from its corresponding theoretical value; The adjustment amount of each welding process parameter is calculated based on the deviation, and the adjusted welding process parameters are obtained. Fine-tune the welding path based on the welding path trajectory correction amount in the adjustment amount of welding process parameters; The adjusted welding process parameters and welding path are then sent out to execute the next layer of welding.

Images