Laser welding method

Abstract

The present invention discloses a laser welding method, which is based on a laser welding system with a visual detection device. The laser welding method includes the following steps: obtaining weld information in front of a laser welding head through the visual detection device; obtaining welding process parameters through a preset welding process library according to the obtained weld information; generating a welding control vector according to the weld information and the welding process parameters; controlling the laser welding head to perform laser welding according to the control vector; quickly responding to detected weld features, realizing real-time laser energy and laser energy density distribution through the control vector, and adjusting the welding process in real time. There is no need to pre-run and record all weld information, calculate the motion trajectory, and then run in a repeated motion mode, and welding can be performed directly; the welding guidance time is compressed by dozens or hundreds of times.

Description

Technical Field

[0001] The present invention relates to the technical field of laser welding, and in particular to a laser welding method. Background Art

[0002] Laser welding involves focusing a laser beam onto the weld location, while simultaneously moving the focused spot through an auxiliary welding device to form the weld. This concentrated and efficient application of energy creates a weld with a high aspect ratio, but the weld width typically does not exceed 0.4 mm, resulting in poor weldability.

[0003] The current technical solutions to address poor laser welding suitability mainly use optical methods (adjusting the energy distribution of the focused spot) or electromechanical methods (using an oscillating focused welding head instead of a fixed-point focused welding head to adjust the distribution of the focused spot energy in the time domain) to obtain a wider weld seam of 0.5 to 0.9 mm to improve welding suitability. However, in actual welding applications, irregular weld gaps due to processing or tooling are the norm. In order to reduce the welding defect rate, the width of the entire weld seam has to be expanded, thereby increasing the welding heat input several times or even dozens of times, losing the main advantage of laser welding.

[0004] Another technical solution is to mount the laser welding head on a welding assist device with vision guidance (such as a multi-axis robot) to achieve better welding suitability. However, the dynamic performance of commonly used welding assist devices cannot meet the high precision requirements of laser welding, and they can only compromise by expanding the width of the entire weld seam. The low dynamic performance also requires a long welding guidance time, making it impossible to respond to localized welding thermal deformation. This requires the design of complex fixtures or further expansion of the weld seam width to address this issue.

[0005] In summary, existing technical solutions often fail to fully utilize the advantages of concentrated and efficient application of laser welding energy. In practice, the following phenomena often occur: 1. Poor weld quality or even false welds; 2. Poor molten pool stability, prone to large amounts of spatter and pores; 3. Severe burnout of alloy elements during welding.

[0006] In the field of laser welding, traditional welding auxiliary devices are not capable of meeting the high-precision and high-dynamic performance requirements. Therefore, the existing technology has introduced a welding head with a front visual inspection device and a welding auxiliary device to form a secondary system. We generally refer to this combination of welding head + auxiliary motion device as flying welding.

[0007] The so-called flight trajectory is an abstract model of the welding auxiliary device (i.e., the trajectory of the welding head) during the flight welding process. To distinguish it from the XYZ visual coordinate system below, the three-dimensional space coordinates are indicated by UVW, and the advance angle is indicated by θ. It should be noted that the flight trajectory is based on an absolute coordinate system, that is, it is obtained from an observation point outside the system.

[0008] The flight trajectory is generally based on the ideal weld, and most of the time it needs to be optimized and generated based on the kinematic characteristics of the welding auxiliary device. Figure 8 As shown in the figure, for example, in an abstract "pipe-to-pipe welding" application, the ideal weld is a circular ring between the circular pipes. Therefore, the motion trajectory of the welding head should be a circle perpendicular to the center axis of the pipe, and its radius is the pipe radius + the welding head working distance. There are three solutions:

[0009] The first solution is to use a six-axis robot with a welding head to move along a circular trajectory while the workpiece is fixed. In this case, the robot controller needs to be synchronized with the welding head, or only the data or even instructions of each axis of the robot can be synchronized to calculate [F(t)] in real time:

[0010] [F(t)]: {[u(t), v(t), w(t), θ(t)]}.

[0011] The second solution is to use a positioner based on the ideal weld, which is the circular trajectory. When working, u(t), v(t), and w(t) are calculated and interpolated from the measured θ(t). The obtained [F(t)] has high accuracy and low calculation pressure.

[0012] The third approach is to keep the welding head stationary while rotating the pipe to be welded. This relative motion also forms a circular trajectory perpendicular to the pipe's centerline. [F(t)] is also calculated and interpolated from the measured θ(t) during operation. Summary of the Invention

[0013] In view of the above-mentioned shortcomings currently existing, the present invention provides a laser welding method that can achieve optimal configuration of laser energy at a microscopic level for heterogeneous joint features.

[0014] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0015] A method for generating a laser welding control vector is provided, based on a laser welding system having a visual detection device, and the method for generating a laser welding control vector comprises the following steps:

[0016] The weld information before the laser welding head is obtained through the visual inspection device;

[0017] Obtain welding process parameters through a preset welding process library based on the obtained weld information;

[0018] The welding control vector is generated according to the weld information and welding process parameters.

[0019] According to one aspect of the present invention, the control vector is a real-time laser energy and laser energy density distribution method based on a welding head coordinate system and generated according to weld information.

[0020] According to one aspect of the present invention, the control vector is:

[0021] [W(t)]:{[P xyz (t)],[LP(t),L c (t)]}

[0022] Among them, P xyz (t) is the laser energy delivery coordinate (x, y, z) based on the welding head coordinate system at time t; LP(t) is the laser output power at time t, L c (t) is the modeling of the laser action mode.

[0023] According to one aspect of the present invention, the weld information includes recognized weld features, and the weld features C(t) are data matrices obtained by recognizing images collected in real time by a visual camera.

[0024] According to one aspect of the present invention, the welding process library is generated by preset parameters including at least joint type, material coefficient, swing reference, and weld morphology.

[0025] According to one aspect of the present invention, it is characterized in that the control vector is used to achieve real-time and precise delivery of laser energy.

[0026] A laser welding method, based on a laser welding system with a visual inspection device, comprises the following steps:

[0027] The weld information before the laser welding head is obtained through the visual inspection device;

[0028] Obtain welding process parameters through a preset welding process library based on the obtained weld information;

[0029] Generate welding control vector according to weld information and welding process parameters;

[0030] The laser welding head is controlled to perform laser welding according to the control vector, specifically, the laser welding head is controlled to perform laser welding according to the control vector corrected in real time.

[0031] According to one aspect of the present invention, the laser welding method includes: obtaining a welding motion trajectory F(t) according to weld information, wherein the welding motion trajectory F(t) is preset or generated in real time according to the weld information.

[0032] According to one aspect of the present invention, at any time t, the welding control vector W(t) is obtained by real-time adjustment based on the real-time collected joint features and the synchronously obtained welding motion trajectory F(t), combined with the obtained welding process parameters.

[0033] According to one aspect of the present invention, the control vector is expressed as:

[0034] W(t)=Φ KB ({[F(τ)]|τ<t},{[B][A] -1 [C(τ)]|τ<tt l})

[0035] Where [A] is the camera installation position, [B] is the welding head installation position, t l is the welding guidance time, Φ KB is the welding process parameter.

[0036] According to one aspect of the present invention, at time t, {[F(τ)]|τ<t} is the entire flight trajectory data before time t, {[B][A] -1 [C(τ)]|τ<tt l} is tt l The generation mechanism of the control vector is based on the continuous calculation of these two sets of input values ​​by the welding process.

[0037] According to one aspect of the present invention, the weld information includes recognized weld features, and the weld features C(t) are data matrices obtained by recognizing images collected in real time by a visual camera.

[0038] Advantages of the present invention:

[0039] The laser welding method described in this invention quickly responds to detected weld characteristics, achieves real-time laser energy and laser energy density distribution through vector control, and simultaneously adjusts the welding process in real time to directly perform welding. It eliminates the need to pre-run and record all weld information, calculate the motion trajectory, and then re-run in a repetitive motion mode, allowing direct welding. This shortens welding guidance time by tens or even hundreds of times. This brings three advantages:

[0040] 1. Reduce unnecessary welding heat input;

[0041] 2. Reduce the welding defect rate caused by irregular gaps;

[0042] 3. Reduce the welding defect rate caused by local thermal deformation during welding.

[0043] Compared to the first solution in the background art, the laser welding method described in this invention eliminates the need for the robot to compensate for guidance signals. Instead, the robot follows a preset trajectory and transmits real-time position data F(t) to the welding head for real-time calculation and compensation. This significantly reduces the demands placed on the robot and is therefore suitable for existing robots, which lack the ability to perform real-time small-step trajectory changes while carrying a load after reaching a certain speed.

[0044] Compared with the second solution of the background technology, the laser welding method described in the invention can work well for relatively regular welds as long as the error between the actual weld and the positioner trajectory is not so large as to exceed the measuring range.

[0045] Compared with the second solution of the background technology, when using the laser welding method described in the invention, when the welding speed is required to be very fast, the first two solutions are not easy to implement, but it is feasible to fix the pipe with the welding head and rotate it by itself. In terms of technical performance, the implicit requirement is that the welding head must be able to respond to changes in weld details in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 This is a schematic diagram of a method for generating a laser welding control vector according to the present invention;

[0048] Figure 2 This is a schematic diagram of common seam types described in the present invention;

[0049] Figure 3 This is a schematic diagram of the long tailor-welded seam according to the present invention;

[0050] Figure 4 This is a schematic diagram of the welding movement process according to the present invention;

[0051] Figure 5 A schematic diagram of generating a control vector according to the present invention;

[0052] Figure 6 This is a schematic diagram of a laser welding method according to the present invention;

[0053] Figure 7 It is a schematic diagram of flying welding in the background technology;

[0054] Figure 8 Schematic diagram of the laser welding system described in the present invention. DETAILED DESCRIPTION

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0056] like Figures 1 to 5 As shown, a laser welding control vector generation method is based on a laser welding system with a visual detection device, and the laser welding control vector generation method includes the following steps:

[0057] Step S1: obtaining weld information before the laser welding head through a visual inspection device;

[0058] like Figure 8 As shown, the laser welding system with a visual detection device includes a laser welding head (i.e., a laser), an optical positioning device, an optical positioning device controller, a machine vision processor, and a controller for controlling the movement of the laser welding head. The optical positioning device described in this embodiment uses an optical scanning galvanometer. In other embodiments, other devices based on refraction and reflection of light to achieve high-speed positioning can also be used.

[0059] In this embodiment, a camera captures the weld image in front of the machine welding head, and a machine vision processor recognizes the image to obtain weld information, which includes but is not limited to weld characteristics, the material of the object to be welded, and other information. The weld characteristics are:

[0060] [C(t)]:{[x n (t), y n (t), z n (t)]|1≤n≤s type}

[0061] The identification of seam features can be performed using common weld CCD image recognition methods, or by extracting weld image features using image block selection. As long as the seam features can be identified, the weld characteristics will be sufficient. Information about the material of the welded objects can be obtained through input or image recognition feature comparison. Different materials require different welding processes. For example, aluminum and iron have different melting points, requiring different laser energies. Similarly, the laser welding processes for aluminum alloys and stainless steel are completely different.

[0062] In practical applications, the above acquisition process is performed continuously in real time. As the laser welding head moves, the weld seam information in front of the laser welding head is acquired in real time. The seam feature C(t) is a data matrix obtained by recognizing the image captured in real time by the visual camera.

[0063] This embodiment combines practical applications to redefine S type ,like Figure 2 As shown in the figure, it is a common type of joint. Taking the second one in the second row as an example, it is a typical V-groove weld, so its S is defined. type=4, then the single set of measurement data C(t) is a 3x4 matrix. The subscripts 1, 2, 3, and 4 represent the four feature points from left to right. The machine vision camera collects data in the following format in real time:

[0064]

[0065] Step S2: obtaining welding process parameters through a preset welding process library according to the obtained weld information;

[0066] A welding process library is preset, and the welding process library is generated by preset parameters including at least joint type, material coefficient, swing reference, and weld morphology. Through the welding process library, welding process parameters of different materials under different joint types and morphologies can be obtained.

[0067] Step S3: Generate a welding control vector according to the weld information and welding process parameters.

[0068] The control vectors are generated based on weld information and are used to distribute laser energy and laser energy density in real time within the weld head coordinate system. This allows for precise, real-time laser energy delivery, minimizing unnecessary heat input and lowering the probability of weld failures caused by localized thermal deformation during welding.

[0069] In practical applications, the control vector can be expressed as:

[0070] [W(t)]:{[P xyz (t)],[LP(t),L c (t)]}

[0071] Among them, P xyz (t) is the laser energy delivery coordinate (x, y, z) based on the welding head coordinate system at time t, which is the real-time coordinate obtained based on the welding motion trajectory F(t) and the real-time weld feature modification; LP(t) is the laser output power at time t, L c (t) is the modeling of the laser action mode.

[0072] P(t) represents the laser energy delivery coordinates (x, y, z) at time t, based on the weld head coordinate system. x and y primarily address weld position offset and weld bead construction, while z primarily addresses height deviation (such as unevenness and rotational eccentricity) and power density control. LP(t) represents the laser output power at time t. Lc(t) models the laser's action mode. For example, a CW laser with stable power output would have a value of 1. For CW lasers requiring fluctuating power due to welding processes, a (0, 1) rise / fall model is constructed based on the laser's technical characteristics.

[0073] Control vector W(t) = Φ KB({[F(τ)]|τ<t},{[B][A] -1 [C(τ)]|τ<tt l The generation mechanism of}) is as follows:

[0074] Where [A] is the camera installation position, [B] is the welding head installation position, t l is the welding guidance time, Φ KB is the welding process parameter.

[0075] The machine vision camera and welding head are fixedly installed on the auxiliary device. The weld detection point is ahead of the welding point in the flight direction. Here, a parameter t is introduced. l At time t, {[F(τ)]|τ<t} is the entire flight trajectory data before time t, and the welding motion trajectory F(t) is a pre-set fixed motion trajectory, or a motion trajectory generated in real time based on the weld information, {[B][A] -1 [C(τ)]|τ<tt l The generation mechanism of the control vector is based on the continuous calculation of the two sets of input values ​​by KB (preset process).

[0076] like Figures 3 to 5 As shown, with an S type =4 is an example of a 304 stainless steel long tailor weld with 4 characteristic points:

[0077] like Figure 3 As shown, (1)(2) and (3)(4) form a height difference of 0.25+ / -0.2mm, and (2)(3) form a gap error of 0.6+ / -0.25mm;

[0078] like Figure 4 As shown, during the welding process, the machine moves along the slide rail with the welding head, and the weld seam keeps changing on a small scale;

[0079] like Figure 5 As shown in the figure, the preset process obtained from the preset welding process library is: CW laser output, preset power of 1800W, base welding speed of 35mm / s, preset welding pattern of a circle with a radius of 0.4mm, repetition rate of 170Hz, anchored by four anchor points {OSC1, OSC2, OSC3, OSC4}. The black envelope is formed by the initial control vector W, set to Φkb. The third axis is not shown in the figure, but in fact, the spatial motion is three-dimensional and includes the welding process. The weld point is not necessarily the surface focal point. Furthermore, if Lc(t) is set to 1, different color depths can represent the power. However, if Lc(t) varies with time, it cannot be described graphically.

[0080] After matrix operation of Φkb×F×C, the dynamic (OSC1′, OSC2′, OSC3′, OSC4′} is obtained, whose envelope is [W(t)], which is the actual control vector set obtained by the preset welding mode according to the real-time measured weld characteristic response.

[0081] Example 2

[0082] like Figures 2 to 6 As shown, a laser welding method is based on a laser welding system with a visual inspection device, and the laser welding method includes the following steps:

[0083] Step S1: obtaining weld information before the laser welding head through a visual inspection device;

[0084] like Figure 8 As shown, the laser welding system with a visual detection device includes a laser welding head (i.e., a laser), an optical positioning device, an optical positioning device controller, a machine vision processor, a welding auxiliary device for controlling the movement of the laser welding head and its controller, etc. The optical positioning device described in this embodiment adopts an optical scanning galvanometer, and in other embodiments, other devices based on refraction and reflection of light to achieve high-speed positioning can also be used.

[0085] In this embodiment, a camera captures the weld image in front of the machine welding head, and a machine vision processor recognizes the image to obtain weld information, which includes but is not limited to weld characteristics, the material of the object to be welded, and other information. The weld characteristics are:

[0086] [C(t)]:{[x n (t), y n (t), z n (t)]|1≤n≤s type}

[0087] The identification of seam features can be performed using common weld CCD image recognition methods, or by extracting weld image features using image block selection. As long as the seam features can be identified, the weld characteristics will be sufficient. Information about the material of the welded objects can be obtained through input or image recognition feature comparison. Different materials require different welding processes. For example, aluminum and iron have different melting points, requiring different laser energies. Similarly, the laser welding processes for aluminum alloys and stainless steel are completely different.

[0088] In practical applications, the above acquisition process is performed continuously in real time. As the laser welding head moves, the weld seam information in front of the laser welding head is acquired in real time. The seam feature C(t) is a data matrix obtained by recognizing the image captured in real time by the visual camera.

[0089] This embodiment combines practical applications to redefine S type ,like Figure 2 As shown in the figure, it is a common type of joint. Taking the second one in the second row as an example, it is a typical V-groove weld, so its S is defined. type =4, then the single set of measurement data C(t) is a 3x4 matrix. The subscripts 1, 2, 3, and 4 represent the four feature points from left to right. The machine vision camera collects data in the following format in real time:

[0090]

[0091] At the same time, a welding motion trajectory [F(t)] is generated based on the weld information to control the movement of the welding auxiliary device: {[u(t), v(t), w(t), θ(t)]}, where U, V, and W are three-dimensional coordinates, distinguished only from x, y, and z, and θ is used to indicate the advance angle. Alternatively, F(t) can be a pre-set welding motion trajectory obtained based on the weld information.

[0092] Step S2: obtaining welding process parameters through a preset welding process library according to the obtained weld information;

[0093] A welding process library is preset, and the welding process library is generated by preset parameters including at least joint type, material coefficient, swing reference, and weld morphology. Through the welding process library, welding process parameters of different materials under different joint types and morphologies can be obtained.

[0094] Step S3: generating a welding control vector according to the weld information and welding process parameters;

[0095] The control vectors are generated based on weld information and are used to distribute laser energy and laser energy density in real time within the weld head coordinate system. This allows for precise, real-time laser energy delivery, minimizing unnecessary heat input and lowering the probability of weld failures caused by localized thermal deformation during welding.

[0096] In practical applications, the control vector can be expressed as:

[0097] [W(t)]:{[P xyz (t)],[LP(t),L c (t)]}

[0098] Among them, P xyz (t) is the laser energy delivery coordinate (x, y, z) based on the welding head coordinate system at time t, which is the real-time coordinate obtained based on the welding motion trajectory F(t) and the real-time weld feature modification; LP(t) is the laser output power at time t, L c (t) is the modeling of the laser action mode.

[0099] P(t) represents the laser energy delivery coordinates (x, y, z) at time t, based on the weld head coordinate system. This is derived from the weld trajectory F(t) and real-time weld characteristics. x and y primarily address weld position offset and bead construction, while z primarily addresses height deviation (surface unevenness, rotational eccentricity, etc.) and power density control. LP(t) represents the laser output power at time t. Lc(t) models the laser's action mode. For example, a CW laser with stable power output has a value of 1. For CW lasers requiring fluctuating power due to the welding process, a (0, 1) rise / fall model is constructed based on the laser's technical characteristics.

[0100] Control vector W(t) = Φ KB ({[F(τ)]|τ<t},{[B][A] -1 [C(τ)]|τ<tt l The generation mechanism of}) is as follows:

[0101] Where [A] is the camera installation position, [B] is the welding head installation position, t l is the welding guidance time, Φ KB is the welding process parameter.

[0102] The machine vision camera and welding head are fixedly installed on the auxiliary device. The weld detection point is ahead of the welding point in the flight direction. Here, a parameter t is introduced. l (leading-time) is quantified. At time t, {[F(τ)]|τ<t} is the entire flight trajectory data before time t, {[B][A] -1 [C(τ)]|τ<tt l The generation mechanism of the control vector is based on the continuous calculation of the two sets of input values ​​by KB (preset process).

[0103] like Figures 3 to 5 As shown, with an S type =4 is an example of a 304 stainless steel long tailor weld with 4 characteristic points:

[0104] like Figure 3 As shown, (1)(2) and (3)(4) form a height difference of 0.25+ / -0.2mm, and (2)(3) form a gap error of 0.6+ / -0.25mm;

[0105] like Figure 4 As shown, during the welding process, the machine moves along the slide rail with the welding head, and the weld seam keeps changing on a small scale;

[0106] like Figure 5As shown in the figure, the preset process obtained from the preset welding process library is: CW laser output, preset power of 1800W, base welding speed of 35mm / s, preset welding pattern of a circle with a radius of 0.4mm, repetition rate of 170Hz, anchored by four anchor points {OSC1, OSC2, OSC3, OSC4}. The black envelope is formed by the initial control vector W, set to Φkb. The third axis is not shown in the figure, but in fact, the spatial motion is three-dimensional and includes the welding process. The weld point is not necessarily the surface focal point. Furthermore, if Lc(t) is set to 1, different color depths can represent the power. However, if Lc(t) varies with time, it cannot be described graphically.

[0107] After matrix operation of Φkb×F×C, the dynamic (OSC1′, OSC2′, OSC3′, OSC4′} is obtained, whose envelope is [W(t)], which is the actual control vector set obtained by the preset welding mode according to the real-time measured weld characteristic response.

[0108] Step S4: controlling the laser welding head to perform laser welding according to the control vector.

[0109] In practical applications, the step S4 of controlling the laser welding head to perform laser welding according to the control vector specifically involves controlling the laser welding head to perform laser welding according to the real-time corrected control vector.

[0110] As the laser welding head moves along the welding trajectory F(t), the control vector controls the energy distribution and energy delivery points of the laser welding head. The laser welding head makes fine adjustments left and right along the larger trajectory, precisely delivering laser energy to the desired point. The welding trajectory F(t) can be preset. For example, to weld a straight line, set the robot to run in a straight line. To weld many short welds on a surface that are not aligned, set an optimized trajectory to run continuously. The galvanometer quickly locates the focal point and welds left and right, eliminating the need to stop each time the robot reaches a short weld. For a weld that appears straight at a macroscopic level but exhibits microscopic deviations or gaps that vary in width, a linear guide can be used to guide the galvanometer welding head in a straight line. The welding trajectory F(t) can also be automatically generated in real time based on the seam characteristics C(t).

[0111] In practical applications, this can be done through a welding auxiliary device, which is used to run the welding motion trajectory F(t), while the control vector accurately controls the laser welding head to accurately distribute and deliver energy.

[0112] Advantages of the present invention:

[0113] The laser welding method described in this invention quickly responds to detected weld characteristics, achieves real-time laser energy and laser energy density distribution through vector control, and simultaneously adjusts the welding process in real time to directly perform welding. It eliminates the need for pre-running to record all weld information, calculate the motion trajectory, and then perform the repetitive motion. This reduces welding guidance time by tens or even hundreds of times. This brings three advantages:

[0114] 1. Reduce unnecessary welding heat input;

[0115] 2. Reduce the welding defect rate caused by irregular gaps;

[0116] 3. Reduce the welding defect rate caused by local thermal deformation during welding.

[0117] Compared to the first solution in the background art, the laser welding method described in this invention eliminates the need for the robot to compensate for guidance signals. Instead, the robot follows a preset trajectory and transmits real-time position data F(t) to the welding head for real-time calculation and compensation. This significantly reduces the demands placed on the robot and is therefore suitable for existing robots, which lack the ability to perform real-time small-step trajectory changes while carrying a load after reaching a certain speed.

[0118] Compared with the second solution of the background technology, the laser welding method described in the invention can work well for relatively regular welds as long as the error between the actual weld and the positioner trajectory is not so large as to exceed the range of the visual inspection device.

[0119] Compared with the second solution of the background technology, when using the laser welding method described in the invention, when the welding speed is required to be very fast, the first two solutions are not easy to implement, but it is feasible to fix the pipe with the welding head and rotate it by itself. In terms of technical performance, the implicit requirement is that the welding head must be able to respond to changes in weld details in real time.

[0120] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for generating a laser welding control vector, based on a laser welding system with a visual detection device, characterized in that: The laser welding control vector is a real-time laser energy and laser energy density distribution method based on the welding head coordinate system generated according to the weld information. The laser welding control vector generation method includes the following steps: Obtain the weld information before the laser welding head through the visual inspection device; Obtain welding process parameters through a preset welding process library based on the obtained weld information; Generate welding control vector according to weld information and welding process parameters; The welding control vector is: in, is the laser energy delivery coordinate based on the welding head coordinate system at time t ; is the laser output power at time t, It is a modeling of the laser action mode, and the control vector is used to achieve real-time and accurate delivery of laser energy.

2. The method for generating a laser welding control vector according to claim 1, wherein: The weld information includes recognized weld features, which are data matrices obtained by recognizing images collected in real time by a visual camera.

3. The method for generating laser welding control vectors according to claim 1, wherein: The welding process library is generated by preset parameters including at least joint type, material coefficient, swing reference, and weld morphology.

4. A laser welding method based on a laser welding system with a visual inspection device, characterized in that: The laser welding method comprises the following steps: The weld information before the laser welding head is obtained through the visual inspection device; Obtain welding process parameters through a preset welding process library based on the obtained weld information; Generate a welding control vector based on the weld information and welding process parameters, wherein the welding control vector is a real-time laser energy and laser energy density distribution method based on the welding head coordinate system generated according to the weld information; The laser welding head is controlled to perform laser welding according to the control vector, and the welding control vector is expressed as: Among them, [A] is the camera installation position, [B] is the welding head installation position is the welding guide time, is the welding process parameter, at time t, is the entire flight trajectory data before time t, for The seam feature data before the moment, the generation mechanism of the control vector is based on the continuous calculation of the two sets of input values ​​by the welding process, the weld information includes the identified seam features, the seam features It is the data matrix obtained by recognizing the images collected by the visual camera in real time.

5. The laser welding method according to claim 4, characterized in that: The laser welding method comprises: obtaining a welding motion trajectory according to weld seam information; , the welding motion trajectory It can be preset or generated in real time based on weld information.

6. The laser welding method according to claim 5, characterized in that: At any time t, the real-time collected seam features and the synchronously acquired welding motion trajectory , combined with the obtained welding process parameters, real-time adjustment to obtain the welding control vector .

Images