Method for identifying annular narrow gap laser welding with filler wire using high-speed photography monitoring

Abstract

This invention discloses a method for identifying annular narrow-gap laser filler wire welding using high-speed photography monitoring, belonging to the field of laser welding technology. The method involves real-time acquisition of images of the molten pool movement, wire melting state, and droplet transition during welding. During acquisition, the acquisition device, laser beam, and wire end remain relatively stationary while the workpiece rotates. The acquired images are preprocessed; image processing algorithms are used to identify the edge contour of the molten droplets; the identified images are processed to obtain a simplified cross-sectional view of the wire feed end of the weld bead; the distances from the wire to both sides of the weld bead, the distance from the wire to the workpiece's welding position, the angle between the molten pool and the sidewall, and the relative position of the laser source and the wire are measured in the simplified image; welding parameters are adjusted based on the measured data. After high-speed photography acquisition and processing, the droplet transition form and transformation are qualitatively determined within the set range of each welding process parameter, and the influence of the molten pool movement state on the weld is assessed. This allows for dynamic adjustment of welding parameters to ensure weld joint quality and welding efficiency.

Description

Technical Field

[0001] This invention relates to a method for narrow-gap laser filler wire welding, specifically a method for identifying annular narrow-gap laser filler wire welding using high-speed photography monitoring, belonging to the field of laser welding technology. Background Technology

[0002] Low-carbon steel has a large application share in industrial production, widely used in the main structural components of mining machinery, pressure vessels, power plants, and bridges. Currently, the engineering machinery industry faces significant challenges in welding medium-thick plates (medium plates with a thickness greater than or equal to 12mm, and thick plates with a thickness greater than 20mm). Traditional arc welding suffers from low efficiency, high energy consumption, high cost, and generally poor quality. Narrow-gap laser wire-filled welding replaces CO2 gas shielded welding, changing the traditional 45° bevel to a narrow-gap bevel (3° angle on one side). The welding wire is automatically fed into the bevel, and the laser beam melts it. After cooling, it forms a weld together with the base metal. A gas guide tube is installed in front of or behind the laser gun, and side-blown inert gas suppresses the metal plasma while simultaneously providing gas protection for the weld. Narrow-gap laser wire-filled welding can reduce the amount of welding wire filler by approximately 80%, increase processing efficiency by 8-10 times, significantly save costs and energy consumption, reduce heat input, improve welding stress and deformation, and enhance product quality. Simultaneously… Under the action of a laser beam, the base metal melts to form a keyhole. Therefore, narrow-gap laser filler wire welding technology combines the advantages of narrow-gap welding and laser welding, making it one of the most promising research directions in thick plate welding.

[0003] Currently, there is limited research on the weld formation law of different welding process parameters (wire distance to workpiece surface, laser power, welding speed, wire feed speed, galvanometer oscillation frequency, galvanometer oscillation amplitude) in narrow-gap laser wire filler welding, as well as on real-time monitoring and adjustment and optimization of welding process parameters for workpieces of different thicknesses. For the processing of complex structural parts, excessive experiments in welding process research lead to high costs.

[0004] On May 29, 2020, Chinese invention patent application 2020100455290 disclosed a monitoring system and method for droplet transition in dual-laser beam dual-sided synchronous welding of filler wire based on high-speed imaging. This system uses a high-speed camera to monitor the droplet transition mode and frequency in real time during dual-laser beam dual-sided synchronous welding of filler wire, aiming to monitor the droplet transition throughout the process at different wire feed speeds. However, this method only monitors droplet transition at different wire feed speeds and cannot monitor and adjust multiple welding parameters, affecting the quality of the weld joint and welding efficiency. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the defects of the prior art and provide a method for high-speed photography monitoring and identification of annular narrow gap laser wire filler welding that can comprehensively collect welding status data during the narrow gap laser wire filler welding process and effectively ensure the quality of the welded joint and welding efficiency.

[0006] To solve the above-mentioned technical problems, the present invention provides a method for identifying annular narrow-gap laser filler wire welding using high-speed photography monitoring, comprising the following steps:

[0007] S1. Real-time acquisition of images showing the movement of the molten pool, the melting state of the welding wire, and the transition of molten droplets during welding; during acquisition, the acquisition device, laser beam, and welding wire end remain relatively stationary while the workpiece to be welded rotates.

[0008] S2. Preprocess the acquired images;

[0009] S3. Identify the edge contour morphology of the molten droplet using image processing algorithms;

[0010] S4. Extract the image after S3 recognition and process it to obtain a simplified cross-sectional view of the wire feed end of the weld bead;

[0011] S5. Measure the distance between the welding wire and the two sides of the weld bead, the distance between the welding wire and the workpiece to be welded position, the angle between the molten pool and the side wall, and the relative position of the laser source and the welding wire in the simplified diagram of S4;

[0012] S6. Adjust the welding parameters based on the data measured in S5.

[0013] In this invention, a laser light source is used as a backlight when acquiring images.

[0014] In this invention, the process of S2 is as follows:

[0015] S21. Using the weighted average method, the components of the R, G, and B channels of the image are weighted and averaged with different weights to obtain a more reasonable grayscale image: L=R*299 / 1000+G*587 / 1000+B*114 / 1000;

[0016] S22. Perform image geometric transformation on the grayscale image;

[0017] S23. Enhance the image processed in S22 using either the frequency domain method or the spatial domain method.

[0018] In this invention, the process of S3 is as follows: after inputting the preprocessed image into the CNN network, the image features are extracted through multiple convolutional pooling operations, and then the image features are sent into the fully connected layer network to complete the image classification and recognition.

[0019] In this invention, the process of S4 is as follows:

[0020] S41. Transmit the image to the recognition system for grayscale processing;

[0021] S42. Perform binary conversion on the grayscale image;

[0022] S43. Find and draw the outline on the processed image.

[0023] The beneficial effects of the present invention are as follows: (1) After high-speed photography and data acquisition and processing, the droplet transition form and transformation are qualitatively judged within the set range of each welding process parameter, the influence of the molten pool movement state on the weld is judged, and then the welding parameters are dynamically adjusted to ensure the quality of the welded joint and the welding efficiency; (2) Through the image recognition system, a simplified diagram of the welding section at the wire feeding end can be formed, and the welding state data in the weld bead can be measured, providing a basis for narrow gap laser welding of different engineering machinery structural parts; (3) By comparing the measurement data of the simplified diagram with the experimental database, the welding process parameters can be adjusted and optimized in a timely manner, avoiding multiple tests, thereby improving efficiency and saving costs; (4) Using a laser light source as a backlight can attenuate the arc light during welding, thereby ensuring that clear images are obtained in the acquisition equipment. Attached Figure Description

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. 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.

[0025] Figure 1 This is a schematic diagram of the welding device structure;

[0026] Figure 2 This is a diagram of the structure of a convolutional neural network model.

[0027] Figure 3 This is a schematic diagram of the structure of an image intelligent recognition system;

[0028] Figure 4 This is a schematic diagram of the comparison process;

[0029] Figure 5 This is a schematic diagram illustrating the effect of the distance between the welding wire and the workpiece surface.

[0030] Figure 6 This is a schematic diagram illustrating the effect of the galvanometer's oscillation frequency.

[0031] Figure 7 This is a schematic diagram illustrating the effect of the galvanometer's oscillation amplitude.

[0032] Figure 8Macroscopic morphology of weld and weld bead morphology under different oscillation amplitudes;

[0033] Figure 9 This is a schematic diagram illustrating the effect of laser welding power.

[0034] Figure 10 This is a schematic diagram illustrating the effect of laser welding speed.

[0035] Figure 11 This diagram illustrates the effect of wire feeding speed on the wire feeder. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0037] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0038] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0039] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this invention is in use. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention. In addition, the terms "first," "second," "third," etc., are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0040] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0041] The following detailed description of some embodiments of the present invention is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0042] like Figure 1 As shown, the welding apparatus used in the method of high-speed photography monitoring and identification of annular narrow gap laser filler wire welding of the present invention includes a laser welding system 1, a digital wire feeding control system 2, a high-speed photography acquisition system 3, and an image intelligent recognition system 4.

[0043] Among them, laser welding system 1 adopts an IPG YLS 4000-S2T laser welder with a maximum output power of 4kW, a fiber core diameter of 100um, a focal length of 300mm, and an IPG D30 Wobble galvanometer device. When the galvanometer is off and the defocus is zero, the spot diameter is 0.15mm. When the galvanometer is on, its deflection amplitude is a maximum of 3mm at zero defocus. The device has a maximum laser power of 6kW and a maximum oscillation frequency of 1000Hz. It is equipped with a FANUC M-20iA robot as a robotic arm to support the laser head. The device has a maximum operating radius of 1811mm, a maximum load capacity of 20kG, and a repeatability accuracy of 0.04mm.

[0044] The digital wire feeding control system 2 uses a WF25i REEL R wire feeder with a maximum wire feeding speed of 6m / min.

[0045] The high-speed photography acquisition system 3 uses a CP80-3-M-540 high-speed camera manufactured by Optronis, and is equipped with a Diode Laser System 40W laser light source 5 manufactured by BWT as the backlight. The resolution is 1696×1708 and the effective screen area is 13.57mm×13.68mm.

[0046] In this embodiment, a 40W laser light source 5 is used as a backlight to illuminate the welding position with the laser beam, which can attenuate the arc light during welding, thereby making the image clearer in the high-speed photography equipment.

[0047] The system architecture of the image intelligent recognition system device 4 is as follows: Figure 3As shown.

[0048] The specific process of the method for identifying annular narrow-gap laser filler wire welding using high-speed photography monitoring in this embodiment is as follows:

[0049] Step 1: Start the entire welding system and the high-speed camera will automatically perform high-frequency video recording;

[0050] Step 2: Turn on the backlight;

[0051] Step 3: Preprocess the acquired images to improve the signal-to-noise ratio of the color images.

[0052] (1) Grayscale Conversion: The R, G, and B channels of the image are processed sequentially. To improve the processing speed of the entire application system, the acquired color image needs to be converted to grayscale to reduce the amount of data to be processed. In this embodiment, a weighted average method is used, where the three components are weighted and averaged according to their importance with different weights. Given that the human eye is most sensitive to green and least sensitive to blue, a more reasonable grayscale image can be obtained by weighting and averaging the RGB components according to the following formula in this embodiment:

[0053] L=R*299 / 1000+G*587 / 1000+B*114 / 1000

[0054] In the formula, L is the image grayscale value, R is the red component, G is the green component, and B is the blue component.

[0055] (2) Geometric Transformation: Geometric transformations such as translation, transpose, mirroring, rotation, and scaling are used to process the grayscale image to correct systematic errors in the high-speed photography acquisition system and random errors in instrument position (imaging angle, perspective, and even lens-related factors). Furthermore, grayscale interpolation algorithms are required. (Grayscale interpolation is a method of redistributing pixels in a grayscale image to change the number of pixels; the "interpolation" program automatically selects pixels with better information to fill in blank pixels, achieving a smoother and cleaner image when magnified.) Because the calculations are performed according to this transformation relationship, the pixels of the output image may be mapped to non-integer coordinates of the input image. Commonly used methods include nearest neighbor interpolation, bilinear interpolation, and bicubic interpolation.

[0056] 3. Image Enhancement: The frequency domain method is used to enhance the image after geometric transformation to improve the visual effect of the image, emphasize the overall or local characteristics of the image, make the originally unclear image clear or emphasize certain features of interest, and amplify the differences between features of different objects in the image, suppress features of no interest, so as to improve the image quality, enrich the amount of information, and enhance the image interpretation and recognition effect to meet the needs of subsequent special analysis.

[0057] In practical applications, spatial domain methods can also be used for image enhancement processing.

[0058] Step 4: Identify the edge contour morphology of the molten droplet using image processing algorithms, such as... Figure 2 As shown:

[0059] (1) Input the preprocessed image into the CNN network and extract the image features through multiple convolution and pooling operations.

[0060] (2) The extracted image features are fed into a fully connected network for classification and recognition;

[0061] (3) The fully connected layer combines multiple pooled data features into a set of signal data for output, and performs image category recognition.

[0062] Step 5: Capture the image obtained in Step 4 and transmit it to the image intelligent recognition system to obtain a simplified cross-sectional view of the wire feeding end of the weld bead;

[0063] (1) Transmit the image to the recognition system for grayscale processing, using cv2.cvtColor from OpenCV (OpenCV is an open-source computer vision library);

[0064] (2) Perform binary conversion on the grayscale image, using cv2.threshold in OpenCV;

[0065] (3) The contour is found using the cv2.findContours method in OpenCV;

[0066] (4) Contour drawing: Use cv2.drawContours in OpenCV to draw contours on the image.

[0067] Step Six: Measure the distance between the welding wire and the two sides of the weld bead, the distance between the welding wire and the workpiece to be welded position, the angle between the molten pool and the sidewall, and the relative position of the laser source and the welding wire in the simplified diagram.

[0068] Import the drawn image into ImageJ (ImageJ is a Java-based public image processing software), select the line tool, draw a straight line of unit length on the ruler, calibrate it, draw a straight line between the two vertices that need to be measured, and press Ctrl+M to measure the length; similarly, select the angle measurement tool, select three points on the angle that needs to be measured, and obtain the angle value.

[0069] Step 7: Compare the data obtained in Step 6 with the database formed by summarizing a large number of previous experiments, and adjust the welding process parameters accordingly.

[0070] like Figure 4 As shown, high-speed photography captures and extracts images (a), then a weld bead parameter diagram (b) is obtained. The image is then processed by an intelligent image recognition system to obtain a simplified cross-sectional diagram of the wire feed end of the weld bead (c). Parameter values ​​measured using this simplified cross-sectional diagram (d) are as follows: First, ensure the welding wire is positioned symmetrically in the middle of the weld; measure the ΔH value and compare it with the previous experimental database to check if the ΔH value is within the acceptable range for good welding; measure whether the angles α and β between the molten pool and the sidewall are acute angles, satisfying the requirement for forming a U-shaped groove conducive to the next weld pass. By measuring these values ​​and comparing them with the experimental database, if large fluctuations in the measured values ​​are found during welding, the welding process parameters are adjusted promptly.

[0071] Figure 5 The upper half of the image shows the workpiece surface, and the lower half shows the internal image during the welding gap. In the image, ΔH represents the distance between the welding wire and the workpiece surface. From... Figure 5 It can be seen that when ΔH is 0.5mm, because the tip of the welding wire is close to the workpiece surface, the laser energy is focused on both the welding wire and the workpiece, causing them to melt simultaneously, resulting in a large molten pool area. During welding, because the distance between the tip of the welding wire and the molten pool is small, the molten pool covers the tip of the welding wire, and the welding wire is inserted into the molten pool. The laser heats the welding wire within the molten pool, causing it to melt and thus weld. At this time, there is no obvious droplet transition during the welding process; the droplet transition is a spreading transition. When ΔH is 1 or 1.5mm, the tip of the welding wire is a certain distance from the work surface. The laser irradiates the welding wire, causing its tip to melt and form a droplet. Part of the energy heats the workpiece, causing it to melt as well. There is a small gap between the molten pool and the tip of the welding wire. During welding, a small droplet is formed at the tip of the welding wire. Before the droplet completely detaches from the welding wire, it comes into contact with the molten pool. Due to surface tension, the dynamically moving molten pool has an adsorption effect on the incomplete droplet. The droplet smoothly transitions from the tip of the welding wire into the interior of the molten pool. This can be observed under high-speed photography, resulting in less spatter during welding. When ΔH is 2mm, a large bulge is generated in the middle of the weld, which is not conducive to the next filler welding and will form defects such as incomplete fusion between layers and porosity.

[0072] Figure 6The upper half shows the internal image during the welding interval, and the lower half is a cross-sectional view of the welded section after welding is completed. From Figure 6 As can be seen, at lower frequencies, the molten pool exhibits smaller fluctuations during oscillation welding. During filler wire welding, the laser's dwell time at the wire tip is short, insufficient to fully melt the wire, resulting in a spreading transition of the molten droplets into the pool. With increasing oscillation frequency, the molten pool becomes more fluid, and the liquid weld metal, under the influence of the oscillating laser, tends to move towards the sidewall. Further increases in oscillation frequency lead to more intense molten pool fluctuations. The droplet transition from spreading to bridging can be observed in high-speed photographs. The oscillation of the light source further promotes the fluidity of the molten pool, and the laser spot moves along a counter-clockwise circular path. At different oscillation frequencies, the dwell time of the heat source per unit area decreases with increasing oscillation frequency. When welding at a 20Hz oscillation frequency, the molten pool shows almost no fluctuation. With increasing oscillation frequency, a more pronounced tendency for the molten pool to move towards the bevel sidewall becomes more apparent.

[0073] Compared to traditional laser filler welding, oscillating laser welding significantly reduces the protrusion above the weld bead after filler welding, resulting in a U-shaped bevel. Research shows that at lower laser oscillation frequencies, the laser heat source heats the same area for a longer time per unit time, leading to more concentrated energy. Furthermore, due to smaller molten pool fluctuations, heat conduction to areas further away from the weld pool is weaker, causing heat concentration in the middle of the weld, resulting in cracks in the center of the filler weld. As the oscillation frequency increases, the molten pool is heated more uniformly, the heat source's residence time at various locations decreases, molten pool fluctuations are more pronounced, and the tendency for heat to transfer to the sidewalls is more significant, eliminating weld cracks. When the oscillation frequency increases to 40Hz and 60Hz, a U-shaped weld bead favorable for subsequent filler welding can be obtained. However, at an oscillation frequency of 40Hz, cold cracking defects appear at the bottom of the molten pool, causing welding defects. Figure 6 The top view of the lower half of the weld shows varying degrees of bulges and depressions at laser beam oscillation frequencies of 20Hz and 80Hz. Combined with the weld cross-section diagram, it can be seen that at an oscillation frequency of 80Hz, excessively high frequencies result in continuous bulges in the middle of the weld, which can easily lead to incomplete fusion and porosity defects in the subsequent filler wire welding. The figure shows that as the laser oscillation frequency increases, the weld width increases linearly, while the weld depth initially decreases and then remains at the same level. This indicates that the oscillation frequency has a significant impact on the weld width, and the weld formation coefficient shows an increasing trend. Furthermore, the weld formation coefficient is approximately the same at laser oscillation frequencies of 40Hz and 60Hz. In conclusion, at an oscillation frequency of 60Hz, no obvious defects appear in the weld, and the resulting U-shaped cross-section is beneficial for the subsequent filler wire welding. The weld formation quality is optimal at an oscillation frequency of 60Hz.

[0074] from Figure 7 and 8It can be seen that when the laser spot diameter is small, the laser source is mainly concentrated at the tip of the welding wire, allowing the welding wire to fully melt and form droplets, which then stably transition to the molten pool. At this time, the molten pool fluctuation amplitude is small, and the heat source area covered by the laser is small. When the laser spot diameter is 1mm and 1.5mm, the melting of the welding wire by the laser beam is similar. While heating the welding wire to melting, the laser beam transfers most of its energy to the interior of the molten pool to obtain a larger melt depth and width. At this time, the droplet transition is mainly a liquid bridge transition. As the laser spot diameter further increases, at a laser spot diameter of 2mm, some of the welding wire tip melts but does not melt, and enters the molten pool with the movement of the welding wire. At this time, the droplet transition is a spreading transition. The area heated by the light source to the molten pool further increases, the droplet transition is more inclined to a spreading transition, the melting range of the molten pool becomes larger, and the molten pool fluctuation amplitude increases significantly. High-speed photography images show that, with a constant oscillation frequency, when welding with a small spot diameter, the molten droplet completely melts before entering the molten pool and enters the molten pool in a bridging manner. When welding with a larger oscillation amplitude, as the oscillation diameter of the spot increases, the laser not only acts on the welding wire but also on the molten pool. Before the molten droplet is fully formed, the molten droplet, along with the unmelted welding wire, is incorporated into the molten pool.

[0075] When the laser spot diameter is small, the laser energy is more concentrated, primarily at the tip of the welding wire. The molten wire droplets directly merge into the weld pool, allowing the laser to penetrate deeper and achieve greater weld depth. This wire melting behavior hinders the escape of internal pores, resulting in smaller pores between weld layers. As the oscillation amplitude increases, the heating time per unit area decreases. The laser first acts on the welding wire, with some laser light penetrating the weld pool, resulting in a shallower weld pool. This effect becomes more pronounced with further increases in the laser spot diameter. In narrow-gap laser wire-filling welding, the change in laser spot diameter should be adapted to the change in bevel width. During welding, as the welding height increases, the bevel width increases accordingly, and the laser spot diameter should also increase accordingly. Figure 8 The top view of the weld shows that when the oscillation amplitude is 0.5 mm, the melting width is relatively narrow, and defects such as undercut and spatter are prone to occur during cap welding. When the oscillation amplitude is 1 mm, 1.5 mm, and 2 mm, the weld formation quality is better, with no obvious defects, and it is suitable for welding.

[0076] Figure 9 High-speed cameras were used to capture images of the effect of welding power on the movement of the molten pool and the melting state of the welding wire under the bevel filling state. Figure 9 The upper half is a schematic diagram taken during the movement of the weld pool within the bevel, and the lower half is a cross-sectional view of the welded section after welding is completed. From Figure 9The upper part of the image shows that at lower welding power, the welding wire is not completely melted before entering the molten pool, and the wire mainly relies on the heat from the laser and the molten pool for melting. When the welding heat input is 3.6 kJ / cm, high-speed photography shows that small droplets form at the tip of the welding wire. As these droplets approach the molten pool, they are attracted by surface tension and enter the pool, resulting in minimal fluctuations in the molten pool. With further increases in laser energy, the melting of the welding wire tip into droplets can be clearly observed above the molten pool. At a laser power of 3.92 kW, the welding wire is completely melted above the molten pool. At this point, the droplets enter the pool under gravity, causing significant fluctuations and large spatter upon entry. Furthermore, the welding heat input (3.92 kJ / cm) is high at this point, exceeding the heat required for melting at this wire feed and welding speed. At a welding power of 3.8 kW, the welding wire melting requirements are fully met, and there is no significant spatter during welding. It can be seen that when the power is low, the droplet transition is a spreading transition. As the power increases, the droplet transition changes from a liquid bridge transition to a particle transition.

[0077] When the power value is low, the weld area is small, and small protrusions easily form in the weld after welding, which is not conducive to the subsequent filler wire welding and easily causes welding defects. When the protrusion is too high, the filler wire is prone to sinking to the bottom before melting, causing welding instability. As the power increases, the cross-sectional area of ​​the weld increases significantly. This is because with higher power, the molten pool can achieve a greater penetration depth and width. As shown in the figure, the penetration depth and width of the weld increase linearly with the increase of welding power, and the power has a greater impact on the width. This indicates that when the power is loaded to a certain extent, the change in the weld is mainly reflected in the melting width. When the power is high, the weld formed after welding gradually becomes smoother. Under the condition that other parameters remain unchanged, increasing the power also has the effect of improving the weld morphology. However, due to the large welding heat input, welding hot cracks appear in the middle of the weld. It can be seen that when the welding heat input is 3.92kJ / cm, internal defects in the weld are easily caused. From the weld formation coefficient in the figure, it can be seen that under the condition of increasing welding power, the weld formation coefficient shows a trend of first decreasing and then increasing. In conclusion, a weld with good forming quality can be obtained when the welding power is 3.8kW.

[0078] Figure 10The upper part is a schematic diagram taken during the movement of the weld pool within the bevel, and the lower part is a cross-sectional view of the welded section after welding is completed. The diagram shows that at lower welding speeds, the welding heat input decreases from 4.89 kJ / cm to 3.42 kJ / cm per unit length due to the decreasing volume of the welding wire per unit length with increasing welding speed, resulting in a significant range of welding heat input. At lower welding speeds, most of the welding wire begins to melt near the weld pool, with a spreading transition in the droplet transfer pattern. Because the welding wire enters the weld pool in an incompletely melted state, the weld pool fluctuates considerably. As the welding speed further increases, the welding wire gradually melts further away from the weld pool, and the transition pattern when the droplet enters the weld pool is a liquid bridge transfer. At welding speeds of 0.48 and 0.54 m / min, the weld pool movement is relatively stable. As the welding speed increases further, the liquid welding wire metal required to fill the bottom molten pool is not completely filled before the welding wire is fully melted. That is, the welding speed is greater than the wire feeding speed required to fill the weld. At this time, the movement state of the molten pool is unstable and there is a tendency to deviate towards the weld. At this time, the droplet transfer form is particle transfer.

[0079] As shown in the figure, the welding quality is good when the welding speed is between 0.42 m / min and 0.6 m / min. After welding, no defects such as incomplete fusion between layers or incomplete fusion of the sidewalls occur on the cross-section. With the increase of welding speed, the height of the filler wire weld layer tends to decrease, and the penetration depth decreases from 3.97 mm to 2.9 mm. The weld width variation range is small, decreasing from 3.87 mm to 3.51 mm. The fusion width fully meets the welding requirements at this time. However, as the number of weld beads increases, the weld width at this time is insufficient to fuse the sidewalls. In the actual welding process, the change in groove width needs to be considered. Moreover, when the welding speed is low, the weld layer stack height is large, forming a large protrusion on the weld surface. This form is not conducive to the subsequent filler wire welding. With the increase of welding speed, the protrusion gradually decreases. As shown in the figure, with the increase of welding speed, the weld penetration depth and weld width show a linear decreasing trend. At this time, the influence of welding speed on the penetration depth and weld width is approximately the same. As the welding speed increases, the weld formation coefficient increases. In the process of multi-layer filler wire welding, the increase in welding speed will lead to a decrease in the weld height. The height of each layer is about 3 to 4 mm. As the number of welding layers increases, the width of the bevel also increases. When welding again using the same welding parameters, the actual stacking height of each weld layer will be reduced. At this time, the wire feed speed should be adjusted appropriately to meet the welding requirements.

[0080] Figure 11The upper part is a schematic diagram taken during the movement of the weld pool within the bevel, and the lower part is a cross-sectional view of the welded section after welding is completed. The schematic diagram taken during the movement of the weld pool within the bevel has the opposite effect of changing the welding speed. For a unit length, the volume of the welding wire fed increases with the increase of the wire feed speed, while the welding heat input remains at a certain value. This has a certain impact on the melting of the welding wire tip. The heat per unit volume of the welding wire increases, but decreases with increasing wire feed speed, which also affects the melting of the welding wire. As shown in the figure, when the wire feed speed is low, most of the welding wire begins to melt away from the molten pool. As the welding wire moves into the molten pool, the droplet transition is a liquid bridge transition. At this time, due to the small fluctuation of the welding molten pool, as the wire feed speed further increases, the welding wire cannot fully form a droplet transition before entering the molten pool. At this time, the droplet transition gradually changes from a liquid bridge transition to a spreading transition, and the molten pool fluctuates discontinuously due to the influence of the unmelted welding wire. The welding process is seriously unstable. It can be concluded that the energy generated by the laser beam is insufficient to support a large wire feed speed.

[0081] from Figure 11 As can be seen, when the wire feed speed is low, sidewall incomplete fusion defects are prone to occur. Due to the small filler amount, the weld metal pool cannot fully melt the sidewall, causing this defect. Furthermore, a high bulge appears in the center of the weld bead after welding, which is detrimental to subsequent filler wire welding. After welding, the weld bead morphology gradually becomes U-shaped, and the bulge caused by the low filler wire speed gradually decreases. As the filler wire amount increases per unit length, the weld bead area expands. However, when the wire feed speed reaches 5.0 m / min, a crack appears in the center of the weld bead. At this point, due to the large volume of the wire stack, a welding defect occurs. Within the range of 4.6 m / min to 4.8 m / min, the welding quality is better, and no defects such as interlayer incomplete fusion or sidewall incomplete fusion occur on the welded cross-section. With the increase of the wire feed speed, the filler wire welding layer height tends to gradually increase, the penetration depth increases from 3.10 mm to 3.5 mm, and the weld width variation range also increases from 3.44 mm to 3.75 mm. At this point, the weld penetration and width fully meet the welding requirements. As the number of weld layers increases, the groove width increases, and appropriately increasing the welding speed is sufficient to meet the weld filling requirements. As shown in the figure, with the increase of wire feed speed, the weld penetration and width increase linearly. At this point, the effect of wire feed speed on the weld penetration and width is approximately the same. Simultaneously, the weld formation coefficient decreases, indicating that increasing the wire feed speed leads to a decrease in the weld formation coefficient. Overall, this has a significant impact on weld formation quality. In conclusion, when other welding parameters remain constant, a wire feed speed between 4.6 and 4.8 m / min results in better weld filling. In single-pass multi-layer welding, the weld groove's requirement for wire filling volume needs to be considered.

[0082] exist Figures 5 to 11In this context, T represents time, T0 represents any time, and ms represents microseconds.

[0083] The above description is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several improvements without departing from the principle of the present invention, and these improvements should also be considered within the scope of protection of the present invention.

Claims

1. A method for identifying annular narrow-gap laser filler wire welding using high-speed photography monitoring, characterized in that... Includes the following steps: S1. Real-time acquisition of images showing the movement of the molten pool, the melting state of the welding wire, and the transition of molten droplets during welding; during acquisition, the acquisition device, laser beam, and welding wire end remain relatively stationary while the workpiece to be welded rotates. S2. Preprocess the acquired images: S21. Using the weighted average method, the components of the R, G, and B channels of the image are weighted and averaged with different weights to obtain a more reasonable grayscale image: L = R * 299 / 1000 + G * 587 / 1000 + B * 114 / 1000; S22. Perform image geometric transformation on the grayscale image; S23. Enhance the image processed in S22 using either the frequency domain method or the spatial domain method; S3. Identify the edge contour of the molten droplet using image processing algorithms: After inputting the preprocessed image into the CNN network, the image features are extracted through multiple convolutional pooling operations, and then the image features are sent into the fully connected layer network to complete the image classification and recognition. S4. Extract the image identified by S3 and process it to obtain a simplified cross-sectional view of the wire feed end of the weld bead: S41. Transmit the image to the recognition system for grayscale processing; S42. Perform binary conversion on the grayscale image; S43. Find and draw outlines on the processed image; S5. Measure the distance between the welding wire and the two sides of the weld bead, the distance between the welding wire and the workpiece to be welded position, the angle between the molten pool and the side wall, and the relative position of the laser source and the welding wire in the simplified diagram of S4; S6. Adjust the welding parameters based on the data measured in S5.

2. The method for identifying annular narrow-gap laser filler wire welding using high-speed photography monitoring according to claim 1, characterized in that: When acquiring images, a laser light source is used as the backlight.

Images