Weld joint three-dimensional structure detection device and method based on vortex laser

Abstract

The invention discloses a weld joint three-dimensional structure detection device based on vortex laser, and relates to the field of laser welding, and the weld joint three-dimensional structure detection device comprises a vortex laser generation system used for generating and emitting vortex laser; the image acquisition unit is used for collecting returned laser data projected to the welding seam and carrying out dynamic scanning and data acquisition; the motion control and feedback system is used for adjusting the relative position of the laser welding head and the workpiece in real time and dynamically adjusting parameters such as welding power and scanning speed according to the collected data result; and the data processing and analyzing platform is used for carrying out image processing and feature extraction on the collected multi-dimensional data and modeling the morphology of the welding sample. The vortex laser serves as a light source for welding seam detection, high-precision, high-depth, multi-parameter and strong-interference-resistance three-dimensional welding seam modeling is achieved, the problems that in the long and thick plate welding process, welding seam detection parameters are few, precision is low, and real-time adjustment cannot be achieved are solved, and meanwhile the device has the performance characteristics of being high in detection precision, high in efficiency and comprehensive in structural parameter.

Description

Technical Field

[0001] The present invention relates to the field of laser welding, and in particular to a device and method for detecting the three-dimensional structure of a weld based on a vortex laser. Background Art

[0002] The welding of long and thick plates (such as ship decks and aerospace structures) has extremely high requirements for the depth-to-width ratio, penetration and three-dimensional morphology of the weld. Due to the large thickness of the plate (usually ≥10mm), defects such as unfused grooves, pores, and thermal cracks are prone to occur during the welding process, and the weld width and molten pool morphology show significant non-uniformity with heat accumulation and process fluctuations. For example, in the welding of thick aluminum alloy plates, high thermal conductivity and light reflection characteristics will lead to poor stability of small holes, further exacerbating the generation of defects. Therefore, it is very important to adjust the welding parameters in real time by detecting the weld structure in the welding of long and thick plates. Traditional welding quality inspection relies on offline sampling or two-dimensional line laser scanning, which cannot provide real-time feedback of three-dimensional morphology data, resulting in delayed adjustment of process parameters and difficulty in meeting high-precision welding requirements. In laser welding, especially thick plate laser welding, the width and defects of the weld are usually more different.

[0003] Existing weld inspection technologies include two-dimensional line laser inspection, structured light three-dimensional inspection, and optical coherence tomography (OCT). Two-dimensional line laser inspection projects a single laser line onto the weld surface and uses triangulation to obtain weld profile height information. Its advantages include simple hardware, strong real-time performance (>1000Hz), and suitability for dynamic tracking. However, its disadvantages are also significant. Two-dimensional line laser inspection can only obtain a single-line two-dimensional profile and cannot reconstruct the three-dimensional structure of the groove or keyhole. It is also sensitive to spatter and smoke, and has limited accuracy (±0.05mm).

[0004] The principle of structured light 3D detection is to project coded stripes (such as sinusoidal gratings or Gray codes) onto the weld and reconstruct a 3D point cloud through phase solution or grating deformation. The advantage of this method is that the 3D topography of the entire surface can be obtained with high detection accuracy (±5μm). The disadvantage of structured light 3D detection is, on the one hand, poor stability. This is because the structured light modulated outside the cavity is easily disturbed by welding spatter and high-temperature airflow, resulting in pattern distortion. On the other hand, the algorithm is complex. This is because structured light 3D detection requires multi-frame phase unwrapping and noise suppression, and has low real-time performance (10-100Hz), which makes it difficult to meet the needs of high-speed welding. In addition, the cost of the structured light 3D detection system is high.

[0005] Optical coherence tomography (OCT) utilizes polarized light interference to determine the internal structure of welds, enabling the detection of keyhole depth and weld pool dynamics. While OCT offers the advantage of high resolution (micrometer-level), suitable for deep penetration weld inspection, it also has significant drawbacks. Its light source is sensitive to multiple reflections from the metal surface, resulting in complex data processing and difficulty in simultaneously detecting both surface topography and internal defects.

[0006] Therefore, it is very important to develop a laser welding weld detection method that can achieve high-precision three-dimensional modeling in real time. Summary of the Invention

[0007] In order to overcome the above-mentioned defects of the prior art, the present invention provides a vortex laser-based weld three-dimensional structure detection device and method thereof to solve the problems existing in the above-mentioned background technology.

[0008] The present invention provides the following technical solution: a vortex laser-based weld three-dimensional structure detection device, comprising:

[0009] A vortex laser generation system includes a vortex laser, a conical mirror, and a beam expander. The vortex laser is used to generate and emit vortex laser light, which is converted into a Bessel beam through the conical mirror and the beam expander.

[0010] Image acquisition unit, including high-speed industrial camera and multi-lens imaging system, is used to collect laser data projected onto the weld and return, and perform dynamic scanning and data acquisition;

[0011] The motion control and feedback system includes a motion control unit and a feedback unit. The free end of the motion control unit is equipped with a laser welding head, which is used to adjust the relative position of the laser welding head and the workpiece in real time to ensure that the scanning path is aligned with the weld. The feedback unit dynamically adjusts parameters such as welding power and scanning speed based on the collected data results.

[0012] The data processing and analysis platform performs image processing and feature extraction on the collected multi-dimensional data and models the morphology of the welding samples.

[0013] Preferably, the vortex laser is one of a chrysoberyl laser, a Nd:YAG laser, a Yb:YAG laser, and a Tm:YLF laser.

[0014] Preferably, the image acquisition unit also includes an image acquisition unit control axis, and the high-speed industrial camera and the multi-eye imaging system are assembled on the image acquisition unit control axis, which is used to drive the high-speed industrial camera and the multi-eye imaging system to move and automatically locate the weld position.

[0015] Preferably, the image acquisition unit further includes a filter for selectively transmitting the target laser signal and suppressing light interference in the environment.

[0016] Preferably, the motion control and feedback system includes a six-axis robotic arm / high-precision displacement platform and a closed-loop control system. The function of the six-axis robotic arm / high-precision displacement platform is to adjust the relative position of the laser welding head and the workpiece in real time to ensure that the scanning path is aligned with the weld. The closed-loop control system is an integrated PLC / industrial PC, which can dynamically adjust parameters such as welding power and scanning speed according to the detection results.

[0017] Preferably, the data processing and analysis platform includes a phase solution algorithm and a deep learning model. The phase solution algorithm uses a four-step phase shift method or sparse Bayesian learning (SBL) to process the phase change of vortex light and extract the sub-pixel morphology of the weld edge. The deep learning model is an improved U-Net model, which directly segments the defect area from the phase image.

[0018] A method for detecting three-dimensional structure of welds based on vortex laser includes the following steps:

[0019] S1. Beam projection: Based on the generated vortex beam, the Bessel beam conversion method is used to expand the vortex beam into a non-diffraction beam to cover the deep area of ​​the weld;

[0020] S2. Dynamic scanning and data acquisition: The vortex laser projected onto the weld surface is reflected to a high-speed industrial camera and a multi-lens imaging system, which dynamically scan and acquire the collected data.

[0021] S3, Image processing and feature extraction: including preprocessing and 3D reconstruction;

[0022] S4. Real-time feedback and process optimization: Input parameters such as weld width and penetration into the process model to dynamically adjust welding parameters.

[0023] Preferably, in step S2, a multimodal data fusion method is adopted to simultaneously collect light intensity distribution, phase change and reflected light polarization information to enhance the robustness of three-dimensional reconstruction.

[0024] Preferably, in step S3, the preprocessing is noise reduction, non-local mean filtering is performed to eliminate splash noise, and phase unwrapping is performed to solve the absolute phase through a four-step phase shift method.

[0025] Preferably, in step S3, the three-dimensional reconstruction includes point cloud generation and defect quantification. During the point cloud generation process, multi-view data is fused, the three-dimensional model is spliced ​​using the iterative closest point (ICP) algorithm, and pits and pore defects are identified based on curvature analysis to perform defect quantification.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] 1. The present invention uses vortex laser as the detection light source. Compared with traditional structured light detection that relies on extracavity modulation devices, the intracavity vortex light generation does not require external modulation devices, thus avoiding energy loss and having a more concentrated light intensity distribution. It is suitable for deep penetration welding keyhole detection. The intracavity generation method suppresses high-order stray modes through mode competition, and the line width can be compressed to below 0.1nm. When working in conjunction with a narrowband filter, the signal-to-noise ratio is improved by more than 5 times, effectively resisting the interference of molten pool plasma radiation. At the same time, the integrated cavity design eliminates the micron-level assembly error of traditional split optical systems and can still maintain phase stability under vibration conditions.

[0028] 2. The present invention uses a vortex laser as the detection light source. The annular intensity distribution of the vortex light enables it to naturally avoid the molten metal droplets at the edge of the keyhole during propagation, reducing the probability of light path blockage. At the same time, the welding pool plasma will spontaneously form a rotating structure under the action of the magnetic field. The direction of its angular momentum has a coupling effect with the OAM of the vortex laser. By analyzing the topological charge changes of the reflected vortex light, the plasma density gradient and electron temperature distribution can be synchronously inverted, providing a new dimension for real-time evaluation of penetration.

[0029] 3. The present invention uses vortex laser as the detection light source to synergistically enhance its phase sensitivity and non-diffraction characteristics. The phase change is solved by the four-step phase shift method to achieve sub-pixel level morphology reconstruction with an accuracy of ±0.01mm, which greatly improves the detection accuracy. At the same time, by introducing a conical mirror in the optical path, the vortex light is converted into a non-diffraction Bessel beam. Its self-healing characteristics can penetrate the vapor layer of the molten pool, and the focal depth is extended to the millimeter level. Compared with the traditional Gaussian beam with a focal depth of 100 microns, the accuracy is significantly improved, thereby completely capturing the keyhole depth structure and greatly improving the three-dimensional morphology reconstruction capability.

[0030] 4. The vortex laser of the present invention integrates single-beam multi-parameter extraction with AI algorithm. Compared with the traditional multi-sensor detection system (such as line laser + thermal imager + acoustic emission sensor), it adopts an improved U-Net model to directly map the vortex light phase image to the three-dimensional morphology parameters, bypassing the time-consuming phase unwrapping and noise suppression steps in traditional structured light detection and simplifying the algorithm.

[0031] 5. The weld detection using vortex laser as the detection laser in the present invention can be coordinated with the real-time dynamic parameter control. By inputting the detection data into the process model, the laser power, defocus amount and other parameters can be adjusted in real time to avoid the formation of hump welds.

[0032] 6. The weld detection of the present invention using vortex laser as the detection laser can integrate vortex light welding and detection, and use ultra-strong vortex laser to achieve welding and detection at the same time. The central area of ​​the annular spot is used for high-power welding, and the outer annular area is used for morphology detection, realizing coaxial synchronization of "processing-detection".

[0033] 7. The present invention adopts vortex laser with high stability and high penetration as the detection light source, and realizes the detection of weld structure in laser welding in a low-cost, high-efficiency and high-precision manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a schematic diagram of the weld detection structure using the vortex laser as the detection light source of the present invention.

[0035] Figure 2 This is a schematic structural diagram of a circular groove shape on the surface of the welding point of the welding workpiece of the present invention.

[0036] Figure 3 This is a schematic structural diagram of a welding workpiece according to the present invention in which the surface of the welding point is in a square groove shape.

[0037] Figure 4 This is a schematic structural diagram of a welding workpiece according to the present invention in which the surface of the welding point is in a non-groove shape.

[0038] Figure 5 This is a flow chart of weld detection using the vortex laser as the detection light source of the present invention.

[0039] The accompanying drawings are marked as follows: 1. Vortex laser; 2. Conical mirror; 3. Beam expander; 4. Motion control unit; 5. Feedback unit; 6. Laser welding head; 7. High-speed industrial camera; 8. Multi-eye imaging system; 9. Image acquisition unit control axis; 10. Data processing and analysis platform; 11. Filter. DETAILED DESCRIPTION

[0040] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making a clearer and more precise definition of the protection scope of the present invention.

[0041] The present invention provides a vortex laser-based weld three-dimensional structure detection device, such as Figure 1 As shown, the vortex laser generation system, image acquisition unit, motion control and feedback unit and data processing and analysis platform.

[0042] The vortex laser generation system provides a high-purity, high-stability detection light source. Its spiral phase characteristics are sensitive to surface deformation and support submicron precision detection. Specifically, it includes a vortex laser, a conical mirror, and a beam expander. The vortex laser, which generates vortex laser light through an intracavity method, emits vortex laser light. The conical mirror and beam expander convert the vortex light into a Bessel beam, extending the focal depth to the millimeter level and improving the ability to detect weld depth.

[0043] The image acquisition and processing unit includes a high-speed industrial camera and a multi-lens imaging system. The high-speed industrial camera requires a narrowband filter to suppress interference from strong light in the molten pool, and a frame rate of 1000 fps or higher to freeze the dynamic welding process. The multi-lens imaging system uses multi-angle cameras to synchronously capture reflected light field data, combining it with the phase information of the vortex light to generate a 3D point cloud.

[0044] The motion control and feedback system includes a motion control unit and a feedback unit. The free end of the motion control unit is equipped with a laser welding head, which is used to adjust the relative position of the laser welding head and the workpiece in real time to ensure that the scanning path is aligned with the weld. The feedback unit dynamically adjusts parameters such as welding power and scanning speed based on the collected data results.

[0045] The data processing and analysis platform performs image processing and feature extraction on the collected multidimensional data, modeling the weld sample's topography. Specifically, the platform includes a phase resolution algorithm and a deep learning model. The phase resolution algorithm uses a four-step phase shift method or sparse Bayesian learning (SBL) to process vortex light phase variations and extract the sub-pixel topography of the weld edge. The deep learning model is a modified U-Net model that directly segments defect areas from phase images.

[0046] In this embodiment, the vortex laser is one of a chrysoberyl laser, a Nd:YAG laser, a Yb:YAG laser, and a Tm:YLF laser.

[0047] In this embodiment, the image acquisition unit also includes an image acquisition unit control axis, and the high-speed industrial camera and the multi-eye imaging system are assembled on the image acquisition unit control axis, which is used to drive the high-speed industrial camera and the multi-eye imaging system to move and automatically locate the weld position.

[0048] In this embodiment, the image acquisition unit further includes a filter for selectively transmitting the target laser signal and suppressing light interference in the environment.

[0049] In this embodiment, the motion control unit includes a six-axis robotic arm / high-precision displacement platform, and the feedback unit includes a closed-loop control system. The function of the six-axis robotic arm / high-precision displacement platform is to adjust the relative position of the laser welding head and the workpiece in real time to ensure that the scanning path is aligned with the weld. The closed-loop control system is an integrated PLC / industrial PC, which can dynamically adjust parameters such as welding power and scanning speed according to the detection results.

[0050] Using vortex laser as the detection light source, compared with traditional structured light detection that relies on extracavity modulation devices (such as spatial light modulators or diffraction optical elements) to phase encode the Gaussian beam, a quasi-continuous coding pattern is generated.

[0051] Traditional structured light detection with extracavity modulation suffers from two issues: First, high energy loss. The laser energy is dispersed during the pixel-by-pixel modulation process, making it difficult to penetrate the highly reflective metal surfaces used in thick plate welding. Second, poor interference immunity. Welding spatter or high-temperature airflow can easily cause thermal distortion of the modulator, resulting in spot distortion or phase noise. For example, in high-power welding at 10kW, the failure rate of extracavity-modulated structured light systems can reach over 30%.

[0052] Vortex lasers, on the other hand, generate vortex beams carrying orbital angular momentum (OAM) directly during the laser resonance phase by controlling the intracavity mode of the resonant cavity (e.g., based on Nd:YVO4 crystal combined with spherical aberration cavity design). This has the following effects:

[0053] 1. Improved energy efficiency: Intracavity vortex light generation does not require external modulation devices, avoiding energy loss (traditional structure light loss >40%), and the light intensity distribution is more concentrated, making it suitable for deep penetration welding keyhole detection.

[0054] 2. Breakthrough in spectral purity: The intracavity generation method suppresses high-order spurious modes through mode competition, and the linewidth can be compressed to below 0.1nm. When working in conjunction with narrowband filters (such as 808±2nm), the signal-to-noise ratio is improved by more than 5 times, effectively resisting interference from molten pool plasma radiation.

[0055] 3. Enhanced mechanical stability: The integrated cavity design eliminates the micron-level assembly errors of traditional split optical systems and can maintain phase stability (phase jitter <λ / 20) under vibrating conditions (such as shipyard welding workshops).

[0056] In terms of penetration capability and detection parameters, vortex laser as a detection light source has achieved the transition from two-dimensional detection to synchronous monitoring inside the molten pool. Traditional line lasers or structured light face two major penetration bottlenecks in thick plate welding:

[0057] One is the scattering attenuation of the detection light by the molten pool vapor and plasma. For example, in aluminum alloy welding, the metal vapor above the molten pool can attenuate the detection light intensity by 90%.

[0058] Secondly, the dynamic fluctuation of the keyhole causes the light path to be blocked. Especially in deep penetration welding, the collapse of the keyhole will instantly block the detection light path.

[0059] The spiral phase structure of the vortex laser (phase distribution is: Where l is the topological charge change number), which gives it a unique penetration mechanism: on the one hand, it can be summarized as the obstacle diffraction effect: the annular intensity distribution of the vortex light allows it to naturally avoid the molten metal droplets at the edge of the keyhole during propagation, reducing the probability of light path blocking. According to tests, in deep penetration welding with a keyhole diameter of 0.5 mm, the effective detection depth of vortex light is 3 times higher than that of Gaussian light. On the other hand, it is plasma vortex matching: the welding pool plasma will spontaneously form a rotating structure under the action of the magnetic field, and its angular momentum direction has a coupling effect with the OAM of the vortex laser. By analyzing the topological charge change number offset (Δl) of the reflected vortex light, the plasma density gradient and electron temperature distribution can be synchronously inverted, providing a new dimension for real-time evaluation of penetration. For example, when Δl suddenly changes from +1 to -1, an unfused defect at the bottom of the keyhole can be warned.

[0060] A vortex laser-based weld 3D structure detection method, using vortex laser as the detection light source weld detection process and implementation steps are shown Figure 5 The following steps are included:

[0061] S1. Beam projection: Based on the generated vortex beam, the Bessel beam conversion method is used to expand the vortex beam into a non-diffraction beam to cover the deep area of ​​the weld;

[0062] S2. Dynamic scanning and data acquisition: The vortex laser projected onto the weld surface is reflected by a high-speed industrial camera and a multi-camera imaging system, which dynamically scans and acquires the collected data. A multimodal data fusion method is used to simultaneously collect information on light intensity distribution, phase change, and reflected light polarization, enhancing the robustness of 3D reconstruction.

[0063] S3. Image processing and feature extraction: including preprocessing and 3D reconstruction; preprocessing is noise reduction, non-local mean filtering is performed to eliminate splash noise, and phase unwrapping is performed, and the absolute phase is solved by the four-step phase shift method; 3D reconstruction includes point cloud generation and defect quantification. During the point cloud generation process, multi-view data is integrated, and the iterative closest point (ICP) algorithm is used to splice the 3D model. Pits and pores are identified based on curvature analysis, and defect quantification is performed.

[0064] S4. Real-time feedback and process optimization: Input parameters such as weld width and penetration into the process model to dynamically adjust welding parameters.

[0065] As a detection light source, vortex laser has a stronger ability to reconstruct three-dimensional morphology, which is manifested in the synergistic enhancement of phase sensitivity and non-diffraction characteristics. The three-dimensional detection capability of vortex laser comes from the dual breakthrough of its physical properties: on the one hand, the spiral phase sensitivity mechanism of vortex laser. When vortex light is projected onto the weld surface, the surface height difference Δh will cause the phase of the reflected light to change. The relationship can be approximated as (λ is the wavelength). By solving the phase change through the four-step phase shift method, sub-pixel morphology reconstruction can be achieved with an accuracy of ±0.01mm, which is an order of magnitude higher than the traditional structured light. On the other hand, there is the Bessel beam fusion design. By introducing a conical mirror in the optical path, the vortex light is converted into a non-diffraction Bessel beam. Its self-healing properties can penetrate the vapor layer of the molten pool, and the focal depth is extended to the millimeter level, which is significantly improved compared to the hundred-micron focal depth accuracy of the traditional Gaussian beam, thereby fully capturing the keyhole depth structure. For example, in the welding of 20mm thick steel plates, the Bessel-vortex composite beam can simultaneously obtain three-dimensional point cloud data of the bottom of the groove (depth 18mm) and the surface residual height.

[0066] Vortex laser as a detection light source can achieve dynamic adaptability and algorithm simplification. This is because, as a detection light source, vortex laser can realize the transition from multi-sensor fusion to single-beam intelligent analysis. Compared with traditional multi-sensor detection systems (such as line laser + thermal imager + acoustic emission sensor), there are problems such as complex calibration and difficult data fusion. Vortex laser has opened up a new path for detection technology by fusing single-beam multi-parameter extraction with AI algorithm. The annular intensity distribution of vortex light will radially shrink or expand with the geometric characteristics of the weld. By analyzing the change in the radius of the light ring (ΔR) and the phase gradient Weld width, penetration, and surface roughness can be simultaneously calculated. For example, in 304 stainless steel welding, the linear correlation coefficient between ΔR and penetration reached 0.98. Using an improved U-Net model, vortex light phase images are directly mapped to three-dimensional topography parameters, bypassing the time-consuming phase unwrapping and noise suppression steps in traditional structured light detection. Experiments have shown that this solution reduces single-frame processing time from 15ms to 2ms for 5mm-thick aluminum alloy weld inspection, meeting 100Hz real-time detection requirements.

[0067] Furthermore, weld inspection using a vortex laser as the probe laser can be coordinated with real-time dynamic parameter control. By inputting inspection data into a process model (such as a PINN-based melt pool dynamics model), parameters such as laser power and defocus can be adjusted in real time. For example, if the keyhole aspect ratio is detected to exceed a critical value, the system automatically reduces the power by 10% to prevent the formation of a hump weld bead.

[0068] Furthermore, weld inspection using vortex lasers as probe lasers can be combined with vortex laser welding and inspection to achieve integrated performance. This approach utilizes ultra-intense vortex lasers (peak power >10GW) for simultaneous welding and inspection. The central area of ​​the annular spot is used for high-power welding, while the outer annular area is used for topography inspection, achieving coaxial synchronization of "processing and inspection." This solution has reduced the defect miss detection rate in nuclear power plant pressure vessel welding from 1.2% to 0.05%.

[0069] Furthermore, weld inspection using vortex laser as the detection laser can be compatible with high-power lasers, while high-power lasers and high-energy lasers are necessary conditions for achieving thick plate welding.

[0070] The following tests were performed using a vortex laser-based weld three-dimensional structure detection device and method:

[0071] Test 1: Thick plate laser welding inspection (aerospace field), welding 20mm thick aluminum alloy cabin structure, length more than 1m, need to detect groove width and keyhole stability. The welding point surface of the welding workpiece is as follows Figure 2-3 The shape is shown. The weld seam width of the welded workpiece is 0-2mm, and the corresponding welding speed is 5cm / s to 2cm / s, and the welding power is 15,000 watts to 20,000 watts. A 2μm vortex laser is used as the detection laser light source, and a multi-lens 3D profilometer is used as the detection equipment, with a scanning speed of 2m / min. The corresponding beam expander expands the diameter of the vortex laser to 3mm, and the corresponding filter has high transmittance for the 2μm laser and high reflectivity for the welding beam. Mechanical adjustment uses a three-dimensional module and includes a rotational dimension. The vortex laser-based weld detection system detects the size of the weld seam and adjusts the movement speed and power during the welding process to achieve full welding at every position of the long and thick plate. It can also detect the plasma state of the keyhole in real time, realize real-time monitoring of the welding situation, and promptly terminate the welding process for workpieces that do not meet the welding requirements. The three-dimensional reconstruction accuracy is ±0.02mm, the detection rate of false welds is increased to 99.5%, and the process adjustment response time is <50ms.

[0072] Test 2: Online detection of spot welding on vehicles (new energy vehicle battery modules). Detection of the diameter and strength of the weld core of three-layer spot welding of galvanized steel plates. The welding method is spot welding. A six-axis robot is used as the control platform, and a 755nm vortex laser is used as the detection beam. The 755nm vortex laser is generated in the cavity of a chrysoberyl laser. The surface of the welding point of the welding workpiece is as follows: Figure 4The groove shape shown in the figure is captured using a high-speed camera. The vortex laser inspection system detects 47 welds per 5 seconds. By measuring the gap width and shape at each point, the optimal weld is found. Process parameters are adjusted based on the test results to achieve a full weld at each location. The effect of temperature on welding parameters is also considered based on the distance between welds. The inspection diameter error is ±0.1mm, and the heat-affected zone quantification accuracy is ±5%.

[0073] In summary, using a vortex laser as a probe beam and a weld detection light source allows for 3D modeling of welds, a better understanding of weld structure, and the use of appropriate welding parameters and processes for welding long and thick plates. Furthermore, the plasma molten pool during welding can be monitored in real time, allowing for timely adjustment of welding parameters. By using a vortex laser with high stability and penetration as a detection light source, weld structure detection during laser welding is achieved in a low-cost, high-efficiency, and high-precision manner. This also provides a method and technical approach for 3D weld structure detection based on vortex lasers.

[0074] Several points should be explained: First, in the description of this application, it should be noted that, unless otherwise specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, which can be mechanical connection or electrical connection, or internal communication between two elements, or direct connection. "Up", "down", "left", "right", etc. are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may change.

[0075] The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiment. Any equivalent modifications or changes made by ordinary technicians in this field based on the contents disclosed in the present invention should be included in the protection scope recorded in the claims.

Claims

1. A vortex laser-based weld three-dimensional structure detection device, characterized in that: include: A vortex laser generation system includes a vortex laser, a conical mirror, and a beam expander. The vortex laser is used to generate and emit vortex laser light, which is converted into a Bessel beam through the conical mirror and the beam expander. Image acquisition unit, including high-speed industrial camera and multi-lens imaging system, is used to collect laser data projected onto the weld and return, and perform dynamic scanning and data acquisition; The motion control and feedback system includes a motion control unit and a feedback unit. The free end of the motion control unit is equipped with a laser welding head, which is used to adjust the relative position of the laser welding head and the workpiece in real time to ensure that the scanning path is aligned with the weld. The feedback unit dynamically adjusts parameters such as welding power and scanning speed based on the collected data results. The data processing and analysis platform performs image processing and feature extraction on the collected multi-dimensional data and models the morphology of the welding samples.

2. The device for detecting three-dimensional weld structures based on vortex laser according to claim 1, characterized in that: The vortex laser is one of a chrysoberyl laser, a Nd:YAG laser, a Yb:YAG laser, and a Tm:YLF laser.

3. The device for detecting three-dimensional weld structures based on vortex laser according to claim 1, characterized in that: The image acquisition unit also includes an image acquisition unit control axis, and a high-speed industrial camera and a multi-eye imaging system are assembled on the image acquisition unit control axis, which is used to drive the high-speed industrial camera and the multi-eye imaging system to move and automatically locate the weld position.

4. The device for detecting three-dimensional weld structures based on vortex laser according to claim 1, characterized in that: The image acquisition unit further includes a filter for selectively transmitting the target laser signal and suppressing light interference in the environment.

5. The device for detecting three-dimensional weld structures based on vortex laser according to claim 1, characterized in that: The motion control and feedback system includes a six-axis robotic arm / high-precision displacement platform and a closed-loop control system. The function of the six-axis robotic arm / high-precision displacement platform is to adjust the relative position of the laser welding head and the workpiece in real time to ensure that the scanning path is aligned with the weld. The closed-loop control system is an integrated PLC / industrial PC, which can dynamically adjust parameters such as welding power and scanning speed based on the detection results.

6. The device and method for detecting three-dimensional structure of welds based on vortex laser according to claim 1, characterized in that: The data processing and analysis platform includes a phase solution algorithm and a deep learning model. The phase solution algorithm uses a four-step phase shift method or sparse Bayesian learning (SBL) to process the phase change of vortex light and extract the sub-pixel level morphology of the weld edge. The deep learning model is an improved U-Net model, which directly segments the defect area from the phase image.

7. A method for detecting three-dimensional structure of welds based on vortex laser, based on a device for detecting three-dimensional structure of welds based on vortex laser according to any one of claims 1 to 6, characterized in that: The following steps are involved: S1. Beam projection: Based on the generated vortex beam, the Bessel beam conversion method is used to expand the vortex beam into a non-diffraction beam to cover the deep area of ​​the weld; S2. Dynamic scanning and data acquisition: The vortex laser projected onto the weld surface is reflected to a high-speed industrial camera and a multi-lens imaging system, which dynamically scan and acquire the collected data. S3, Image processing and feature extraction: including preprocessing and 3D reconstruction; S4. Real-time feedback and process optimization: Input parameters such as weld width and penetration into the process model to dynamically adjust welding parameters.

8. The method for detecting three-dimensional structure of welds based on vortex laser according to claim 7, characterized in that: In step S2, a multimodal data fusion method is used to simultaneously collect light intensity distribution, phase change and reflected light polarization information to enhance the robustness of three-dimensional reconstruction.

9. The method for detecting three-dimensional structure of welds based on vortex laser according to claim 7, characterized in that: In step S3, the preprocessing is noise reduction, non-local mean filtering is performed to eliminate splash noise, and phase unwrapping is performed to calculate the absolute phase through a four-step phase shift method.

10. The method for detecting three-dimensional structure of welds based on vortex laser according to claim 7, characterized in that: In step S3, the three-dimensional reconstruction includes point cloud generation and defect quantification. During the point cloud generation process, multi-view data is fused, the three-dimensional model is spliced ​​using the iterative closest point (ICP) algorithm, and pits and pores are identified based on curvature analysis to perform defect quantification.

Images