A straight seam welded pipe arc extinguishing plate robot intelligent assembly welding method and system

Abstract

The application discloses a straight-seam welded pipe arc striking plate robot intelligent assembly welding method and system, and relates to the technical field of straight-seam welded pipe welding. In the method, contact force field distribution data collected by a robot on a welded pipe end face is acquired; a normal force component is extracted and a normal force gradient change rate is calculated; a radial distance distribution data, an inclination angle deviation and a normal vector deflection angle deviation are determined accordingly; the radial distance distribution data is corrected based on an elastic rebound amount to obtain target data; a space pose deviation matrix is constructed according to the angle deviations and the target data, a joint angle increment sequence and a motion trajectory curve are generated through decomposition and adjustment; finally, the robot is driven to adjust the posture of the arc striking plate, and the contact force field data is updated in real time until the standard deviation of the normal force component meets a preset pose convergence threshold. The technical scheme provided by the application can improve the contact uniformity and precision of the arc striking plate and the welded pipe end face assembly, and ensure the stability of the welding quality.

Description

Technical Field

[0001] This application belongs to the field of straight seam welded pipe welding technology, specifically a robotic intelligent assembly welding method and system for straight seam welded pipe arc extinguishing plates. Background Technology

[0002] With the development of oil and gas pipeline projects towards larger diameters, higher steel grades, and longer distances, circumferential welding of straight seam welded pipes has become a critical process in pipeline connections. During circumferential welding, arc-starting and arc-extinguishing plates need to be installed at the start and end positions of the weld to ensure the weld formation quality of the arc-starting and arc-extinguishing sections, preventing defects in arc-starting and arc-extinguishing from affecting the overall performance of the welded joint. The alignment quality of the arc-starting and arc-extinguishing plates with the pipe end face directly affects the uniformity of the welding gap, the alignment accuracy, and the stability of the weld quality, and is a prerequisite for achieving high-quality circumferential welding.

[0003] Currently, robot assembly technology uses visual sensors or laser scanning equipment to acquire spatial position information of the arc-extinguishing plate and the welded pipe end face. Through image processing or point cloud data analysis, it identifies the edge features and position coordinates of the workpiece, calculates the pose deviation of the arc-extinguishing plate relative to the welded pipe end face, and then controls the robot's end effector to adjust the position and orientation of the arc-extinguishing plate based on the calculated deviation, completing the assembly operation. Some technical solutions also employ contact measuring probes to measure the gap at multiple positions in the circumferential direction, calculate the deviation based on the measurement data, and execute adjustment actions.

[0004] However, existing technologies lack effective means to assess the uniformity of the overall contact state of the contact interface after the pose adjustment is completed. In particular, when there are manufacturing tolerances, local deformations, or uneven surfaces on the welded pipe end face, the adjusted contact interface is prone to the phenomenon that some areas are too tight while other areas are not enough. Existing methods cannot accurately determine whether the current contact state has reached the optimal uniformity, making further fine adjustments difficult. This results in insufficient stability of the assembly accuracy and affects the subsequent welding quality.

[0005] Application content

[0006] To address the above problems, this application provides a robotic intelligent assembly welding method and system for straight seam welded pipe arc extinguishing plates, which solves the problem of difficulty in accurately assessing the uniformity of the contact state at the contact interface, resulting in insufficient assembly accuracy and stability. It can improve the contact uniformity and accuracy of the assembly between the arc extinguishing plate and the welded pipe end face, and ensure the stability of welding quality.

[0007] To achieve the above objectives, the technical solution adopted in this application is as follows: The contact force field distribution data is obtained, which is force information collected by the robot in multiple sampling areas in the circumferential direction of the welded pipe end face; The normal force component is extracted based on the contact force field distribution data, and the normal force gradient change rate between adjacent sampling areas is calculated. Based on the normal force gradient change rate, the centroid coordinates of the over-compression area and the centroid coordinates of the local suspended area are determined. Based on the centroid coordinates of the over-compressed region and the centroid coordinates of the locally suspended region, determine the radial distance distribution data, tilt angle deviation, and normal vector deflection angle deviation; Based on the radial distance distribution data, the elastic rebound amount is predicted to obtain the predicted rebound displacement, and the radial distance distribution data is corrected using the predicted rebound displacement to obtain the target radial distance distribution data. A spatial pose deviation matrix is ​​constructed based on the tilt angle deviation, the normal vector deflection angle deviation, and the target radial distance distribution data. The spatial pose deviation matrix is ​​then decomposed and its gain is adjusted to generate a joint angle increment sequence. The motion trajectory curve is generated based on the joint angle increment sequence, and the robot is driven to adjust the attitude of the extinguishing plate based on the motion trajectory curve. During the adjustment process, the contact force field distribution data is updated in real time until the standard deviation of the normal force component corresponding to the updated contact force field distribution data is less than or equal to the preset pose convergence criterion threshold.

[0008] By employing the above technical solution, contact force field distribution data is acquired from multiple sampling areas along the circumference of the welded pipe end face, transforming the contact state of the contact interface into quantifiable force information. By extracting the normal force component and calculating the rate of change of the normal force gradient between adjacent sampling areas, abrupt changes in the normal force distribution at the contact interface are identified. Based on the rate of change of the normal force gradient, the centroid coordinates of the over-compression area and the local suspended area are determined, enabling spatial localization of areas with excessively tight or insufficient contact. Radial distance distribution data, tilt angle deviation, and normal vector deflection angle deviation are determined based on the centroid coordinates of the over-compression area and the local suspended area, resolving the uneven contact phenomenon into specific spatial pose deviation parameters. The predicted springback displacement is obtained by predicting the elastic rebound amount based on the radial distance distribution data, and the predicted springback displacement is used to correct the radial distance distribution data to obtain the target radial distance distribution data. This compensates for the springback effect caused by the elastic deformation of the material during the contact adjustment process between the arc extinguishing plate and the welded pipe end face, avoiding over-adjustment or under-adjustment problems caused by not considering elastic rebound. A spatial pose deviation matrix is ​​constructed based on tilt angle deviation, normal vector deflection angle deviation, and target radial distance distribution data. This matrix is ​​then decomposed and its gain adjusted to generate a joint angle increment sequence, which is converted into control commands for each robot joint. Motion trajectory curves generated from the joint angle increment sequence drive the robot to adjust the attitude of the arc-extinguishing plate. During the adjustment process, the contact force field distribution data is updated in real time until the standard deviation of the normal force component corresponding to the updated contact force field distribution data is less than or equal to a preset pose convergence criterion threshold. This forms an iterative optimization closed-loop control mechanism based on force feedback, making the normal force distribution at the contact interface tend towards a uniform state. This solution solves the problem in existing technologies where the uniformity of the overall contact state at the contact interface is difficult to accurately assess and finely adjust when there are manufacturing tolerances, local deformations, or surface unevenness on the welded pipe end face. It improves the contact uniformity and precision stability of the assembly between the arc-extinguishing plate and the welded pipe end face, ensuring the quality of subsequent circumferential welding.

[0009] In a second aspect, embodiments of this application provide a robotic intelligent assembly welding system for straight seam welded pipe arc extinguishing plates. The robotic intelligent assembly welding system for straight seam welded pipe arc extinguishing plates includes: one or more processors and a memory; the memory is coupled to the one or more processors and is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the robotic intelligent assembly welding system for straight seam welded pipe arc extinguishing plates to perform the method described in the first aspect and any possible implementation thereof.

[0010] In summary, one or more technical solutions provided in this application have at least the following technical effects or advantages: By acquiring contact force field distribution data, a comprehensive perception of the overall stress state of the contact interface between the arc extinguishing plate and the welded pipe end face is achieved, overcoming the limitation of existing technologies in effectively assessing the uniformity of the contact state. Based on this, by extracting the normal force component and calculating its normal force gradient change rate between adjacent sampling areas, the centroid coordinates of the over-extrusion area and the centroid coordinates of the locally suspended area can be accurately determined, thereby deriving radial distance distribution data, tilt angle deviation, and normal vector deflection angle deviation. This process accurately quantifies the complex contact unevenness caused by manufacturing tolerances or local deformation under actual working conditions into multi-dimensional spatial geometric deviation parameters, providing an accurate basis for fine adjustment. Furthermore, this application predicts and corrects the elastic rebound amount based on the radial distance distribution data to obtain the target radial distance distribution data, effectively overcoming the elastic rebound error caused by workpiece compression and improving the reliability of the target parameters. Subsequently, by constructing and decomposing the spatial pose deviation matrix, and combining gain adjustment to generate a joint angle increment sequence and motion trajectory curve, the robot is smoothly driven to adjust the attitude of the arc extinguishing plate. During the adjustment process, a dynamic closed-loop feedback is formed by updating the contact force field distribution data in real time until the standard deviation of the normal force component meets the preset pose convergence criterion threshold. This closed-loop control mechanism not only realizes the adaptive and fine adjustment of the pose of the arc extinguishing plate, but also ensures that the stress state of the final contact interface reaches optimal uniformity. It solves the problem of insufficient assembly accuracy stability in the prior art, improves the contact uniformity and alignment accuracy of the assembly between the arc extinguishing plate and the welded pipe end face, and thus provides a guarantee for high-quality circumferential welding. Attached Figure Description

[0011] Figure 1 This is a schematic flowchart of a robotic intelligent assembly welding method for straight seam welded pipes with arc-extinguishing plates, as disclosed in an embodiment of this application. Figure 2 This is another schematic diagram of the process of a robotic intelligent assembly welding method for straight seam welded pipes with arc extinguishing plates disclosed in the embodiments of this application; Figure 3 This is a structural schematic diagram of a robotic intelligent assembly welding system for straight seam welded pipe arc extinguishing plates provided in an embodiment of this application.

[0012] In the diagram: 301, Central Processing Unit; 302, Read-Only Memory; 303, Random Access Memory; 304, Bus; 305, Input / Output Interface; 306, Input Section; 307, Output Section; 308, Storage Section; 309, Communication Section; 310, Driver; 311, Removable Media. Detailed Implementation

[0013] To enable those skilled in the art to better understand the technical solution, the present application will be described in detail below with reference to the embodiments. The description in this section is only exemplary and explanatory, and should not be used to limit the scope of protection of the present application in any way.

[0014] It should be noted that, in this document, the terms "comprising," "including," and any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Specific examples have been used in this document to illustrate the principles and implementation methods of the technical solutions of this application. The above examples are only for the purpose of helping to understand the methods and core ideas of this application. The above are merely preferred embodiments of this application. It should be pointed out that, due to the limitations of written expression and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes, or combinations, or the direct application of the concept and technical solutions of this application to other situations without modification, should all be considered within the scope of protection of this application.

[0015] This application provides a robotic intelligent assembly welding method for straight seam welded pipe arc-starting and extinguishing plates, referring to... Figure 1 , Figure 1 This is a flowchart illustrating a robotic intelligent assembly welding method for straight seam welded pipes using an arc-extinguishing plate, as provided in an embodiment of this application. The method is applied to a system, which refers to a hardware and software integrated platform capable of executing a robotic intelligent assembly welding program for straight seam welded pipes using an arc-extinguishing plate. The system can execute a robotic intelligent assembly welding program for straight seam welded pipes using an arc-extinguishing plate. The method includes steps 101 to 106, as follows: Step 101: Obtain contact force field distribution data. The contact force field distribution data is force information collected by the robot in multiple sampling areas along the circumference of the welded pipe end face.

[0016] In this embodiment, the contact force field distribution data refers to the collection of force measurement information sensed at various sampling positions distributed along the circumferential direction during the contact between the arc extinguishing plate and the welded pipe end face. The welded pipe end face represents the axial cross-section of the pipe to be welded, which has a circular annular structure. The circumferential direction is used to represent the circumferential angle range along the annular contour of the welded pipe end face, covering the complete angle range from zero degrees to 360 degrees. The sampling area refers to multiple discrete measurement positions divided along the circumferential direction according to a preset angular interval. For example, if a sampling area is set every 30 degrees, then twelve sampling areas are set along the circumferential direction. The force information represents the three-dimensional force vector and three-dimensional torque vector data measured by the force sensor in each sampling area, including normal force, tangential force, and corresponding torque components.

[0017] Specifically, a force sensor array performs distributed measurements along the circumference of the welded pipe end face. When the robot, holding the extinguishing plate, approaches and initially contacts the welded pipe end face, the force sensor's data acquisition program is initiated. Measurement points are set at equal angular intervals along the circumference, dividing the complete circumference into several sampling areas, each corresponding to a force sensor measurement unit. The sensors synchronously collect the current mechanical state information in each sampling area, acquiring six-degree-of-freedom force data including normal force components, tangential force components, and torque components. The collected force data from each sampling area are marked and stored according to circumferential angular positions, establishing a mapping relationship between circumferential angles and force data. The force data from all sampling areas are aggregated to form a complete contact force field distribution dataset. This dataset reflects the overall mechanical distribution characteristics of the extinguishing plate and the welded pipe end face under the current contact state, providing fundamental data support for subsequent deviation identification and adjustment control.

[0018] Step 102: Extract the normal force component based on the contact force field distribution data and calculate the normal force gradient change rate between adjacent sampling areas. Determine the centroid coordinates of the over-compressed area and the local suspended area based on the normal force gradient change rate.

[0019] In this embodiment, the normal force component refers to the mechanical component perpendicular to the surface of the welded pipe end face in the contact force field distribution data. This component reflects the magnitude of the normal pressure or tension exerted by the arc-extinguishing plate on the welded pipe end face. The normal force gradient change rate is used to represent the ratio of the difference in normal force components between two adjacent sampling areas to the circumferential angle interval between the two areas, reflecting the rate of change of the normal force along the circumferential direction. The centroid coordinates of the excessively compressed area represent the weighted geometric center coordinates of a continuous sampling area with a positive and large normal force gradient change rate. This area corresponds to the location where the arc-extinguishing plate applies excessive pressure to the welded pipe end face. The centroid coordinates of the locally suspended area refer to the weighted geometric center coordinates of a continuous sampling area with a negative and large absolute value of the normal force gradient change rate. This area corresponds to the location where there is a gap or insufficient contact force between the arc-extinguishing plate and the welded pipe end face.

[0020] Specifically, the normal force components of each sampling region are extracted from the contact force field distribution data and arranged in circumferential angle order to form a normal force sequence. Adjacent element difference operations are performed on this sequence to calculate the difference in normal force components between the i-th sampling region and the i-th plus one sampling region. This difference is divided by the angular interval between the two sampling regions to obtain the normal force gradient rate of change for the i-th interval. This process is repeated for all adjacent intervals to obtain a complete normal force gradient rate of change sequence. A positive gradient threshold is set to identify continuous intervals with gradient rates of change greater than this threshold, and these intervals are marked as over-compression regions. The circumferential angle and radial position of each sampling point within the over-compression region are extracted, and a weighted average coordinate is calculated using the absolute value of the gradient rate of change at each point as the weight, serving as the centroid coordinates of the over-compression region. A negative gradient threshold is set to identify continuous intervals with gradient rates of change less than this threshold, and these are marked as locally suspended regions. The same weighted calculation method is used to determine the centroid coordinates of these locally suspended regions.

[0021] In one possible implementation, the normal force component is extracted based on the contact force field distribution data, and the rate of change of the normal force gradient between adjacent sampling areas is calculated. The centroid coordinates of the over-compressed region and the centroid coordinates of the locally suspended region are determined based on the rate of change of the normal force gradient. Specifically, this includes steps 1021-1024, as follows: Step 1021: According to the arrangement order of each sampling area in the circumferential direction of the welded pipe end face, calculate the ratio of the difference of the normal force component between two adjacent sampling areas to the corresponding arc length, and obtain the normal force gradient change rate between each adjacent sampling area.

[0022] In this embodiment, the arrangement order refers to the spatial distribution order of each sampling area along the circumference of the welded pipe end face from zero degrees to 360 degrees. The difference in normal force components represents the numerical difference between the normal force values ​​measured in two adjacent sampling areas. For example, when the normal force in the third sampling area is twelve Newtons and the normal force in the fourth sampling area is ten Newtons, the difference is negative two Newtons. The corresponding arc length is used to represent the arc distance between two adjacent sampling areas in the circumferential direction, which is calculated by the product of the circumferential radius and the central angle of the two sampling areas. The rate of change of normal force gradient refers to the ratio of the difference in normal force components to the corresponding arc length. This ratio reflects the rate of change of normal force along the circumferential direction, and the unit is Newtons per millimeter.

[0023] Specifically, the circumferential angle markers of each sampling region in the contact force field distribution data are read, and the sampling regions are sorted in ascending order of angle. The normal force component values ​​of each sampling region after sorting are extracted to form a normal force sequence. Starting from the first element of the sequence, the normal force component of the i-th plus one sampling region is calculated by subtracting the normal force component of the i-th sampling region from that of the i-th sampling region, yielding the normal force difference between the i-th and i+1 sampling regions. The radius parameter of the welded pipe end face is obtained, and the circumferential angle difference between the i-th and i+1 sampling regions is calculated. The angle difference is converted from degrees to radians and multiplied by the radius to obtain the corresponding arc length. The normal force difference is divided by the corresponding arc length to calculate the normal force gradient rate of change between the i-th and i+1 sampling regions. The above calculation is repeated for all adjacent sampling region pairs, with the gradient also calculated for the last sampling region forming a closed connection with the first sampling region. All calculated normal force gradient rates of change are stored in the order of the sampling region pairs to form a complete normal force gradient rate of change dataset.

[0024] Step 1022: The sampling region where the rate of change of normal force gradient is greater than or equal to the preset extrusion gradient threshold is determined as the over-extrusion candidate region, and the sampling region where the rate of change of normal force gradient is less than the preset dangling gradient threshold is determined as the local dangling candidate region.

[0025] In this embodiment, the preset extrusion gradient threshold refers to a critical value for the rate of change of the normal force gradient used to determine whether an excessive extrusion state exists in the sampling area. When the rate of change of the gradient is greater than or equal to this threshold, it is determined that the extrusion force has increased abnormally. For example, the threshold is set to 0.5 Newtons per millimeter. The excessive extrusion candidate region represents the set of sampling regions where the rate of change of the normal force gradient meets the extrusion determination condition. These regions correspond to the locations where the ignition and extinguishing plate applies excessive pressure to the end face of the welded pipe. The preset suspending gradient threshold is used to represent a critical value for the rate of change of the normal force gradient used to determine whether a local suspending state exists in the sampling area. This threshold is a negative value, for example, -0.4 Newtons per millimeter. The local suspending candidate region refers to the set of sampling regions where the rate of change of the normal force gradient is less than the suspending determination threshold. These regions correspond to the locations where the ignition and extinguishing plate does not make sufficient contact with the end face of the welded pipe.

[0026] Specifically, the preset compression gradient threshold parameter and the preset dangling gradient threshold parameter are read from memory. Each numerical element in the normal force gradient change rate dataset is traversed, and the current gradient change rate value is compared with the preset compression gradient threshold. If the current gradient change rate value is greater than or equal to the preset compression gradient threshold, the sampling region corresponding to that gradient is marked as an over-compression candidate region, and the circumferential angle, radial coordinate, and gradient value of that sampling region are recorded. The current gradient change rate value is compared with the preset dangling gradient threshold. If the current gradient change rate value is less than the preset dangling gradient threshold, the sampling region corresponding to that gradient is marked as a local dangling candidate region, and the location and gradient information of that region are also recorded. The information of all sampling regions marked as over-compression candidate regions is summarized in the over-compression candidate region data table, and the information of all sampling regions marked as local dangling candidate regions is summarized in the local dangling candidate region data table, providing a data foundation for subsequent centroid coordinate calculation.

[0027] Step 1023: Using the absolute value of the rate of change of the normal force gradient corresponding to each over-compression candidate region as the weight, perform a weighted calculation on the centroid coordinates of each over-compression candidate region to obtain the centroid coordinates of the over-compression region; using the absolute value of the rate of change of the normal force gradient corresponding to each locally suspended candidate region as the weight, perform a weighted calculation on the centroid coordinates of each locally suspended candidate region to obtain the centroid coordinates of the locally suspended region.

[0028] In this embodiment, the absolute value of the rate of change of the normal force gradient represents the non-negative value of the gradient rate of change after removing the positive and negative signs. It is used to measure the severity of the change in mechanical state without considering the direction of change. Weight refers to the contribution coefficient of each data point to the final result during the weighted calculation process; the larger the absolute value of the gradient, the more significant the influence of the region on the centroid position. Centroid coordinates represent the geometric center position of the candidate region in space, expressed using a Cartesian coordinate system or a cylindrical coordinate system. Weighted calculation refers to the mathematical operation of multiplying the coordinate values ​​of each candidate region by its corresponding weight, summing the results, and then dividing by the total weight. The centroid coordinates of the over-compressed region represent the comprehensive geometric center position obtained after weighted calculation of all over-compressed candidate regions. The centroid coordinates of the locally suspended region represent the comprehensive geometric center position obtained after weighted calculation of all locally suspended candidate regions.

[0029] Specifically, the data table of over-compression candidate regions is used to read the normal force gradient change rate values ​​for each region, and the absolute value of each value is calculated. The calculated absolute value sequence is used as a weight sequence, and the sum of all elements in the weight sequence is calculated. The centroid coordinates of each over-compression candidate region are extracted, including X-axis, Y-axis, and Z-axis coordinate components. For the X-axis direction, the X-coordinate value of each candidate region is multiplied by its corresponding weight, and all products are summed and divided by the total weights to obtain the weighted X-coordinate. The same calculation method is used to obtain the weighted Y-coordinate and weighted Z-coordinate, and the three coordinates are combined to form the centroid coordinates of the over-compression region. The data table of locally suspended candidate regions is read, and the absolute value of the gradient change rate for each region is used as the weight. The centroid coordinate components of each locally suspended candidate region are extracted, and a weighted average is calculated in the X, Y, and Z directions to obtain the centroid coordinates of the locally suspended region. The two centroid coordinates are stored as input data for subsequent deviation calculations.

[0030] Step 103: Based on the centroid coordinates of the over-compressed area and the centroid coordinates of the locally suspended area, determine the radial distance distribution data, tilt angle deviation, and normal vector deflection angle deviation.

[0031] In this embodiment, the radial distance distribution data refers to the set of radial distance measurements from the center of the welded pipe end face to each centroid position, reflecting the spatial distribution characteristics of the over-extended area and the locally suspended area in the radial direction. The tilt angle deviation represents the tilt angle value of the ignition plate mounting plane relative to the ideal plane of the welded pipe end face. It is calculated through the spatial geometric relationship between the centroids of the over-extended area and the locally suspended area. For example, when the two centroids are 180 degrees apart in the circumferential direction and the difference in normal force is 15 Newtons, the corresponding tilt angle is 0.3 degrees. The normal vector deflection angle deviation represents the deflection angle value of the normal vector direction of the ignition plate surface relative to the normal vector direction of the welded pipe end face. This deviation reflects the degree of directional deviation of the ignition plate attitude in three-dimensional space.

[0032] Specifically, a cylindrical coordinate system is established with the center of the welded pipe end face as the origin, and the coordinates of the centroids of the over-extended region and the locally suspended region are converted into cylindrical coordinate representations. The radial distances of the two centroids relative to the center of the circle are calculated and recorded as the radial distance of the over-extended region and the radial distance of the locally suspended region, respectively. These two values ​​are combined to form radial distance distribution data. The angular difference between the two centroids in the circumferential direction is calculated. Combined with the difference in the normal force components at the two centroids, the tilt angle of the extinguishing plate relative to the welded pipe end face is calculated according to the geometric projection relationship. The difference between the tilt angle and the ideal zero tilt state is determined as the tilt angle deviation. The tangential force components and torque components at the two centroids are extracted. The direction vector of the surface normal vector of the extinguishing plate in three-dimensional space is calculated through the torque balance equation. The angle between this direction vector and the ideal normal vector of the welded pipe end face is calculated, and this angle is determined as the normal vector deflection angle deviation.

[0033] Step 104: Based on the radial distance distribution data, predict the elastic rebound amount to obtain the predicted rebound displacement, and use the predicted rebound displacement to correct the radial distance distribution data to obtain the target radial distance distribution data.

[0034] In this embodiment, elastic rebound prediction refers to calculating the elastic deformation of the extinguishing plate and the welded pipe end face under contact pressure based on the principles of material mechanics, and predicting the displacement change caused by the elastic recovery of the material after the pressure is released. The predicted rebound displacement represents the radial displacement value that the extinguishing plate will experience after the external force is removed following the adjustment, calculated using elastic rebound prediction. For example, the predicted rebound displacement is 0.08 mm when the radial distance is 0.5 mm and the contact stiffness is 10 Newtons per millimeter. The target radial distance distribution data represents the set of radial distance values ​​after elastic rebound compensation correction. This dataset considers the material elastic recovery effect and represents the actual radial position distribution that needs to be achieved during adjustment.

[0035] Specifically, the radial penetration depth values ​​of the over-compression region and the radial clearance values ​​of the locally suspended region are extracted from the radial distance distribution data. Based on the material properties of the extinguishing plate and the welded pipe end face, including elastic modulus, Poisson's ratio, and material thickness, a correspondence between contact stiffness and radial displacement is established. The radial penetration depth value is multiplied by the contact stiffness coefficient to obtain the elastic restoring force value generated by the current contact pressure. The elastic restoring force value is divided by the product of the material's equivalent elastic modulus and thickness coefficient to calculate the predicted springback displacement. Each radial distance value in the radial distance distribution data is corrected: the radial penetration depth of the over-compression region is increased by the predicted springback displacement, and the radial clearance of the locally suspended region is decreased by the predicted springback displacement, resulting in corrected radial distance values. All corrected radial distance values ​​are recombined to form the target radial distance distribution data, which serves as a reference for the target position of subsequent robot adjustment and control.

[0036] In one possible implementation, the elastic rebound amount is predicted based on the radial distance distribution data to obtain the predicted rebound displacement, and the radial distance distribution data is corrected using the predicted rebound displacement to obtain the target radial distance distribution data. Specifically, this includes steps 1041-1043, as follows: Step 1041: Obtain the preset elastic modulus parameters of the arc extinguishing plate material and the equivalent contact stiffness of the end face of the welded pipe.

[0037] In this embodiment, the elastic modulus parameter of the arc-extinguishing plate material refers to the ratio of stress to strain when the arc-extinguishing plate material is subjected to force, reflecting the material's ability to resist elastic deformation. The unit is gigapascal; for example, the elastic modulus of an aluminum alloy arc-extinguishing plate is seventy gigapascals. Equivalent contact stiffness represents the force required to achieve a unit deformation of the welded pipe end face under normal pressure, taking into account both the welded pipe material properties and the end face geometry. The unit is newtons per millimeter. Preset refers to parameter values ​​that are pre-set and stored during the system configuration phase based on material specifications and experimental measurement data. Acquisition refers to the process of reading and loading specified parameter values ​​from the system parameter database or configuration file.

[0038] Specifically, the system accesses the material property configuration table in the system parameter database and locates the corresponding material parameter record entry based on the material model identifier of the arc-extinguishing plate. The elastic modulus parameter value of the arc-extinguishing plate material is read from this record entry, converted from its storage format to the required calculation type, and unit conversion is performed to ensure the unit is uniformly Pascal. The welded pipe parameter configuration table is accessed, and the corresponding contact mechanics parameters are found based on the currently operating welded pipe specification. The equivalent contact stiffness value of the welded pipe end face is extracted from the contact mechanics parameters; this value is typically calculated based on the elastic modulus of the welded pipe material, wall thickness, and end face geometry. The read elastic modulus parameter and equivalent contact stiffness value are validated to ensure their range is within a reasonable physical range. The validated elastic modulus parameter and equivalent contact stiffness value of the arc-extinguishing plate material are stored in the current task's runtime memory variable for subsequent springback displacement calculation.

[0039] Step 1042: Calculate the product of the radial penetration depth of the over-compression region and the equivalent contact stiffness in the radial distance distribution data to obtain the elastic restoring force; calculate the ratio of the elastic restoring force to the elastic modulus parameter of the arc extinguishing plate material to obtain the predicted rebound displacement of the over-compression region after the pressure is released.

[0040] In this embodiment, the radial penetration depth of the over-extrusion region refers to the depth of the arc-extinguishing plate penetrating the ideal plane of the welded pipe end face at the centroid of the over-extrusion region in the radial distance distribution data, expressed in millimeters, for example, 0.6 millimeters. Elastic restoring force represents the restoring force generated by the elastic deformation of the material, calculated by multiplying the penetration depth by the equivalent contact stiffness, expressed in Newtons. A ratio is the result of dividing two physical quantities, used to establish a quantitative relationship between different physical quantities. The predicted springback displacement represents the radial springback distance caused by the elastic recovery of the material after the pressure is completely released in the over-extrusion region, expressed in millimeters. Pressure release refers to the process of gradually reducing the normal pressure applied by the arc-extinguishing plate to the welded pipe end face to zero.

[0041] Specifically, the data items corresponding to the over-compression region are identified from the radial distance distribution data, and the radial penetration depth value of this region is extracted. The extracted radial penetration depth value is multiplied by the equivalent contact stiffness value, and the product is calculated. The product is used as the elastic restoring force value, which represents the magnitude of the restoring force corresponding to the elastic potential energy accumulated inside the material under the current deformation state. The calculated elastic restoring force value is divided by the elastic modulus parameter of the arc extinguishing plate material, and the ratio of elastic restoring force to elastic modulus parameter is calculated. According to the material mechanics relationship, this ratio is multiplied by the correction coefficient of material thickness and contact area to obtain the predicted rebound displacement of the over-compression region after the pressure is released. The predicted rebound displacement value is stored as the input parameter for radial distance correction calculation. This predicted rebound displacement reflects the actual position change caused by the elastic recovery of the material after the arc extinguishing plate is adjusted and the external force is removed, providing a basis for subsequent position compensation.

[0042] Step 1043: Add the predicted rebound displacement to the radial penetration depth of the over-compression area in the radial distance distribution data, and update the radial distance distribution data based on the addition result to obtain the target radial distance distribution data.

[0043] In this embodiment, "addition" refers to the mathematical operation of summing two values. The summation result refers to the sum of the predicted rebound displacement and the radial penetration depth of the over-compression region. "Update" represents the data modification operation of replacing the corresponding value in the original data item with the newly calculated value. The target radial distance distribution data refers to the radial distance distribution data after elastic rebound compensation correction. This data takes into account the material's elastic recovery effect and reflects the actual radial position distribution that needs to be achieved during adjustment. The radial penetration depth of the over-compression region in the radial distance distribution data represents the penetration depth value of the data item marked as the over-compression region in the original radial distance data set.

[0044] Specifically, the predicted rebound displacement value and the radial penetration depth value of the over-compression region in the radial distance distribution data are read. The predicted rebound displacement value and the radial penetration depth value are added together to calculate the sum. The sign of the sum is determined: a positive sum indicates that radial deviation still exists after correction, while a sum close to zero indicates that the ideal position has been achieved after correction. The data storage location corresponding to the over-compression region is located in the radial distance distribution data structure, and the original radial penetration depth value at that location is replaced with the sum value. Data items in other regions of the radial distance distribution data are retained, and only the values ​​of the over-compression region are modified. The radial distance distribution data after numerical update is renamed or marked as the target radial distance distribution data. The integrity of the target radial distance distribution data is checked to confirm that the radial distance values ​​of all sampled regions have been correctly updated and that the value range is reasonable. The target radial distance distribution data is output to the subsequent spatial pose deviation matrix construction process as target reference data for pose adjustment.

[0045] Step 105: Construct a spatial pose deviation matrix based on tilt angle deviation, normal vector deflection angle deviation, and target radial distance distribution data. Decompose and adjust the gain of the spatial pose deviation matrix to generate a joint angle increment sequence.

[0046] In this embodiment, the spatial pose deviation matrix refers to a six-dimensional matrix containing the deviation values ​​of the extinguishing arc plate relative to the welded pipe end face in three translational degrees of freedom and three rotational degrees of freedom. Decomposition and gain adjustment represent the process of performing coordinate transformation and amplitude adjustment on the spatial pose deviation matrix, including decomposing the six-dimensional pose deviation into components in the coordinate systems of each robot joint, and adjusting the gain coefficients of each component according to the system's dynamic response characteristics. The joint angle increment sequence represents the ordered arrangement of the angle values ​​that each rotary joint of the robot needs to change to achieve the target pose adjustment; for example, the joint angle increment sequence of a six-joint robot contains six angle increment values.

[0047] Specifically, the tilt angle deviation, normal vector deflection angle deviation, and target radial distance distribution data are converted into a six-dimensional pose deviation vector in the robot's end effector coordinate system. The tilt angle deviation is decomposed into rotational components around the X and Y axes, the normal vector deflection angle deviation is decomposed into a rotational component around the Z axis, and the target radial distance distribution data is converted into translational components along the X, Y, and Z axes. The six deviation components are arranged in the order of the three translational components and the three rotational components to construct a six-dimensional spatial pose deviation matrix. The spatial pose deviation matrix is ​​mapped to the joint space using the robot's inverse kinematics Jacobian matrix, and the required changes in each joint angle are calculated. The calculated changes in joint angles are then adjusted by gain, with different gain coefficients set based on the differences in inertia and response speed of each joint. Each joint angle change is multiplied by the corresponding gain coefficient. The adjusted joint angle changes are arranged in order of joint number to generate a joint angle increment sequence.

[0048] In one possible implementation, a spatial pose deviation matrix is ​​constructed based on tilt angle deviation, normal vector deflection angle deviation, and target radial distance distribution data. The spatial pose deviation matrix is ​​then decomposed and its gain adjusted to generate a joint angle increment sequence. Specifically, this includes steps 1051-1054, as follows: Step 1051: Convert the tilt angle deviation, normal vector deflection angle deviation, and target radial distance distribution data into spatial translation component deviation values ​​and spatial rotation component deviation values, and determine the spatial pose deviation matrix based on the spatial translation component deviation values ​​and spatial rotation component deviation values.

[0049] In this embodiment, the spatial translation component deviation value refers to the positional deviation of the arc-extinguishing plate relative to the weld pipe end face in the three directions of the Cartesian coordinate system (X-axis, Y-axis, Z-axis), measured in millimeters. The spatial rotation component deviation value represents the rotational angle deviation of the arc-extinguishing plate relative to the weld pipe end face around the three coordinate axes, including the pitch angle deviation around the X-axis, the roll angle deviation around the Y-axis, and the yaw angle deviation around the Z-axis, measured in degrees or radians. "Transformation" represents the calculation process of converting the angular deviation and radial distance data into translational and rotational components through coordinate transformation and geometric decomposition. The spatial pose deviation matrix refers to a six-dimensional vector or matrix data structure containing three translational components and three rotational components. The target radial distance distribution data represents the target radial distance distribution data after elastic rebound compensation correction.

[0050] Specifically, the tilt angle deviation, normal vector deflection angle deviation, and target radial distance distribution data are read. The tilt angle deviation is decomposed into two rotational components around the X-axis and Y-axis, and the values ​​of each rotational component are calculated through geometric projection relationships. The normal vector deflection angle deviation is determined as the rotational component around the Z-axis. The radial coordinates of the centroids of the over-compressed region and the locally suspended region are extracted from the target radial distance distribution data, and the difference in the radial coordinates of the two centroids is converted into X-axis and Y-axis translational components in Cartesian coordinates. The height difference between the two centroids in the Z-axis direction is calculated to determine the Z-axis translational component. The X-axis translational component, Y-axis translational component, and Z-axis translational component are summarized to form a spatial translational component deviation value vector. The rotational components around the X-axis, Y-axis, and Z-axis are summarized to form a spatial rotational component deviation value vector. The six components are combined to construct a six-dimensional spatial pose deviation matrix in the order of translational components first and rotational components last. The units of the spatial pose deviation matrix are standardized to ensure that the unit of translational components is meters and the unit of rotational components is radians.

[0051] Step 1052: Perform orthogonal decomposition on each component in the spatial pose deviation matrix to obtain the position deviation decomposition and force deviation decomposition.

[0052] In this embodiment, orthogonal decomposition refers to the independent projection decomposition of each component in the spatial pose deviation matrix into the position control space and the force control space. The position deviation decomposition quantity represents the component of the spatial pose deviation matrix in the position control dimension, reflecting the deviation value that needs to be adjusted in pure geometric position, including the difference between the expected position value and the current value for translation and rotation. The force deviation decomposition quantity represents the component of the spatial pose deviation matrix in the force control dimension, reflecting the difference between the target value of the contact force and the current measured value. Each component refers to the numerical element in the six dimensions of the spatial pose deviation matrix. The operation that yields the output data after calculation and processing is then performed.

[0053] Specifically, the six dimensions of the spatial pose deviation matrix are read and divided into translational and rotational components. Orthogonal basis vectors for the position control space and force control space are established, with the two sets of basis vectors perpendicular to each other to form a complete six-dimensional spatial basis. The translational components of the spatial pose deviation matrix are projected onto the translational basis vectors of the position control space, and the projection coefficients are calculated as the translational part of the position deviation decomposition. The rotational components of the spatial pose deviation matrix are projected onto the rotational basis vectors of the position control space, and the projection coefficients are calculated as the rotational part of the position deviation decomposition. The translational components are projected onto the basis vectors of the force control space, and the force deviation decomposition is calculated based on the normal force information in the contact force field distribution data. The rotational components are converted into equivalent torque deviations using the torque balance equation, serving as the rotational part of the force deviation decomposition. The translational and rotational parts of the position deviation decomposition are combined to form a complete position deviation decomposition vector, and the various parts of the force deviation decomposition are combined to form a complete force deviation decomposition vector. The orthogonality of the two decomposition vectors was verified, confirming that the inner product of the position deviation decomposition and the force deviation decomposition is zero.

[0054] Step 1053: Calculate the contact stiffness parameters between the ignition plate and the end of the welded pipe based on the contact force field distribution data, and dynamically adjust the proportional coefficient between the position loop gain and the force feedback closed loop gain based on the contact stiffness parameters.

[0055] In this embodiment, the contact stiffness parameter refers to the comprehensive stiffness characteristic value at the contact interface between the arc-extinguishing plate and the welded pipe end face, calculated through the force-displacement relationship in the contact force field distribution data, with units of Newtons per millimeter. The position loop gain represents the amplification factor from the position error to the control output in the position control loop; a larger value results in faster position adjustment but reduced stability. The force feedback closed-loop gain represents the amplification factor from the force error to the control output in the force control loop; a larger value results in faster force control response but is prone to oscillation. The proportional coefficient is the relative ratio between the position loop gain and the force feedback closed-loop gain, determining the weight distribution of position control and force control in the hybrid control. Dynamic adjustment refers to a control strategy that adaptively changes the proportional coefficient value based on real-time measured changes in the contact stiffness parameter.

[0056] Specifically, the normal force component and corresponding radial displacement data for each sampling region are extracted from the contact force field distribution data. A linear regression is performed on the normal force component and radial displacement data, and the slope of the fitted line is used as a preliminary estimate of the contact stiffness parameter. The estimated contact stiffness values ​​for each sampling region are weighted and averaged, with the weights determined by the contact area ratio of each region, to obtain the global contact stiffness parameter. According to the preset control strategy formula, the contact stiffness parameter is substituted into the target proportionality coefficient for calculating the position loop gain and the force feedback closed-loop gain. The position loop gain value in the current system configuration is read, and this value is divided by the target proportionality coefficient to calculate the adjusted force feedback closed-loop gain value. If the contact stiffness parameter increases, indicating tighter contact, the weight of the force feedback closed-loop gain is increased to improve force control accuracy. If the contact stiffness parameter decreases, indicating looser contact, the weight of the position loop gain is increased to prioritize position tracking performance. The adjusted position loop gain and force feedback closed-loop gain values ​​are updated in the controller parameter table to provide gain parameters for subsequent joint angle increment sequence calculations.

[0057] Step 1054: Based on the position loop gain, force feedback closed loop gain, position deviation decomposition, and force deviation decomposition after adjusting the proportional coefficient, the joint angle increment sequence of multiple joint axes of the robot is calculated through the mapping relationship matrix between the robot end pose and joint angles.

[0058] In this embodiment, the position loop gain after adjusting the scaling factor refers to the new position loop gain value obtained through dynamic adjustment of the contact stiffness parameter. The force feedback closed-loop gain after adjusting the scaling factor represents the new force feedback closed-loop gain value obtained through dynamic adjustment of the contact stiffness parameter. The mapping relationship matrix between the robot end-effector pose and joint angles refers to the robot inverse kinematics Jacobian matrix, establishing a linear mapping relationship between the six-dimensional pose change of the end-effector and the changes in each joint angle. The joint angle increment sequence is used to represent an ordered array of angle values ​​that need to be changed by each joint of the robot to achieve the target pose adjustment, arranged by joint number. Multiple joint axes represent multiple rotational or translational joints of the robot; for example, a six-degree-of-freedom robot contains six joint axes.

[0059] Specifically, the adjusted position loop gain and force feedback closed-loop gain values ​​are read. The position deviation decomposition is multiplied by the position loop gain to obtain the position control contribution. The force deviation decomposition is multiplied by the force feedback closed-loop gain to obtain the force control contribution. The position control contribution and the force control contribution are vector-added to obtain the end effector pose adjustment vector under hybrid control. This vector contains six components, corresponding to the values ​​that the end effector needs to adjust in three translational and three rotational directions. The Jacobian matrix under the current robot configuration is obtained; this matrix has a dimension of six rows multiplied by the number of joints. The Jacobian matrix is ​​inverted or pseudo-inverted to obtain the inverse Jacobian matrix. The end effector pose adjustment vector is multiplied by the inverse Jacobian matrix to obtain the angle adjustment vector in joint space. The dimension of this angle adjustment vector is equal to the number of robot joints, and each element corresponds to the angle value that needs to be changed for each joint axis. The elements of the angle adjustment vector are arranged in order from joint number one to the total number of joints to form a joint angle increment sequence. The amplitude of the joint angle increment sequence is limited to ensure that the increment of each joint angle does not exceed the joint range of motion limit and the single-step adjustment safety threshold.

[0060] Step 106: Generate motion trajectory curves based on the joint angle increment sequence, and drive the robot to adjust the attitude of the extinguishing arc plate according to the motion trajectory curves. During the adjustment process, update the contact force field distribution data in real time until the standard deviation of the normal force component corresponding to the updated contact force field distribution data is less than or equal to the preset pose convergence criterion threshold.

[0061] In this embodiment, the motion trajectory curve refers to the continuous path curve in three-dimensional space of the robot end effector during the process of adjusting from the current pose to the target pose. The standard deviation of the normal force component represents the statistical measure of the dispersion of the normal force component values ​​of all sampled areas in the contact force field distribution data relative to the average normal force. For example, when the normal force values ​​at twelve sampling points are between ten and twelve Newtons, the standard deviation is 0.8 Newtons. The pose convergence criterion threshold is used to represent the critical value of the standard deviation of the normal force component when determining whether the adjustment of the pair is complete. When the real-time updated standard deviation of the normal force component is less than or equal to this threshold, the contact state is considered to have met the uniformity requirements.

[0062] Specifically, the robot's current joint angles are read, and the incremental sequence of joint angle values ​​is added to the corresponding current joint angle to obtain the target angle. A time-discrete intermediate angle sequence is generated between the current angle and the target angle, and a cubic spline interpolation algorithm is used to smooth the angle sequence. The joint angle curves are converted into the trajectory of the end effector's pose in Cartesian space over time using forward kinematics; this trajectory is the motion trajectory curve. Instructions are sent to the servo controllers of each joint according to the motion trajectory curve, driving the robot to perform attitude adjustment of the extinguishing arc plate. During the adjustment process, force sensor data is continuously collected to update the contact force field distribution data in real time. The arithmetic mean of the normal force components in each sampling area after the update is calculated. The sum of squares of the differences between the normal force components at each sampling point and the mean is calculated. The square root of this sum of squares divided by the number of sampling points is used to obtain the standard deviation of the normal force components. The standard deviation is compared with the pose convergence criterion threshold. The adjustment stops when the standard deviation is less than or equal to the threshold; otherwise, the adjustment process is repeated.

[0063] In one possible implementation, the motion trajectory curve is generated based on the joint angle increment sequence, specifically including steps 1061-1062, as follows: Step 1061: Obtain the current joint angle values ​​of the robot; superimpose the joint angle increment sequence with the joint angle values ​​to obtain the target angle command for each joint axis.

[0064] In this embodiment, the current joint angle value of the robot refers to the real-time angular position value of each joint axis before the adjustment command is executed, which is read by the encoder of each joint and is expressed in degrees or radians. The superposition operation represents a mathematical operation that sums each element in the joint angle increment sequence with the corresponding current joint angle value. Each joint axis represents all rotational or translational joints of the robot, including all kinematic pairs from the base joint to the end effector. The target angle command refers to the target angle value that should be achieved after the adjustment of each joint axis, serving as the position command input for motion control.

[0065] Specifically, a joint angle query command is sent to the robot controller, triggering position sampling by the encoders of each joint. The current joint angle measurement value is read from the feedback signals of each joint servo driver, and the read angle values ​​are stored as an array of current joint angle values ​​in joint number order. The dimensionality consistency between the joint angle increment sequence and the current joint angle value array is checked to confirm that the number of elements is equal and the correspondence is correct. Starting from the first joint, the i-th element in the joint angle increment sequence is added to the i-th element in the current joint angle value array. The target angle of the i-th joint is calculated as the current angle value plus the joint angle increment. This superposition operation is repeated for all joints to obtain the target angle value for each joint axis. The calculated target angle values ​​are combined into a target angle command array in joint order. An out-of-bounds check is performed on the target angle commands to determine whether the target angle of each joint is within the mechanical limit range of each joint. If it exceeds the limit, the target angle is clamped to the limit boundary value. The verified target angle command is stored in the command buffer of the motion planning module, awaiting trajectory interpolation processing.

[0066] Step 1062: Extract the timestamp information of the trajectory interpolation points; perform spline interpolation on the target angle command based on the timestamp information to generate a smooth motion trajectory curve.

[0067] In this embodiment, trajectory interpolation points refer to intermediate points discretely distributed on the motion trajectory curve according to time or spatial intervals, used to convert continuous trajectories into discrete control command sequences. Timestamp information represents the time value corresponding to each trajectory interpolation point, used to determine the temporal relationship of the motion process, in seconds or milliseconds. Spline interpolation processing refers to a numerical calculation method that uses a polynomial spline function to fit curves to the discrete interpolation points, generating smooth and continuous motion trajectory curves. Smoothing is used to indicate that the motion trajectory curve remains continuous and differentiable in both time and spatial dimensions, avoiding abrupt changes in velocity and acceleration. The motion trajectory curve represents a continuous function curve of the robot's joint angles changing over time, describing the complete motion process from the current angle to the target angle.

[0068] Specifically, the motion duration from the current joint angle value to the target angle command is determined based on the motion planning algorithm, and the motion duration is divided into several time periods according to a preset interpolation period. Trajectory interpolation points are set at the endpoints of each time period, and the timestamp information of each interpolation point is recorded to form a timestamp sequence. The current joint angle value is set at the interpolation point at the start time, the target angle command is set at the interpolation point at the end time, and the interpolation points remain undefined at intermediate times. A cubic spline interpolation algorithm is used to construct the functional relationship between each joint angle and time, with boundary conditions set to zero velocities at the start and end points to ensure smooth start and stop. Spline interpolation is performed independently for each joint, and the coefficient matrix of the spline function is solved based on the start angle, end angle, and timestamp sequence. The obtained spline function is then calculated at each timestamp to obtain the median joint angle value at each interpolation point. The first and second derivatives of the spline function are performed to verify that the angular velocity and angular acceleration at each interpolation point are continuous and within acceptable ranges. The spline interpolation functions of each joint are aligned according to the timestamps to form a multi-joint collaborative motion trajectory curve.

[0069] In one possible implementation, the robot is driven to adjust the posture of the extinguishing plate according to the motion trajectory curve, and the contact force field distribution data is updated in real time during the adjustment process. Specifically, this includes steps 1063-1064, which are as follows: Step 1063: Drive the servo motors of each joint of the robot to perform angle tracking control according to the motion trajectory curve, so as to adjust the posture of the extinguishing plate.

[0070] In this embodiment, a servo motor refers to an electric actuator with position and speed feedback functions configured on each joint of the robot, which receives angle commands and drives the joints to rotate. Angle tracking control refers to a closed-loop control process in which the servo motor adjusts its output angle in real time according to the angle commands in the motion trajectory curve, so that the actual angle follows the target angle. Driving refers to the operation of sending command signals to the servo motor controller and causing the motor to generate rotational torque. Adjusting the attitude of the extinguishing plate refers to changing the spatial position and rotation direction of the extinguishing plate held by the end effector by changing the angles of each joint of the robot.

[0071] Specifically, the motion trajectory curve data is read, and the target angle values ​​of each joint corresponding to the first time point in the curve are extracted. Position commands are sent to the servo motor controllers of each joint, containing the target angle value and the expected arrival time. Upon receiving the commands, the servo motor controllers initiate the position loop control algorithm, read the current actual angle fed back by the joint encoders, and calculate the deviation between the target angle and the actual angle. Based on the deviation, a motor drive current command is calculated, and current is applied to the motor coils through a current driver to generate electromagnetic torque. This electromagnetic torque drives the joint rotation, and the encoder continuously feeds back angle changes, forming a closed-loop negative feedback adjustment. The target angles at subsequent time points on the motion trajectory curve are read sequentially according to the interpolation cycle, and the position commands sent to the servo motors are periodically updated. Each joint synchronously performs angle tracking control, causing the robot's end effector to move along the planned trajectory. The arc-extinguishing plate held by the end effector changes its position coordinates and attitude angle in space as the robot moves, achieving pose adjustment relative to the welded pipe end face. Throughout the entire motion process, the servo system continuously monitors the tracking error between the actual trajectory and the planned trajectory, controlling the tracking error within an allowable range through closed-loop feedback.

[0072] Step 1064: During the angle tracking control process, monitor the output torque of each joint axis; if the output torque of any joint axis exceeds the preset torque limit, pause the current angle tracking control, reduce the interpolation step size of the motion trajectory curve according to the preset deceleration gradient, and continue to execute angle tracking control based on the reduced interpolation step size; if the output torque of each joint axis does not exceed the preset torque limit, continuously adjust the attitude of the extinguishing plate and obtain the updated contact force field distribution data.

[0073] In this embodiment, output torque refers to the actual rotational torque output by the joint servo motor during angle tracking control, calculated by a motor current sensor, and measured in Newton-meters. Monitoring refers to the process of real-time acquisition of the output torque signals of each joint servo motor and comparison with a preset threshold. The preset torque limit value represents the maximum safe torque threshold allowed for joint output; exceeding this threshold may result in mechanical damage or excessive contact force. Pausing current angle tracking control means stopping the transmission of new position commands to the servo motor and maintaining the current joint angle unchanged. The preset deceleration gradient represents the rate of change of motion speed, measured in milliseconds per step or degrees per second squared. The interpolation step size represents the time interval or angular interval between trajectory interpolation points; a smaller step size results in slower and smoother motion. Reduction refers to the operation of proportionally reducing the interpolation step size value.

[0074] Specifically, when the servo motors of each joint perform angle tracking control, the current values ​​of the motor drive current are read in real time from the current sensors of each motor driver. The drive current value is multiplied by the motor torque constant to calculate the real-time output torque of each joint. The calculated output torque is compared with a preset torque limit to determine if any joint output torque exceeds the limit. If the output torque of any joint exceeds the preset torque limit, new position commands are immediately stopped from being sent to all joints, and the current angle position of each joint remains unchanged. The interpolation step size parameter of the current motion trajectory curve is read, and this parameter is multiplied by the reduction factor corresponding to the preset deceleration gradient to calculate the reduced interpolation step size. The motion trajectory curve is resampled in time using the reduced interpolation step size to generate a new trajectory curve with denser interpolation points. Position commands are continued to be sent to the servo motors based on the new trajectory curve, and angle tracking control continues to be performed at a slower speed. If the output torque of each joint does not exceed the preset torque limit, the current interpolation step size is maintained to continue attitude adjustment, while triggering the force sensor to collect the latest contact force field distribution data and update the data storage area.

[0075] In one possible implementation, if the output torque of each joint axis does not exceed the preset torque limit value, after continuously adjusting the attitude of the extinguishing plate and obtaining the updated contact force field distribution data, the process further includes steps 10641-10642, as follows: Step 10641: Calculate the standard deviation of the normal force component corresponding to the updated contact force field distribution data.

[0076] In this embodiment, the updated contact force field distribution data refers to the latest force sensor measurement data collected in real time during the robot's posture adjustment process, reflecting the contact mechanical state between the extinguishing plate and the welded pipe end face at the current moment. The corresponding numerical relationship associated with the updated contact force field distribution data is represented. The standard deviation of the normal force component is used to represent the statistical degree of dispersion of the normal force component values ​​in each sampling area. It is calculated by taking the square root of the variance; the smaller the standard deviation, the more uniform the normal force distribution. The unit is Newtons. Calculation refers to the process of performing mathematical operations to obtain numerical results.

[0077] Specifically, the normal force components of each sampling region are extracted from the real-time updated contact force field distribution data and arranged in circumferential order to form a normal force array. The arithmetic mean of all elements in the normal force array is calculated, and the sum of the element values ​​is divided by the total number of sampling regions. Each element in the normal force array is iterated over, and the difference between that element's value and the mean is calculated. Each difference is squared to obtain a sequence of squared differences. All elements in the sequence of squared differences are summed to obtain the sum of squared differences. The sum of squared differences is divided by the total number of sampling regions to obtain the variance of the normal force components. The square root of the variance is then taken to calculate the standard deviation of the normal force components. The calculated standard deviation is stored in the state variable of the current iteration, serving as a quantitative indicator for evaluating contact uniformity and pose adjustment convergence. A large standard deviation indicates significant differences in normal forces across different sampling regions, indicating non-uniform contact; a small standard deviation indicates that the normal forces in each region are similar, indicating more uniform contact. The standard deviation of the normal force component is output to the convergence judgment module for decision-making on whether to continue iterative adjustment.

[0078] Step 10642: Determine whether the standard deviation of the normal force component is less than or equal to the preset pose convergence criterion threshold; if the standard deviation of the normal force component is greater than the preset pose convergence criterion threshold, then multiply the standard deviation of the normal force component as a compensation factor with the predicted springback displacement to obtain the compensated predicted springback displacement, update the radial distance distribution data based on the compensated predicted springback displacement, and generate an updated motion trajectory curve based on the updated radial distance distribution data to continue driving the robot to adjust the attitude of the arc extinguishing plate; if the standard deviation of the normal force component is less than or equal to the preset pose convergence criterion threshold, then complete the assembly and positioning of the arc extinguishing plate and the welded pipe end.

[0079] In this embodiment, the judgment refers to the logical decision-making process of comparing the standard deviation of the normal force component with a preset pose convergence criterion threshold and selecting the subsequent execution path based on the comparison result. The compensation factor represents a correction coefficient used to adjust the predicted springback displacement, and its value is the standard deviation of the normal force component. The product operation refers to the mathematical operation of multiplying the compensation factor and the predicted springback displacement to obtain the compensated value. The compensated predicted springback displacement represents the new predicted springback displacement value after correction by the standard deviation of the normal force, taking into account the influence of the current contact non-uniformity on the springback. The updated motion trajectory curve represents the robot motion path curve regenerated based on the updated radial distance distribution data. Completing the alignment and positioning of the arc-extinguishing plate and the welded pipe end means achieving an ideal alignment state with uniform contact force distribution between the arc-extinguishing plate and the welded pipe end face, thus ending the pose adjustment process.

[0080] Specifically, the standard deviation of the normal force component and the preset pose convergence threshold parameter are read, and the standard deviation is compared with the threshold. If the standard deviation is greater than the threshold, it is determined that the current pose has not yet converged and needs further adjustment, and the standard deviation is assigned to the compensation factor variable. The compensation factor is multiplied by the original predicted rebound displacement, and the product is calculated as the compensated predicted rebound displacement. The compensated predicted rebound displacement replaces the original predicted rebound displacement, and the radial distance correction calculation is re-executed. The compensated predicted rebound displacement is added to the radial intrusion depth of the over-compression region in the current radial distance distribution data to obtain the updated target radial distance distribution data. Based on the updated target radial distance distribution data, the spatial pose deviation matrix construction, orthogonal decomposition, gain adjustment, and joint angle increment calculation are re-executed. Based on the newly calculated joint angle increment sequence, the updated motion trajectory curve is generated, driving the robot to continue to perform pose adjustment. If the standard deviation is less than or equal to the threshold, it is determined that the pose has converged and reached the adjustment target, the motion command is stopped, the current extinguishing arc plate pose is maintained, the group positioning completion flag is set, and the entire pose adjustment control process ends.

[0081] In the above embodiments, by setting multiple sampling areas in the circumferential direction of the welded pipe end face and combining them with a six-dimensional force sensor to collect contact force field distribution data, the contact state between the arc-extinguishing plate and the end of the welded pipe is accurately characterized. Furthermore, the centroid coordinates of the over-extrusion area and the locally suspended area are determined by calculating the normal force gradient change rate. To further address the difficulties of traditional assembly methods in accurately quantifying the coupling relationship between spatial geometric deviations and mechanical deviations at the contact interface, and the lack of collaborative identification capabilities for tilt angle deviations and normal vector deflection angle deviations, and to ensure the accurate acquisition of pose deviation parameters and the accurate generation of spatial pose adjustment commands, this application also provides a robotic intelligent assembly welding method for straight seam welded pipe arc-extinguishing plates. This method calculates the radial intrusion depth and radial separation depth by measuring the radial offset of the centroid coordinates of the over-extrusion region and the centroid coordinates of the locally suspended region relative to the target contact circle of the welded pipe end face. These are then combined to form radial distance distribution data. By analyzing the distribution of the two centroid coordinates in the circumferential direction, the circumferential angle difference is calculated. Furthermore, by combining the tangential moment component in the contact force field distribution data and the ratio of the moment difference to the circumferential angle difference, the tilt angle deviation of the ignition plate plane relative to the welded pipe end face is determined. Based on the tilt angle deviation and the initial normal vector of the welded pipe end face, the deflection angle deviation of the ignition plate normal vector relative to the welded pipe end face normal vector is calculated. This allows the system to more accurately and comprehensively handle the requirements of radial distance measurement, tilt angle identification, and normal vector deflection angle determination in the assembly and positioning scenario of the ignition plate and the welded pipe end face. The following section combines... Figure 2 The following describes the welding method using a robotic intelligent group for initiating and extinguishing arc plates for straight seam welded pipes, as described in another embodiment of this application: Please see Figure 2This is another schematic diagram of the process of a robotic intelligent assembly welding method for straight seam welded pipe arc extinguishing plates in an embodiment of this application.

[0082] Step 201: Calculate the radial penetration depth of the over-extended region and the radial separation depth of the local suspended region relative to the target contact circle of the welded pipe end face, based on the radial offset of the centroid coordinates of the over-extended region and the centroid coordinates of the local suspended region.

[0083] In this embodiment, the target contact circle refers to the circular contact trajectory formed between the arc-extinguishing plate and the end face of the welded pipe under ideal contact conditions, and this circle is located on the end face of the welded pipe. The radial offset represents the distance the centroid coordinates deviate from the target contact circle in the radial direction; a positive value indicates an outward offset, and a negative value indicates an inward offset. For example, when the centroid is 52mm from the center of the circle and the radius of the target contact circle is 50mm, the offset is 2mm. The radial penetration depth of the over-extrusion region represents the distance the centroid of the over-extrusion region penetrates radially into the inside of the welded pipe. The radial separation depth of the locally suspended region refers to the separation distance between the centroid of the locally suspended region and the target contact circle in the radial direction.

[0084] Specifically, the three-dimensional spatial position information of the centroid coordinates of the over-extended region and the locally suspended region is read. The geometric parameters of the welded pipe end face are obtained, including the coordinates of the end face center and the radius of the target contact circle. The centroid coordinates of the over-extended region are transformed into a cylindrical coordinate system with the center of the welded pipe end face as the origin, and the radial distance from the centroid to the center of the circle is calculated. The radial offset of the over-extended region is obtained by subtracting the radius of the target contact circle from this radial distance. If the radial offset is positive, it is directly recorded as the radial penetration depth of the over-extended region. If the radial offset is negative, its absolute value is taken as the radial penetration depth. The same method is used to process the centroid coordinates of the locally suspended region, transforming them into a cylindrical coordinate system and calculating the radial distance to the center of the circle. The radial offset of the locally suspended region is obtained by subtracting the radius of the target contact circle from the radial distance of the centroid of the locally suspended region. If the radial offset is negative, its absolute value is recorded as the radial separation depth of the locally suspended region. If the radial offset is positive, its absolute value is also taken as the radial separation depth. The calculated radial penetration depth of the over-extrusion area and the radial separation depth of the local suspended area are stored separately. These two values ​​describe the contact deviation characteristics of the arc extinguishing plate and the welded pipe end face in different areas in the radial direction.

[0085] Step 202: Combine the radial intrusion depth of the over-compressed area with the radial separation depth of the local suspended area to obtain radial distance distribution data.

[0086] In this embodiment, "combination" refers to the operation of integrating radial distance data from multiple different sources according to their spatial positional relationships. The radial distance distribution data represents the set of radial distance values ​​at various positions along the circumferential direction of the welded pipe end face. This data includes radial distance information for the over-extended region, the locally suspended region, and other regions. The radial penetration depth of the over-extended region and the radial separation depth of the locally suspended region refer to the radial distance characteristic values ​​of the two characteristic regions calculated previously.

[0087] Specifically, a data structure for radial distance distribution data is created, indexed according to the angular position of the welded pipe end face circumferentially. The angular coordinates of each sampling position in the contact force field distribution data are read, establishing a correspondence between the sampling position and the circumferential angle. Based on the angular position of the centroid coordinates of the over-extrusion region in the circumferential direction, the corresponding data item is located in the radial distance distribution data structure. The calculated radial penetration depth value of the over-extrusion region is filled into this data item, and the data item is marked as penetration type. Based on the angular position of the centroid coordinates of the local suspended region in the circumferential direction, the corresponding data item is located in the radial distance distribution data structure. The calculated radial separation depth value of the local suspended region is filled into this data item, and the data item is marked as separation type. For other sampling positions that do not belong to either the over-extrusion region or the local suspended region, the radial displacement measurement value of the corresponding position is extracted from the contact force field distribution data. These radial displacement measurement values ​​are sequentially filled into the corresponding positions in the radial distance distribution data and marked as normal contact type. All sampling positions in the circumferential direction are traversed to ensure that each position has a corresponding radial distance value. Perform an integrity check on the completed radial distance distribution data to check for any missing data items. Output the complete radial distance distribution data, which is arranged in circumferential angle order and fully describes the radial distance distribution characteristics of the contact interface between the arc extinguishing plate and the welded pipe end face.

[0088] Step 203: Based on the distribution of the centroid coordinates of the over-extended region and the centroid coordinates of the locally suspended region in the circumferential direction of the welded pipe end face, calculate the circumferential angle difference between the centroid coordinates of the over-extended region and the centroid coordinates of the locally suspended region.

[0089] In this embodiment, the circumferential direction of the welded pipe end face refers to the direction extending along the tangent of the circumference with the center of the welded pipe end face as the center, and the angle range is from 0 degrees to 360 degrees. The distribution position is used to represent the angular coordinates corresponding to the centroid coordinates in the circumferential direction, which is determined by the azimuth angle of the centroid relative to the center of the circle. The circumferential angle difference refers to the angular interval between the centroid coordinates of the over-extended area and the centroid coordinates of the locally suspended area in the circumferential direction. For example, when the centroid of the over-extended area is located at 30 degrees and the centroid of the locally suspended area is located at 210 degrees, the angle difference is 180 degrees.

[0090] Specifically, the X-axis and Y-axis components of the centroid coordinates of the over-extended region are read. The X-axis and Y-axis components of the center coordinates of the welded pipe end face are obtained. The X-direction offset of the centroid of the over-extended region relative to the center of the circle is calculated by subtracting the X-coordinate of the center of the circle from the X-coordinate of the centroid. The Y-direction offset of the centroid of the over-extended region relative to the center of the circle is calculated by subtracting the Y-coordinate of the center of the circle from the Y-coordinate of the centroid. The arctangent of the ratio of the Y-direction offset to the X-direction offset is performed to obtain the angular coordinates of the centroid of the over-extended region in the circumferential direction. The same method is used to calculate the X-direction and Y-direction offsets of the centroid of the locally suspended region relative to the center of the circle. The arctangent of the ratio of the Y-direction offset to the X-direction offset of the locally suspended region is performed to obtain the angular coordinates of the centroid of the locally suspended region in the circumferential direction. The two angular coordinates are uniformly converted to the standard range of 0 degrees to 360 degrees. The angular coordinates of the centroid of the locally suspended region are subtracted from the angular coordinates of the centroid of the over-extended region to obtain the preliminary angular difference. If the initial angle difference is greater than 180 degrees, subtract this difference from 360 degrees to obtain the final angle difference. If the initial angle difference is less than -180 degrees, add 360 degrees to this difference to obtain the final angle difference. If the initial angle difference is between -180 degrees and 180 degrees, use it directly as the final circumferential angle difference. Store the calculated circumferential angle difference; this value describes the relative positional relationship of the centroids of the two feature regions in the circumferential direction.

[0091] Step 204: Calculate the tilt angle deviation and the normal vector deflection angle deviation based on the circumferential angle difference and the tangential moment component in the contact force field distribution data.

[0092] In this embodiment, the tangential moment component represents the torque value generated tangentially along the end face of the welded pipe in the contact force field. This torque is generated by the product of the tangential force and the lever arm. The tilt angle deviation refers to the tilt angle of the arc-extinguishing plate plane relative to the end face plane of the welded pipe, reflecting the degree of non-parallelism between the two planes. The normal vector deflection angle deviation is used to represent the spatial angle between the normal vector of the arc-extinguishing plate and the normal vector of the end face of the welded pipe.

[0093] Specifically, the tangential moment components at each sampling location are read from the contact force field distribution data. The corresponding sampling locations are located based on the circumferential angle positions of the centroid coordinates of the over-compression region and the centroid coordinates of the locally suspended region. The tangential moment components corresponding to the centroid positions of the over-compression region and the locally suspended region are extracted. The tangential moment component of the over-compression region is subtracted from the tangential moment component of the locally suspended region to calculate the moment difference. The moment difference is divided by the circumferential angle difference to obtain the rate of change of moment per unit angle. The radius parameters of the welded pipe end face and the equivalent radius parameters of the contact area are obtained. The rate of change of moment is divided by the square of the radius of the welded pipe end face to obtain the force eccentricity per unit area. Based on the ratio of force eccentricity to the contact stiffness coefficient, the tangent value of the inclination angle of the extinguishing plate is calculated. The arctangent of the inclination angle tangent is calculated to obtain the inclination angle deviation of the extinguishing plate plane relative to the welded pipe end face. The three coordinate components of the initial normal vector of the welded pipe end face are read. A spatial rotation matrix is ​​constructed based on the tilt angle deviation and tilt direction. The initial normal vector of the welded pipe end face is multiplied by the rotation matrix to obtain the three coordinate components of the arc-extinguishing plate normal vector. The dot product of the arc-extinguishing plate normal vector and the initial normal vector of the welded pipe end face is calculated; this dot product equals the cosine of the angle between the two vectors. An inverse cosine operation is performed on the cosine of the angle to obtain the deflection angle deviation of the arc-extinguishing plate normal vector relative to the welded pipe end face normal vector. The tilt angle deviation and deflection angle deviation are stored and output.

[0094] In one possible implementation, the tilt angle deviation and the normal vector deflection angle deviation are calculated based on the circumferential angle difference and the tangential moment component in the contact force field distribution data, specifically including steps 2041-2043, as follows: Step 2041: Extract the tangential moment components corresponding to the centroid coordinates of the over-compressed region and the centroid coordinates of the locally suspended region from the contact force field distribution data.

[0095] In this embodiment, extraction refers to the operation of finding and reading the corresponding data value from the contact force field distribution data according to the specified coordinate position. The tangential moment component represents the measured moment value along the tangential direction of the welded pipe end face at the sampling position. The tangential moment component corresponding to the centroid coordinates of the over-extended region is used to represent the tangential moment value at the centroid position of the over-extended region or the nearest sampling position. The tangential moment component corresponding to the centroid coordinates of the locally suspended region refers to the tangential moment value at the centroid position of the locally suspended region or the nearest sampling position.

[0096] Specifically, the complete data structure of the contact force field distribution data is read, which includes the coordinate information of each sampling location and the corresponding tangential moment component value. The three-dimensional spatial position of the centroid coordinates of the over-compression region is obtained. The coordinates of all sampling locations in the contact force field distribution data are traversed, and the spatial distance between each sampling location and the centroid coordinates of the over-compression region is calculated. The sampling location with the smallest spatial distance is found and marked as the sampling location corresponding to the centroid of the over-compression region. The value of the tangential moment component is read from the data item of this sampling location and recorded as the tangential moment component corresponding to the centroid coordinates of the over-compression region. The same method is used to process the centroid coordinates of the local suspended region to obtain its three-dimensional spatial position information. The spatial distance between each sampling location in the contact force field distribution data and the centroid coordinates of the local suspended region is calculated. The sampling location closest to the centroid of the local suspended region is found and marked as the sampling location corresponding to the centroid of the local suspended region. The value of the tangential moment component is read from the data item of this sampling location and recorded as the tangential moment component corresponding to the centroid coordinates of the local suspended region. The two extracted tangential moment components are checked for unit consistency to confirm that they use the same moment unit. The extracted tangential moment components of the over-compression region and the locally suspended region are output as input data for subsequent moment difference calculation.

[0097] Step 2042: Calculate the torque difference between the tangential torque component corresponding to the centroid coordinates of the over-compressed region and the tangential torque component corresponding to the centroid coordinates of the locally suspended region.

[0098] In this embodiment, the torque difference refers to the numerical difference between the tangential torque component corresponding to the centroid coordinates of the over-compressed region and the tangential torque component corresponding to the centroid coordinates of the locally suspended region. This difference reflects the degree of imbalance in tangential torque between the two regions. For example, when the tangential torque of the over-compressed region is 0.8 N·m and the tangential torque of the locally suspended region is -0.6 N·m, the difference is 1.4 N·m. Calculation refers to the mathematical operation of performing subtraction to obtain the difference.

[0099] Specifically, read the tangential moment component value corresponding to the centroid coordinates of the over-compressed region, and record its sign and magnitude. Read the tangential moment component value corresponding to the centroid coordinates of the locally suspended region, and similarly record its sign and magnitude. Check the units of measurement for the two tangential moment components to ensure they are consistent. Subtract the tangential moment component value corresponding to the centroid coordinates of the locally suspended region from the tangential moment component value corresponding to the centroid coordinates of the over-compressed region. If the two moment components have the same sign and the absolute value of the minuend is larger, the difference is positive and its absolute value equals the difference between the two absolute values. If the two moment components have the same sign and the absolute value of the minuend is larger, the difference is negative and its absolute value equals the difference between the two absolute values. If the two moment components have opposite signs, the absolute value of the difference equals the sum of their absolute values, and the sign of the difference is the same as the sign of the minuend. Store the calculated difference result as a moment difference value. The torque difference is determined by sign: a positive value indicates that the over-compressed area experiences a larger positive tangential torque, while a negative value indicates that the over-compressed area experiences a larger negative tangential torque or that the tangential torque dominates in a locally suspended area. The torque difference is output, and this value, along with the circumferential angle difference, is used in subsequent calculations of tilt angle deviation and normal vector deflection angle deviation. The order of magnitude of the torque difference is checked to confirm that it is within a reasonable engineering range.

[0100] Step 2043: Calculate the tilt angle deviation of the arc extinguishing plate plane relative to the welded pipe end face based on the ratio of the torque difference to the circumferential angle difference; calculate the deflection angle deviation of the arc extinguishing plate normal vector relative to the welded pipe end face normal vector based on the tilt angle deviation and the initial normal vector of the welded pipe end face.

[0101] In this embodiment, the ratio relationship refers to the quotient obtained by dividing the torque difference by the circumferential angle difference and its physical meaning. The ignition and extinguishing plate plane represents the main contact surface of the ignition and extinguishing plate, which ideally fits perfectly against the welded pipe end face. The tilt angle deviation represents the tilt angle of the ignition and extinguishing plate plane relative to the welded pipe end face plane. The initial normal vector of the welded pipe end face refers to a unit vector perpendicular to the welded pipe end face and along the welded pipe axis. The ignition and extinguishing plate normal vector represents a unit vector perpendicular to the ignition and extinguishing plate plane. The deflection angle deviation refers to the spatial angle between the ignition and extinguishing plate normal vector and the welded pipe end face normal vector.

[0102] Specifically, the values ​​of the torque difference and the circumferential angle difference are read. The torque difference is divided by the circumferential angle difference to obtain the torque change per unit angle. The radius parameters of the welded pipe end face and the effective radius parameters of the contact area of ​​the ignition and extinguishing plate are obtained. The torque change per unit angle is divided by the square of the radius of the welded pipe end face to obtain the normalized force eccentricity coefficient. Based on the ratio of the stiffness coefficient of the contact area to the force eccentricity coefficient, the tangent value of the tilt angle of the ignition and extinguishing plate is calculated. The arctangent function is applied to the tilt angle tangent value to obtain the tilt angle deviation of the ignition and extinguishing plate plane relative to the welded pipe end face. The three components of the initial normal vector of the welded pipe end face in the X, Y, and Z axes are read. Based on the magnitude of the tilt angle deviation and the tilt direction angle, a three-dimensional rotation transformation matrix is ​​constructed. The three components of the initial normal vector of the welded pipe end face are multiplied by the rotation transformation matrix to obtain the X-axis, Y-axis, and Z-axis components of the ignition and extinguishing plate normal vector. Calculate the dot product between the normal vector of the arc-extinguishing plate and the initial normal vector of the welded pipe end face, and sum the corresponding components of the two vectors. Since both vectors are unit vectors, the dot product is equal to the cosine of the angle between them. Perform an inverse cosine function operation on the cosine of the angle to obtain the deflection angle deviation of the arc-extinguishing plate normal vector relative to the normal vector of the welded pipe end face. Output and store the tilt angle deviation and the deflection angle deviation.

[0103] The following describes a welding system based on a robotic intelligent assembly of arc-quenching and arc-extinguishing plates for straight seam welded pipes, as described in this application, from a hardware processing perspective. Please refer to [link to relevant documentation]. Figure 3 This is a schematic diagram of the structure of a robotic intelligent assembly welding system for straight seam welded pipe arc extinguishing plates in an embodiment of this application.

[0104] It should be noted that, Figure 3 The structure of the robotic intelligent assembly welding system for straight seam welded pipe arc extinguishing plate shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0105] like Figure 3 As shown, a robotic intelligent assembly welding system for straight seam welded pipes includes a Central Processing Unit (CPU) 301, which can perform various appropriate actions and processes according to a program stored in a Read-Only Memory (ROM) 302 or a program loaded from a storage section 308 into a Random Access Memory (RAM) 303, such as the method described in the above embodiment. The RAM 303 also stores various programs and data required for system operation. The CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An Input / Output (I / O) interface 305 is also connected to the bus 304.

[0106] The following components are connected to I / O interface 305: input section 306 including audio input devices, push-button switches, etc.; output section 307 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 308 including a hard disk, etc.; and communication section 309 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 309 performs communication processing via a network such as the Internet. Drive 310 is also connected to I / O interface 305 as needed. Removable media 311, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 310 as needed so that computer programs read from them can be installed into storage section 308 as needed.

[0107] The above description is merely a specific embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A robotic intelligent assembly welding method for straight seam welded pipe arc-extinguishing plates, characterized in that, The method includes: The contact force field distribution data is obtained, which is force information collected by the robot in multiple sampling areas in the circumferential direction of the welded pipe end face; The normal force component is extracted based on the contact force field distribution data, and the normal force gradient change rate between adjacent sampling areas is calculated. Based on the normal force gradient change rate, the centroid coordinates of the over-compression area and the centroid coordinates of the local suspended area are determined. Based on the centroid coordinates of the over-compressed region and the centroid coordinates of the locally suspended region, determine the radial distance distribution data, tilt angle deviation, and normal vector deflection angle deviation. Based on the radial distance distribution data, the elastic rebound amount is predicted to obtain the predicted rebound displacement, and the radial distance distribution data is corrected using the predicted rebound displacement to obtain the target radial distance distribution data. A spatial pose deviation matrix is ​​constructed based on the tilt angle deviation, the normal vector deflection angle deviation, and the target radial distance distribution data. The spatial pose deviation matrix is ​​then decomposed and its gain is adjusted to generate a joint angle increment sequence. The motion trajectory curve is generated based on the joint angle increment sequence, and the robot is driven to adjust the attitude of the extinguishing plate based on the motion trajectory curve. During the adjustment process, the contact force field distribution data is updated in real time until the standard deviation of the normal force component corresponding to the updated contact force field distribution data is less than or equal to the preset pose convergence criterion threshold.

2. The method according to claim 1, characterized in that, The step of extracting the normal force component based on the contact force field distribution data and calculating the normal force gradient change rate between adjacent sampling areas, and determining the centroid coordinates of the over-compression region and the local suspended region based on the normal force gradient change rate, includes: According to the arrangement order of each sampling area in the circumferential direction of the welded pipe end face, the ratio of the difference of the normal force component between two adjacent sampling areas to the corresponding arc length is calculated to obtain the normal force gradient change rate between each adjacent sampling area. The sampling region where the rate of change of the normal force gradient is greater than or equal to the preset extrusion gradient threshold is determined as an over-extrusion candidate region, and the sampling region where the rate of change of the normal force gradient is less than the preset dangling gradient threshold is determined as a local dangling candidate region. The centroid coordinates of each of the candidate regions for over-compression are calculated by weighting the absolute value of the rate of change of the normal force gradient corresponding to each candidate region for over-compression, thereby obtaining the centroid coordinates of the over-compression region. The centroid coordinates of each of the local suspended candidate regions are calculated by weighting the absolute value of the rate of change of the normal force gradient corresponding to each of the local suspended candidate regions, thereby obtaining the centroid coordinates of the local suspended region.

3. The method according to claim 1, characterized in that, The step of determining radial distance distribution data, tilt angle deviation, and normal vector deflection angle deviation based on the centroid coordinates of the over-compressed region and the centroid coordinates of the locally suspended region includes: Based on the radial offset of the centroid coordinates of the over-extended region and the centroid coordinates of the locally suspended region relative to the target contact circle of the welded pipe end face, the radial penetration depth of the over-extended region and the radial separation depth of the locally suspended region are calculated respectively. The radial intrusion depth of the over-compressed region and the radial separation depth of the local suspended region are combined to obtain the radial distance distribution data; Based on the distribution positions of the centroid coordinates of the over-extended region and the centroid coordinates of the locally suspended region in the circumferential direction of the welded pipe end face, calculate the circumferential angle difference between the centroid coordinates of the over-extended region and the centroid coordinates of the locally suspended region. The tilt angle deviation and the normal vector deflection angle deviation are determined based on the circumferential angle difference and the tangential moment component in the contact force field distribution data.

4. The method according to claim 3, characterized in that, The step of calculating the tilt angle deviation and the normal vector deflection angle deviation based on the circumferential angle difference and the tangential moment component in the contact force field distribution data includes: Extract the tangential moment components corresponding to the centroid coordinates of the over-compressed region and the centroid coordinates of the locally suspended region from the contact force field distribution data; Calculate the torque difference between the tangential torque component corresponding to the centroid coordinates of the over-compressed region and the tangential torque component corresponding to the centroid coordinates of the locally suspended region; Based on the ratio of the torque difference to the circumferential angle difference, the tilt angle deviation of the arc extinguishing plate plane relative to the end face of the welded pipe is calculated. Based on the tilt angle deviation and the initial normal vector of the welded pipe end face, calculate the deflection angle deviation of the arc extinguishing plate normal vector relative to the normal vector of the welded pipe end face.

5. The method according to claim 1, characterized in that, The process of predicting elastic rebound amount based on the radial distance distribution data to obtain predicted rebound displacement, and then using the predicted rebound displacement to correct the radial distance distribution data to obtain target radial distance distribution data, includes: Obtain the preset elastic modulus parameters of the arc ignition plate material and the equivalent contact stiffness of the end face of the welded pipe. The elastic restoring force is obtained by multiplying the radial penetration depth of the excessive compression region in the radial distance distribution data with the equivalent contact stiffness. Calculate the ratio of the elastic restoring force to the elastic modulus parameter of the arc extinguishing plate material to obtain the predicted rebound displacement of the over-compression region after the pressure is released; The predicted rebound displacement is added to the radial intrusion depth of the over-compression region in the radial distance distribution data, and the radial distance distribution data is updated based on the addition result to obtain the target radial distance distribution data.

6. The method according to claim 1, characterized in that, The process involves constructing a spatial pose deviation matrix based on the tilt angle deviation, the normal vector deflection angle deviation, and the target radial distance distribution data; decomposing and adjusting the gain of the spatial pose deviation matrix to generate a joint angle increment sequence, including: The tilt angle deviation, the normal vector deflection angle deviation, and the target radial distance distribution data are converted into spatial translation component deviation values ​​and spatial rotation component deviation values. The spatial pose deviation matrix is ​​determined based on the spatial translation component deviation values ​​and the spatial rotation component deviation values. Orthogonally decompose each component in the spatial pose deviation matrix to obtain the position deviation decomposition and force deviation decomposition. The contact stiffness parameters between the ignition plate and the end of the welded pipe are calculated based on the contact force field distribution data, and the proportional coefficient between the position loop gain and the force feedback closed loop gain is dynamically adjusted based on the contact stiffness parameters. Based on the position loop gain, force feedback closed loop gain, position deviation decomposition, and force deviation decomposition after adjusting the proportional coefficient, the incremental sequence of joint angles for multiple joint axes of the robot is calculated through the mapping relationship matrix between the robot's end-effector pose and joint angles.

7. The method according to claim 1, characterized in that, The step of generating a motion trajectory curve based on the joint angle increment sequence includes: Obtain the current joint angle values ​​of the robot; The joint angle increment sequence is superimposed with the joint angle value to obtain the target angle command for each joint axis; Extract the timestamp information of the trajectory interpolation points; Based on the timestamp information, spline interpolation is performed on the target angle command to generate a smooth motion trajectory curve.

8. The method according to claim 1, characterized in that, The step of driving the robot to adjust the attitude of the extinguishing plate according to the motion trajectory curve, and updating the contact force field distribution data in real time during the adjustment process, includes: The robot's joint servo motors are driven to perform angle tracking control according to the motion trajectory curve, so as to adjust the posture of the extinguishing plate. During the execution of the angle tracking control, the output torque of each joint axis is monitored; If the output torque of any of the joint axes exceeds the preset torque limit, the current angle tracking control is paused, and the interpolation step size of the motion trajectory curve is reduced according to the preset deceleration gradient. The angle tracking control is then continued based on the reduced interpolation step size. If the output torque of each joint axis does not exceed the preset torque limit value, the attitude of the extinguishing plate is continuously adjusted, and the updated contact force field distribution data is obtained.

9. The method according to claim 8, characterized in that, If the output torque of each of the joint axes does not exceed the preset torque limit value, then after continuously adjusting the attitude of the extinguishing plate and obtaining the updated contact force field distribution data, the method further includes: Calculate the standard deviation of the normal force component corresponding to the updated contact force field distribution data; Determine whether the standard deviation of the normal force component is less than or equal to the preset pose convergence criterion threshold; If the standard deviation of the normal force component is greater than the preset pose convergence criterion threshold, the standard deviation of the normal force component is used as a compensation factor and multiplied with the predicted rebound displacement to obtain the compensated predicted rebound displacement. The radial distance distribution data is updated based on the compensated predicted rebound displacement, and an updated motion trajectory curve is generated based on the updated radial distance distribution data to continue driving the robot to adjust the attitude of the extinguishing arc plate. If the standard deviation of the normal force component is less than or equal to the preset pose convergence criterion threshold, then the assembly and positioning of the arc extinguishing plate and the end of the welded pipe is completed.

10. A robotic intelligent assembly welding system for straight seam welded pipe arc-starting and extinguishing plates, characterized in that, The straight seam welded pipe arc-quenching and extinguishing plate robot intelligent assembly welding system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the straight seam welded pipe arc-quenching and extinguishing plate robot intelligent assembly welding system to perform the method as described in any one of claims 1-9.

Images