Laser welding seam automatic tracing method, detection method and device

Abstract

The present invention relates to an automatic tracing method for laser welding welds, comprising the following steps: installing a focus tool, fixing a tool tip tool, determining a reference point of the focus tool and a fixed point of the tool tip tool, and establishing a laser welding gun focus coordinate system using a four-point method; establishing a robot base coordinate system based on a mounting base of a robot, and determining a coordinate relationship between the robot base coordinate system and the laser welding gun focus coordinate system according to a posture matrix of the robot base coordinate system; scanning an edge of a rib plate to obtain a surface contour image of a test piece to be welded, extracting weld feature points on both side edges of a laser strip from the obtained surface contour image of the test piece to be welded, and respectively making line segments one and two for the weld feature points on the two side edges; fitting line segments one and two of the weld feature, and solving their intersection as the weld center trajectory coordinates. The method aims to replace manual teaching, automatically complete weld welding and post-weld inspection, and greatly improve welding efficiency and quality stability.

Description

Technical Field

[0001] The present invention belongs to the technical field of laser welding automated processing of thin-walled structural parts with complex spatial curved surfaces, and in particular relates to a laser welding weld seam automatic tracing method, detection method and device. Background Art

[0002] The stiffener panels of aviation, automobiles, etc. usually have T-shaped weld structures. T-shaped welds are generally welded by dual-beam laser welding from the left and right sides. The laser welding quality and efficiency of T-shaped weld structures directly affect the development speed of these industries and the promotion of laser welding applications. Robotic dual-beam laser welding is one of the main production and processing methods for titanium alloy assemblies. In order to meet the requirements of stealth and lightweight, riveting and spot welding cannot be used. Due to the hot forming deformation of the workpiece, stiffener processing errors and clamping errors, the consistency of the welds to be welded is poor, and the offline programming trajectory cannot be directly used for welding. Each weld requires manual teaching and calibration. The manual teaching welding quality consistency is poor and the efficiency is low, which is not adapted to the rapid development of my country's new generation of aircraft, and seriously affects the production quality and efficiency of my country's aircraft titanium alloy assemblies.

[0003] Laser welding of thin-walled structural parts with complex spatial curves made of aviation titanium alloys is difficult. First, the welding space is narrow, the ribs are crisscrossed, and the spatial curves are complex. The measuring device needs to measure at a distance without interfering with the tooling and workpiece. The accuracy of the measuring device decreases as its own measurement range increases. Laser welding requires very high precision, with weld quality reaching Level I and weld center deviation no more than ±0.1mm. Therefore, the system requires the measuring device to be small and compact in size while ensuring high measurement accuracy. Secondly, the titanium alloy wall panels must be laser cleaned before welding, and the ribs are vertically assembled on the base plate, forming an effect similar to two mirrors arranged vertically. Multiple reflections are easily generated during sensor measurement, seriously affecting measurement accuracy.

[0004] Universal structured light sensors experience significant loss in measurement accuracy as their field of view and distance increase, failing to meet process requirements. Structured laser sensors with narrow fields of view, however, lack the accessibility and field of view required for titanium alloy laser welding of T-joints. Furthermore, at laser welding speeds of 6 to 12 m / min, the robotic weld measurement system's slow dynamic tracking and dynamic adjustment of the weld trajectory results in poor weld trajectory quality, failing to meet process requirements.

[0005] In terms of post-weld weld inspection, dual-beam welding welds are required to meet the Class I weld standard for navigation marks, primarily including requirements for internal weld quality, mechanical properties, and appearance. X-ray flaw detection is the primary method for detecting internal weld defects, with common defects including cracks, incomplete penetration, lack of fusion, porosity, and slag inclusions. Mechanical properties are primarily tested using tensile testing, with tensile strength and shear strength meeting technical requirements. Appearance quality requirements include uniform weld transitions, the absence of pits, weld bumps, undercuts, and other defects at arc start and end points, and a weld leg height difference of ≤0.3mm. Appearance quality is a key evaluation criterion for dual-beam welding process quality. Currently, manual visual inspection for surface defects is the primary method, with metallographic methods used to measure weld leg height differences. Metallographic methods for measuring weld leg height require cutting the T-shaped weld to produce metallographic specimens. This method is inefficient and cannot inspect actual welded workpieces. Summary of the Invention

[0006] The present invention mainly addresses the above problems and proposes an automatic tracking method, detection method and device for laser welding welds, which replace manual teaching, automatically complete weld welding and post-weld detection, and greatly improve welding efficiency and quality stability.

[0007] To achieve the above object, the present invention provides a method for automatically tracking a laser welding seam, comprising the following steps:

[0008] Step 1: Install a focus tool at the end of the laser welding gun, determine the reference point of the focus tool, fix the tool tip on the workbench, determine the fixed point of the tool tip, use the four-point method to make the reference point just touch the fixed point, and establish the laser welding gun focus coordinate system based on the data of the four position points;

[0009] Step 2: With the line laser sensor installed on the robot's mounting base as a reference, establish a robot base coordinate system, and determine the coordinate relationship between the robot base coordinate system and the laser welding gun focus coordinate system according to the pose matrix of the robot base coordinate system;

[0010] Step 3: The robot drives the line laser sensor to move along the welding position of the welded specimen, scans the edge of the rib plate, and obtains the surface contour image of the welded specimen.

[0011] Step 4: Extracting weld feature points on both sides of the laser band from the obtained surface contour image of the welded specimen, and creating line segments 1 and 2 for the weld feature points on both sides of the laser band, respectively, with the created line segments 1 and 2 being weld features;

[0012] Step 5: Fit the line segment 1 and line segment 2 of the weld feature, and solve their intersection point to obtain the coordinates of the weld center trajectory.

[0013] Furthermore, in step 1, the calibration process of establishing the laser welding gun focus coordinate system is:

[0014] Tool center point position calibration: Manually operate the robot to move the laser welding gun from four different directions to a fixed point, record the four points P1, P2, P3, and P4, and the robot end flange pose data (X, Y, Z, θ1, θ2, θ3). The coordinates of the four tool center points in the world coordinate system are equal. Calculation is completed on the teaching pendant to obtain the tool coordinate system TCP (X, Y, Z).

[0015] Tool coordinate system posture calibration: Move the tool coordinate system TCP to any fixed point for measurement, move a point in the negative direction of the Y axis to the fixed point for measurement, then move any point in the XY plane with a negative X value to the fixed point for measurement, record the data, and complete the calculation calibration on the teaching pendant.

[0016] Furthermore, in step 2, the step of determining the coordinate relationship between the robot base coordinate system and the laser welding gun focus coordinate system includes:

[0017] When the robot is in a certain posture, the posture matrix of the weld seam measurement end effector relative to the robot base coordinate is T1, the posture matrix of the line sensor coordinate system to the end flange is X1, and the posture of the world coordinate system in the robot base coordinate system is P1=T1X1M1. When the robot moves to another posture, the above parameters become T2, X2, and P2. Since the line sensor is fixed on the end, X1=X2. Since the world coordinate system and the end coordinate system are stationary, P1=P2. Let Then AX=XB,

[0018] Assume A is an m×n matrix, B is an n×n matrix, and X is an m×n matrix. From the definition of matrix direct product, we know that:

[0019]

[0020] When A or B is the identity matrix, we have

[0021]

[0022] Let a and b be constants, then we have,

[0023] vec(aA+bB)=a×vec(A)+b×vec(B)

[0024] Assumptions

[0025]

[0026] Then AX=XB can be expressed as:

[0027]

[0028]

[0029] To determine the unique solution of the above equation, two sets of motions with non-parallel rotation axes are required. The simultaneous equations obtained from these two sets of motions are:

[0030]

[0031]

[0032] Use the least squares method to solve the eigenvector v=[v1v2…v 13 ], then the relationship matrix is:

[0033]

[0034] Furthermore, the step 3 includes making the YZ plane of the laser welding gun and the vertical surface of the rib plate form an angle of about 45 degrees, setting several points S0, S1..., S n 、S n+1 , so that the robot drives the line laser sensor along S0, S1..., S n 、S n+1 The scanning is performed in sequence. At point S0, the PLC sends a pulsating instruction to the line laser sensor to trigger data acquisition. The three-dimensional line structure laser vision system emits a line structure laser to scan the edge of the rib plate to obtain the surface contour image of the welded specimen.

[0035] Furthermore, in step 4, before forming line segment 1 and line segment 2 for the weld feature points on both side edges respectively, filtering processing is also included for the weld feature points.

[0036] Furthermore, in step 5, the step of fitting the line segment 1 and the line segment 2 of the weld feature and solving the intersection thereof as the coordinates of the weld center trajectory includes:

[0037] According to the known fitting function of the weld feature points, the square of the distance from all weld feature points to the straight line is minimized, and the sum of the squares of the errors from the weld feature points to the straight line is calculated, that is:

[0038]

[0039] In the formula, k and b are coefficients of the equation to be solved, z i 、x i For the collected coordinate data, we take the derivatives of k and b respectively, and we can get:

[0040]

[0041] make get:

[0042]

[0043]

[0044] Where A and B are the weld feature points in line segment 1, and C and D are the weld feature points in line segment 2. Line segment 1: z1 = k1x + b1 and line segment 2: z2 = k2x + b2 can be determined. The intersection point X of the two line segments can be obtained. A , Z A ; Perform linear interpolation on all sampling points to obtain the actual weld center trajectory coordinates.

[0045] Furthermore, after step 5, the method further includes correcting the welding trajectory based on the coordinates of the weld center trajectory.

[0046] To achieve the above-mentioned object, the present invention provides a laser welding seam detection method, comprising: collecting a laser weld fringe image of a surface of a test piece after welding; preprocessing the laser weld fringe image; the preprocessing step comprising: extracting key points of the preprocessed laser weld fringe image, wherein the step of extracting key points of the laser weld fringe image comprises:

[0047] Extract a weld contour curve, divide the original data into regions of interest, and set the weld detection starting point A and weld detection end point D;

[0048] Traverse all weld profile feature data points within the divided area of ​​interest, find the corresponding horizontal coordinate of the highest point E of the arc in the laser 2D coordinates, and use this point as a reference to search in the vertical or horizontal coordinate direction until the weld key feature inflection point B and weld key feature inflection point C are found when the distance from the base plate and the vertical plate is zero, and mark the coordinates of these points;

[0049] The least squares method is used to fit the line segments AB and CD. The intersection O of the two line segments is the center of the weld. The weld foot height difference |OB-OC| is calculated. It is determined whether the weld foot height difference |OB-OC| is less than or equal to a preset value. If it exceeds the threshold, the weld is judged to be unqualified.

[0050] Furthermore, the pre-processing step also includes: grayscale processing, noise reduction, binarization, image enhancement, and key area extraction of the graphics.

[0051] To achieve the above-mentioned objectives, the present invention provides a device for an automatic tracing method of a laser welding weld, the device comprising: a robot motion device, a robot, a weld seam measurement end effector, a work platform, and a laser; robot motion devices are provided on both sides of the work platform, and the robot is mounted on a slide of the robot motion device; the weld seam measurement end effector is mounted on the end flange of the robot; wherein the weld seam measurement end effector comprises at least an integrated line laser sensor, a wire feeding device, a laser welding gun, and a CCD module.

[0052] The above technical solution of the present invention has the following advantages:

[0053] The 3D line structure laser vision system emits a line structure laser to scan the surface of the workpiece to be measured, obtaining a contour image of the workpiece surface. After processing through an image filter and a data acquisition card, the 3D point cloud coordinates of the workpiece surface are obtained. The system automatically calculates and analyzes the center coordinates of the weld seam of the workpiece to be welded. The robot drives the sensor to scan the welding path, automatically solves the actual weld trajectory coordinates, obtains the tool coordinate position of the robot end effector through matrix transformation, and automatically corrects the robot's motion trajectory. After welding, the sensor scans the weld seam, automatically identifies the weld seam, calculates the weld leg height, and analyzes the quality of common morphologies, providing a basis for workpiece inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a general structural diagram of a device for an automatic tracking method of laser welding welds disclosed in the present invention.

[0055] Figure 2 This is a three-dimensional diagram of the weld measurement end effector disclosed in the present invention.

[0056] Figure 3 This is a front view of the weld measurement end effector disclosed in the present invention.

[0057] Figure 4 This is a side view of the weld measurement end effector disclosed in the present invention.

[0058] Figure 5 This is a schematic diagram of the welding direction disclosed in the present invention.

[0059] Figure 6 This is a diagram of the sensor measurement principle disclosed in the present invention.

[0060] Figure 7 This is a flow chart of the weld trajectory detection and post-weld morphology detection method disclosed in the present invention.

[0061] Figure 8 Create a diagram for the welding gun focal point TCP disclosed in the present invention.

[0062] Figure 9This is a schematic diagram of the TCP calibration of the welding gun disclosed in the present invention.

[0063] Figure 10 This is a schematic diagram of TCF calibration disclosed in the present invention.

[0064] Figure 11 Scanning sampling points of the test piece disclosed in the present invention.

[0065] Figure 12 This is a schematic diagram of the cross section of the T-shaped weld disclosed in the present invention.

[0066] Figure 13 This is a schematic diagram of the weld morphology after welding disclosed in the present invention.

[0067] In the figure: 001, robot motion device; 002, robot control cabinet; 003, dust removal cabinet; 004, robot; 005, wire feeder; 006, working platform; 007, water cooler; 008, laser; 009, ribbed wall panel workpiece; 010, control console; 100, weld seam measurement end effector; 101, line laser sensor; 102, sensor fixing seat; 103, wire feeding device; 104, connecting flange; 105, adapter flange; 106, coaxial shielding gas kit; 107, wire feeding head; 108, wire feeding tube; 109, laser welding gun; 110, base; 111, CCD module; 112, wire feeding fixing plate; 200, bottom plate; 201, vertical plate; 202, tooling pressure plate 1; 203, tooling pressure plate 2. DETAILED DESCRIPTION

[0068] The following embodiments of the present invention are described in further detail with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0069] In the description of the present invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like, indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or components referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

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

[0071] See also Figure 1 , which is a schematic diagram of the assembly structure of a device for an automatic tracking method of a laser welding weld provided in a preferred embodiment of the present application.

[0072] The laser welding seam automatic tracking method in this application is applied in laser welding, laser detection and other related fields. Figure 1 , Figure 1 This is a schematic diagram of the assembly structure of a device for an automatic tracking method for laser welding welds according to one embodiment of the present invention.

[0073] exist Figure 1 In the embodiment shown, the assembly structure mainly consists of two sets of weld seam measurement end effectors 100, two sets of robots 004 (six-degree-of-freedom robots), two sets of robot motion devices 001, a robot control cabinet 002, a dust removal cabinet 003, a work platform 006, a laser 008, a water cooler 007, a console 010 and measurement and control analysis software (see Figure 1 The work platform 007 is arranged in the middle of the system. The work platform 006 is fixed to the ground foundation by expansion bolts and is used to install welding tools and workpieces. Two sets of robot motion devices 001 are symmetrically arranged on both sides of the work platform 007. Two sets of robots 004 are respectively installed on the slides of the two sets of robot motion devices 001. The robot motion devices 001 serve as the external axes of the robots 004, expanding the motion processing range of the robots 004. Two sets of weld seam measurement end effectors 100 are respectively installed on the end flanges of the robots 004, symmetrical on the left and right sides, see, 2. Figure 3 Two laser welding guns 109 are integrated and mounted on two weld seam measurement end effectors 100. Two lasers 008 are connected to the laser welding guns 109 via optical fibers, providing them with welding laser energy. Each laser 008 is equipped with a water chiller 007, which is placed next to the laser 008. The lasers 008 and water chiller 007 are arranged at the inner end of the system, with the moving parts and the stationary parts separated. The console 010 is placed outside the system to facilitate operator control.

[0074] like Figure 1-Figure 4The weld seam measurement end effector 100 consists of a line laser sensor 101, a sensor fixing base 102, a wire feeding device 103, a connecting flange 104, an adapter flange 105, a coaxial shielding gas kit 106, a wire feeding head 107, a wire feeding tube 108, a laser welding gun 109, a base 110, a CCD module 111 and a wire feeding fixing plate 112.

[0075] The connection relationship of the weld seam detection end effector 100 will be described in detail below.

[0076] The connecting flange 104 is connected to the sensor fixing base 102 by screws, and the adapter flange 105 is installed on the top of the connecting flange 104 and docked with the end flange of the robot 004. The adapter flange 105 can be designed to have different sizes according to different robot interfaces for easy installation. The base 110 is connected and fixed to the connecting flange 104 by screws, and the laser welding gun 109 is installed on the base 110. The coaxial shielding gas kit 106 is installed at the lower end of the laser welding gun 109. The coaxial shielding gas kit 106 has an air inlet on the side, which is used to connect the shielding gas to provide gas protection for welding and improve the surface forming quality of the parts.

[0077] The CCD module 111 is installed at the top of the laser welding gun 109. The CCD module 11 is a high-definition camera. The weld can be clearly seen through the internal lens group. A filter lens is configured to monitor the laser welding process.

[0078] The QBH interface 112 is installed on the side of the laser welding gun 109 and connected to the optical fiber. The other end of the optical fiber is connected to the laser 008. The sensor fixing seat 102 is installed on the base 110. The line laser sensor 101 is installed on the sensor fixing seat 102. The wire feeding fixing plate 112 is installed on the side of the base 110. The wire feeding device 103 is installed on the wire feeding fixing plate 112. One end of the wire feeding device 103 is connected to the wire feeder 005 through the wire feeding tube 108, and the other end is connected to the wire feeding head 107 through the wire feeding tube 108. The wire feeding head 107 is fixed on the coaxial gas shielding kit 106. The wire feeding device 103 is the power device of the wire feeding system, which can realize the forward and backward movement of the welding wire. The welding wire reel is installed inside the wire feeder 005. The welding direction is that the line laser sensor 101 is in front, the welding wire is in the middle, and the laser welding gun 109 is at the back (such as Figure 5 As shown in the figure). The line laser sensor 101, wire feeder 103 and laser welding gun 109 are integrated together and mounted on the flange end of robot 003 through flange, with symmetrical arrangement on both sides. The system first performs weld position tracking and automatically corrects the robot welding trajectory.

[0079] The measuring laser axis is parallel to the center of the welding gun. During measurement, the center of the welding gun and the reinforcement (vertical plate) form an angle of about 45°. The welding gun is appropriately retracted along the Z axis to avoid interference with the tooling. During welding, the welding gun is extended along the Z axis so that the welding gun head is in the welding position with the welding wire, and the welding gun forms an angle of 60° with the reinforcement.

[0080] The weld seam tracking system based on a line laser sensor is mainly composed of a PLC, an industrial computer, a data acquisition card, a motion controller, a robot system, and a communication module. The PLC communicates with the robot system and the industrial computer via Modbus-TCP. The robot and laser are controlled by the PLC for I / O respectively. The main console is connected to the left and right measurement welding systems via the EtherCat interface, thereby realizing the coordinated control of the dual-robot system.

[0081] The line structure laser vision system consists of laser driver, line structure laser, image sensor, image filter, image processor, control logic board, etc. (such as Figure 6 As shown in the figure, a 3D line-structured laser vision system emits a line-structured laser to scan the surface of the workpiece being measured (i.e., the test piece to be welded), generating a contour image of the workpiece surface. This image is then processed using an image filter and a data acquisition card to generate a 3D point cloud coordinate of the workpiece surface. The system automatically calculates and analyzes the center coordinates of the weld seam of the workpiece to be welded. A robot drives the sensor to scan the welding path, automatically calculating the actual weld trajectory coordinates and using matrix transformation to determine the tool coordinates of the robot's end effector. This automatically corrects the robot's motion trajectory. After welding, the sensor scans the weld seam, automatically identifying the weld, calculating the weld leg height, and analyzing common topography quality, providing a basis for workpiece inspection.

[0082] Those skilled in the art will understand that Figures 1-6 The assembly structure of the device for an automatic tracking method for laser welding welds shown in the figure does not constitute a limitation on the device. The device may include more or fewer components than shown in the figure, and some components are not necessary components of the device. They can be omitted or combined as needed without changing the essence of the invention.

[0083] This embodiment develops an automatic recognition system for the weld trajectory of a T-shaped structure in laser welding based on a customized high-precision line structured light sensor with a large field of view and long depth of field. It realizes autonomous planning of the weld path program in the dual-beam laser welding process, replaces manual teaching, and develops a post-weld weld morphology analysis system to form a complete system unit with hardware and software, meeting product quality requirements and realizing the engineering application of automatic laser welding of T-shaped welds on titanium alloy wall panels.

[0084] like Figure 7-Figure 9 As shown, an embodiment of the present invention provides a method for automatically tracking a laser welding seam, comprising the following steps:

[0085] Step 1: The laser welding gun tool coordinate system is established, which is also the tool center point (TCP) calibration.

[0086] Install a focus tool at the end of the laser welding gun, determine the reference point of the focus tool, fix the tool tip tool on the workbench, determine the fixed point of the tool tip tool, use the four-point method to make the reference point just touch the fixed point, and establish the laser welding gun focus coordinate system through the data of the four position points.

[0087] It can be understood that the focal length position of the laser welding gun is first determined, and the tool coordinate system TCP is established based on this position. The focus tool is installed on the end of the laser welding gun, and the distance from the tip of the focus to the center of the focusing mirror of the laser welding gun is the focal length of the welding gun (such as Figure 8 shown).

[0088] The calibration tool tip tool is fixed on the workbench, the tip of the calibration tool tip tool is used as the reference point, and the four-point method is used to calibrate the laser welding gun focus coordinate system TCP (tool center point).

[0089] Define the base coordinate system {B}, the end flange coordinate system {F} and the tool coordinate system {T}. The pose relationship between the end flange coordinate system {F} and the base coordinate system {B} can be expressed using the matrix express,

[0090]

[0091] in, is the rotation matrix of the end flange coordinate system {F} relative to the base coordinate system {B}, which contains three position vectors They represent the direction cosines of the three unit principal vectors of {F} relative to {B}, is the position vector of the origin of {F} relative to {B}, It can be obtained from the direct solution of robot kinematics. represents the transformation matrix of {T} relative to the base coordinates {B}, Represents the transformation matrix of the tool coordinate {T} relative to the end flange coordinate {F}. Since the tool is installed on the robot flange, It is a fixed value, and the tool TCP calibration is determined Parameters, and F P tcp .

[0092]

[0093] Calibration process: (a) tool center point (TCP) calibration; (b) tool coordinate system pose (TCF) calibration.

[0094] like Figure 9 , Tool Center Point (TCP) calibration: Manually operate the robot to move the laser welding gun from four different directions to a fixed point, record the four points P1, P2, P3, and P4, and the robot end flange posture data (X, Y, Z, θ1, θ2, θ3). The coordinates of the four tool center points in the world coordinate system are equal. The calculation is completed on the operation teach pendant to obtain the tool coordinate system TCP (X, Y, Z).

[0095] like Figure 10 , Tool Coordinate System Posture (TCF) calibration: Move the tool coordinate system TCP to any fixed point for measurement, move a point in the negative direction of the Y axis to the fixed point for measurement, and then move any point in the XY plane with a negative X value to the fixed point for measurement, record the data, and complete the calculation calibration on the operation teach pendant.

[0096] Step 2: Line laser sensor coordinate calibration;

[0097] A robot base coordinate system is established based on the installation base of the line laser sensor installed on the robot. The coordinate relationship between the robot base coordinate system and the laser welding gun focus coordinate system is determined according to the posture matrix of the robot base coordinate system.

[0098] Specifically, the coordinate calibration of the line laser sensor is mainly to determine the coordinate relationship between the line laser sensor coordinate system and the laser welding gun focus. When the robot is in a certain posture, the posture matrix of the weld seam measurement end effector relative to the robot base coordinate is T1, the posture matrix of the line sensor coordinate system to the end flange is X1, and the posture of the world coordinate system in the robot base coordinate system is P1=T1X1M1. When the robot moves to another posture, the above parameters become T2, X2, and P2. Since the line sensor is fixed on the end, X1=X2. Since the world coordinate system and the end coordinate system are stationary, P1=P2. Let Then AX=XB,

[0099] Assume A is an m×n matrix, B is an n×n matrix, and X is an m×n matrix. From the definition of matrix direct product, we know that:

[0100]

[0101] When A or B is the identity matrix, we have

[0102]

[0103] Let a and b be constants, then we have,

[0104] vec(aA+bB)=a×vec(A)+b×vec(B)

[0105] Assumptions

[0106]

[0107] Then AX=XB can be expressed as:

[0108]

[0109]

[0110] To determine the unique solution of the above equation, two sets of motions with non-parallel rotation axes are required. The simultaneous equations obtained from these two sets of motions are:

[0111]

[0112]

[0113] Use the least squares method to solve the eigenvector v=[v1 v2 … v 13 ], then the relationship matrix is:

[0114]

[0115] Step 3: Scanning program planning;

[0116] The robot drives the line laser sensor to move along the welding position of the test piece to be welded, scans the edge of the rib plate, and obtains a surface contour image of the test piece to be welded.

[0117] In this step, first write the robot measurement program in the offline programming software so that the YZ plane of the laser welding gun forms an angle of about 45 degrees with the vertical surface of the rib plate ( Figure 11 As shown), several points S0, S1, S n 、S n+1 , so that the robot drives the line laser sensor along S0, S1..., S n 、S n+1 The scanning is performed in sequence. At point S0, the PLC sends a pulse command to the line laser sensor to trigger data acquisition. The three-dimensional line structure laser vision system emits a line structure laser to scan the edge of the rib plate to obtain the surface contour image of the welded specimen, and then obtain the three-dimensional point cloud coordinates of the welded specimen surface. The line laser sensor moves to position S1, the measurement command stops, and the line laser sensor returns the position information of the rib plate edge. The line laser sensor moves to S2~S n+1 Position, PLC sends out pulse measurement instruction to obtain each sampling point S i3D point cloud coordinate information, each sampling point pauses for 1 to 3 seconds to reduce the loss of measurement accuracy caused by robot movement, the line laser sensor moves to S n Position, PLC trigger pulse measurement instruction, line laser sensor motion S n+1 Position, acquisition instruction stops, and the line laser sensor returns the edge information of the vertical rib plate.

[0118] Step 4: Weld feature recognition;

[0119] From the obtained surface contour image of the welded specimen, the weld feature points on both sides of the laser band are extracted, and line segments 1 and 2 are drawn for the weld feature points on both sides of the edges, and the drawn line segments 1 and 2 are used as weld features.

[0120] like Figure 12 As shown, taking a single side of a T-shaped weld as an example, the vertical plate 201 is placed vertically on the bottom plate 200 under the action of the tooling pressure plate 1 202 and the tooling pressure plate 2 203. The tooling features are excluded, and the line segments AB and CD are taken as the weld features.

[0121] Step 5: Calculate the center coordinates of the T-type weld trajectory;

[0122] The line segment 1 and the line segment 2 of the weld feature are fitted, and the intersection point is the coordinate of the weld center trajectory.

[0123] Continue to refer to Figure 12 , use the least squares method to fit line segments AB and CD respectively, that is, according to the point fitting function, make the square of the distance from all points to the straight line minimum, and calculate the sum of the squares of the errors from the points to the straight line, that is:

[0124]

[0125] In the formula, k and b are coefficients of the equation to be solved, z i 、x i For the collected coordinate data, we take the derivatives of k and b respectively, and we can get:

[0126]

[0127] make

[0128] get:

[0129]

[0130]

[0131] Where A and B are the weld feature points in line segment 1, and C and D are the weld feature points in line segment 2. Line segment 1: z1 = k1x + b1 and line segment 2: z2 = k2x + b2 can be determined. The intersection point X of the two line segments can be obtained. A , Z A ; Perform linear interpolation on all sampling points to obtain the actual weld center trajectory coordinates.

[0132] Step 6: Robot trajectory correction;

[0133] On the basis of calibrating the welding trajectory, the robot applies the trajectory correction data obtained through measurement and calculation on the tool coordinate system TCP, calls the trajectory offset instruction, and realizes the correction of the welding trajectory.

[0134] The Modbus-TCP communication function is mainly used for communication between the weld positioning software and the robot. The data sent by the weld positioning software mainly includes the XYZ coordinate offset of the trajectory correction, the measurement and calculation completion flag, etc.; the data sent by the robot includes the robot's position and the current point number. Using a distributed transmission method, after each point is measured, the correction data is sent and stored in a specific register of the robot system, and is uniformly called when executing the welding program to ensure that the communication transmission bandwidth and rate will not affect the welding process.

[0135] Step 7: Robot automatic welding;

[0136] According to the process requirements, add signal instructions such as light output, attenuation, and wire output, add dual-beam synchronous welding signals, test run the welding program, disconnect the laser and wire feeder, check and guide the laser welding gun spot trajectory, run the welding program, and complete the part welding.

[0137] After completing steps 1-7 above, post-weld inspection is necessary because dual-beam welding welds must meet Class I weld standards, primarily including internal weld quality, mechanical properties, and appearance requirements. Currently, X-ray flaw detection is the primary method for detecting internal weld defects. Common defects include cracks, incomplete penetration, lack of fusion, porosity, and slag inclusions. Mechanical properties are primarily assessed using tensile testing, with tensile and shear strengths meeting technical requirements. Appearance quality requirements include uniform weld transitions, the absence of pits, weld bumps, or undercuts at arc start and end points, and a weld leg height difference of ≤0.3mm. Appearance quality is a key evaluation criterion for dual-beam welding process quality. Currently, manual visual inspection for surface defects is the primary method, with weld leg height measured using metallographic methods. Metallographic measurement of weld leg height requires cutting the T-shaped weld to produce metallographic test specimens, a method that is inefficient and unable to inspect the welded workpiece.

[0138] Therefore, this embodiment also provides a laser welding weld inspection method that uses laser vision to detect the height difference of the weld legs of T-shaped welds and performs online detection of common morphological defects. This method is fast, accurate, and effective, improving the level of weld quality evaluation. The method includes the following steps 8-9, specifically as follows:

[0139] Step 8: Scan the weld after welding;

[0140] Before testing, a line laser sensor collects three-dimensional surface data of the welded specimen and transmits the three-dimensional surface data to an industrial computer via Ethernet.

[0141] Step 9: Weld Quality Analysis

[0142] After the quality online detection system software obtains the line laser sensor data, it performs image preprocessing on it, and then converts the filtered data into grayscale images and three-dimensional point cloud images, displays the 2D / 3D images on the software interface, and can display output and weld morphology data.

[0143] Among them, the preprocessing of the laser weld stripe image includes the following methods, but is not limited to: grayscale processing, noise reduction, binarization, image enhancement, and key area extraction, among which:

[0144] Grayscale processing uses weighted average method to grayscale the original image to maintain the original image shape and improve measurement accuracy.

[0145] Filtering and noise reduction: Line laser sensors are affected by environmental disturbances and light during data image acquisition, which hinders image processing and analysis. Therefore, a suitable filtering algorithm is used for filtering and noise reduction. Referring to the weld positioning filtering method, a convolution template is used for Gaussian filtering to eliminate the influence of noise.

[0146] Image binarization is to reduce the amount of image data and improve computational efficiency. The maximum space method is used for binarization. The grayscale range of D1 is [0, g], accounting for the image proportion m0, and the average grayscale value is p0. The grayscale range of D2 is [g+1, H], accounting for the image proportion m1, and the average grayscale value is p1. The total average grayscale value of the image is p, s is the variance of D1 and D2, the number of pixels with grayscale values ​​less than g is K0, and the number of pixels with grayscale values ​​greater than g is K1. Then:

[0147] s=m0m1(p0-p1) 2

[0148] The maximum inter-class variance method is based on the threshold g corresponding to the maximum value s in the image grayscale distribution. The weld image after noise reduction is binarized and the threshold is automatically obtained. This not only retains the edge information of the original image, but also reduces the number of fractures.

[0149] Image enhancement, when performing corrosion operation on the original image, the target image gradually shrinks and some areas become blurred, which may cause image distortion. Image enhancement processing, performing closing operation on the image, enhances the effective part of the image and facilitates the extraction of key areas.

[0150] Extract key image areas and select valid line segments AB and CD to reduce the number of operations and improve efficiency.

[0151] Extraction of key points of weld images. After obtaining the weld laser image, it is necessary to further extract the key points in the image in order to detect and judge the weld morphology quality. The extraction of key points is as follows: Figure 13 As shown, the following steps are included:

[0152] Extract a weld contour curve, divide the original data into regions of interest, and set the weld detection starting point A and weld detection end point D;

[0153] Traverse all weld profile feature data points within the divided area of ​​interest, find the corresponding horizontal coordinate of the highest point E of the arc in the laser 2D coordinates, and use this point as a reference to search in the vertical or horizontal coordinate direction until the weld key feature inflection point B and weld key feature inflection point C are found when the distance from the base plate and the vertical plate is zero, and mark the coordinates of these points;

[0154] The least squares method is used to fit line segments AB and CD. The intersection point O of the two segments is the weld center. The weld leg height difference |OB-OC| is calculated. It is determined whether this difference is less than or equal to a preset value of 0.3mm. If it exceeds the threshold, the weld is deemed unqualified and rework is performed. If it passes, the next step is carried out.

[0155] Similarly, the high points and concave points on the BEC curve can be judged, and a threshold can be set. If the threshold is exceeded, it is judged as unqualified.

[0156] In the description and claims of this application, the words "include / comprise" and the words "have / include" and their variations are used to specify the existence of stated features, values, steps or components, but do not exclude the existence or addition of one or more other features, values, steps, components or their combinations.

[0157] Some features of the present invention are described in separate embodiments for clarity of explanation, however, these features may also be described in combination in a single embodiment. Conversely, some features of the present invention are described in a single embodiment for brevity, however, these features may also be described in different embodiments individually or in any suitable combination.

[0158] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A laser welding seam automatic tracking method, characterized in that: The following steps are involved: Step 1: Install a focus tool at the end of the laser welding gun, determine the reference point of the focus tool, fix the tool tip on the workbench, determine the fixed point of the tool tip, use the four-point method to make the reference point just touch the fixed point, and establish the laser welding gun focus coordinate system based on the data of the four position points; Step 2: With the line laser sensor installed on the robot's mounting base as a reference, establish a robot base coordinate system, and determine the coordinate relationship between the robot base coordinate system and the laser welding gun focus coordinate system according to the pose matrix of the robot base coordinate system; Step 3: The robot drives the line laser sensor to move along the welding position of the welded specimen, scans the edge of the rib plate, and obtains a surface contour image of the welded specimen; Step 4: Extracting weld feature points on both sides of the laser band from the obtained surface contour image of the welded specimen, and creating line segments 1 and 2 for the weld feature points on both sides of the laser band, respectively, with the created line segments 1 and 2 being weld features; Step 5: Fit the line segment 1 and the line segment 2 of the weld feature, and solve their intersection point to obtain the coordinates of the weld center trajectory; In step 2, the step of determining the coordinate relationship between the robot base coordinate system and the laser welding gun focus coordinate system includes: When the robot is in a certain posture, the posture matrix of the weld seam measurement end effector relative to the robot base coordinate is T1, the posture matrix of the line sensor coordinate system to the end flange is X1, and the posture of the world coordinate system in the robot base coordinate system is P1=T1X1M1. When the robot moves to another posture, the above parameters become T2, X2, and P2. Since the line sensor is fixed on the end, X1=X2. Since the world coordinate system and the end coordinate system are stationary, P1=P2. Let , , then AX=XB, Assume A is an m×n matrix, B is an n×n matrix, and X is an m×n matrix. From the definition of matrix direct product, we know that: When A or B is the identity matrix, we have Let a and b be constants, then we have, vec(aA+bB)=a×vec(A)+b×vec(B) Assumptions Then AX=XB can be expressed as: To determine the unique solution of the above equation, two sets of motions with non-parallel rotation axes are required. The simultaneous equations obtained from these two sets of motions are: Use the least squares method to solve the eigenvector v=[v1 v2 … v 13 ], then the relationship matrix is: 。 2. The laser welding seam automatic tracking method according to claim 1, characterized in that: In step 1, the calibration process for establishing the laser welding gun focus coordinate system is as follows: Tool center point position calibration: Manually operate the robot to move the laser welding gun from four different directions to a fixed point, record the four points P1, P2, P3, and P4, and the robot end flange pose data (X, Y, Z, θ1, θ2, θ3). The coordinates of the four tool center points in the world coordinate system are equal. Calculation is completed on the teaching pendant to obtain the tool coordinate system TCP (X, Y, Z). Tool coordinate system posture calibration: Move the tool coordinate system TCP to any fixed point for measurement, move a point in the negative direction of the Y axis to the fixed point for measurement, then move any point in the XY plane with a negative X value to the fixed point for measurement, record the data, and complete the calculation calibration on the teaching pendant.

3. The laser welding seam automatic tracking method according to claim 1, characterized in that: The steps of step 3 include making the YZ plane of the laser welding gun and the vertical surface of the rib plate form an angle of about 45 degrees, setting several points S0, S1..., S1 on the welded specimen. n 、S n+1 , so that the robot drives the line laser sensor along S0, S1..., S n 、S n+1 The scanning is performed in sequence. At point S0, the PLC sends a pulsating instruction to the line laser sensor to trigger data acquisition. The three-dimensional line structure laser vision system emits a line structure laser to scan the edge of the rib plate to obtain the surface contour image of the welded specimen.

4. The laser welding seam automatic tracking method according to claim 1, characterized in that: In step 4, before forming line segment 1 and line segment 2 for the weld feature points on both side edges, filtering processing is also performed on the weld feature points.

5. The laser welding seam automatic tracking method according to claim 1, characterized in that: In step 5, the steps of fitting the line segment 1 and the line segment 2 of the weld feature and solving the intersection thereof as the coordinates of the weld center trajectory include: According to the known fitting function of the weld feature points, the square of the distance from all weld feature points to the straight line is minimized, and the sum of the squares of the errors from the weld feature points to the straight line is calculated, that is: In the formula, k and b are coefficients of the equation to be solved, z i 、x i For the collected coordinate data, we take the derivatives of k and b respectively, and we can get: make , get: Where A and B are the weld feature points in line segment 1, and C and D are the weld feature points in line segment 2. Line segment 1: z1 = k1x + b1 and line segment 2: z2 = k2x + b2 can be determined. The intersection point X of the two line segments can be obtained. A , Z A ; Perform linear interpolation on all sampling points to obtain the actual weld center trajectory coordinates.

6. The laser welding seam automatic tracking method according to claim 1, characterized in that: After step 5, the method further includes correcting the welding trajectory based on the coordinates of the weld center trajectory.

7. A device for the automatic tracking method of laser welding seams according to any one of claims 1 to 6, characterized in that: The device includes: a robot motion device, a robot, a weld seam measurement end effector, a work platform, and a laser; the robot motion device is provided on both sides of the work platform, and the robot is mounted on a slide of the robot motion device; the weld seam measurement end effector is mounted on the end flange of the robot; wherein the weld seam measurement end effector includes at least an integrated line laser sensor, a wire feeding device, a laser welding gun, and a CCD module.

Images