Real-time monitoring system and method for processing status of laser cladding head

Abstract

The present invention discloses a real-time monitoring system and method for the processing state of a laser cladding head. The system includes: a laser head unit, which performs laser cladding processing in the form of powder feeding or wire feeding; a coaxial monitoring unit, which transmits image signals emitted by a molten pool; an industrial camera unit, which collects molten pool images; an image online processing unit, which removes noise interference in the molten pool images collected by the industrial camera unit and extracts key image features; a protective lens state recognition unit, which identifies and judges the degree of contamination of the laser head protective lens; a powder feeding state recognition unit, which identifies the state of the laser head powder feeding nozzle, and judges the powder feeding state and whether there is a powder blocking problem. The present invention can monitor the state of the laser cladding head protective lens and the powder feeding nozzle in real time, accurately identify the lens contamination and nozzle powder blocking caused by long-term processing and accidents, and then make timely adjustments to avoid affecting the processing quality. It has strong applicability, high degree of integration, and low implementation cost.

Description

Technical Field

[0001] The present invention belongs to the technical field of online monitoring of laser processing, and specifically, is a real-time monitoring system and identification method for the processing state of a laser cladding head used in processing technologies such as surface cladding and deposition forming. Background Art

[0002] Laser cladding is essentially a directional energy deposition technology based on laser beams. It is an important additive manufacturing technology, mainly used in the fields of direct forming of metal materials, functional coating cladding and high-end parts repair. In current industrial applications, laser cladding is often a layer-by-layer process with a high number of cycles over a long period of time, which makes real-time monitoring of the status of the laser cladding head very necessary. If the laser head can be adjusted in time when it fails, it can avoid waste of resources, improve production efficiency and enhance processing reliability.

[0003] When using a laser cladding head for processing, the state of the protective lens and the state of the powder feeding nozzle are two important factors that affect the use effect of the laser head, leading to a decline in processing quality or even failure. Factor 1: The protective lens is placed between the laser focusing lens and the processing molten pool to prevent powder splashing and steam contamination of the focusing lens during processing. It is a consumable part and needs to be replaced according to the state of the lens. When the protective lens of the laser head is dirty, on the one hand, it will interfere with the energy transfer of the laser beam, causing the actual laser power used for processing to deviate, interfering with the processing quality. On the other hand, it will cause energy to accumulate in the protective lens, causing the lens to heat up and break, and may even damage important optical components such as the focusing lens. Factor 2: The state of the powder feeding nozzle is directly related to the nozzle powder discharge and powder convergence. Especially when processing high-reflective materials such as copper alloys and aluminum alloys, nozzle blockage is more common and often occurs during the processing process. When the nozzle is blocked and the powder cannot be gathered, the cladding layer will collapse and the processing quality will be significantly reduced. If it is not cleared for a long time, it will cause the overall wear of the outer nozzle of the laser head or even scrap.

[0004] Although there are some methods that can be used to monitor the status of the laser head protective lens, most of these methods require additional configuration of dedicated modules to implement, which is costly and system redundant. The status of the powder feeding nozzle is still difficult to monitor in real time. Therefore, in the context of the rapid development of laser head coaxial monitoring technology, a new method that can integrate laser cladding head status monitoring into coaxial monitoring is urgently needed.

[0005] In summary, the state of the cladding head during laser cladding will affect the processing quality and even lead to processing failure and waste of resources. In addition, the current online monitoring cost of the laser cladding head state is high and the system is redundant. It is necessary to propose a low-cost, fast, and widely applicable real-time monitoring system and method for the laser cladding head processing state that can utilize the laser head coaxial monitoring technology, so as to visualize the cladding head state during continuous processing, adjust the process path parameters in time, and troubleshoot. Summary of the invention

[0006] In view of this, the purpose of the present invention is to provide a real-time monitoring system and method for the processing status of a laser cladding head, which does not require an additional dedicated module and can be combined with the current coaxial monitoring technology. It has a low cost, a simple system, a fast processing speed, and a wider applicability. It can ensure the processing quality and avoid waste of resources to solve the deficiencies in the prior art. Moreover, the protective lens status recognition unit is not only suitable for laser cladding, but can also be used in laser beam processing technologies such as laser marking, laser welding, and laser cutting, with a wider applicability; the powder feeding status recognition unit is particularly suitable for laser processing of high-reflective materials such as aluminum alloys and copper alloys, and accurately monitors whether the powder feeding port is blocked, ensuring the processing quality and avoiding waste of resources.

[0007] In order to achieve the above object, the object of the present invention is achieved through the following technical solutions:

[0008] On the one hand, a real-time monitoring system for processing status of a laser cladding head is provided, comprising:

[0009] The laser head unit includes a laser head, which performs laser cladding processing in the form of powder feeding or wire feeding to generate a liquid metal molten pool;

[0010] A coaxial monitoring unit transmits image signals from the liquid metal pool;

[0011] An industrial camera unit for collecting the molten pool image transmitted by the coaxial monitoring unit;

[0012] An image online processing unit removes noise interference in the molten pool image collected by the industrial camera unit and extracts required key image features;

[0013] a protective lens state recognition unit, which recognizes and determines the degree of contamination of the laser head protective lens according to the image data processed by the image online processing unit; and

[0014] The powder feeding status recognition unit recognizes the status of the powder feeding nozzle of the laser head according to the data of the image online processing unit, and determines the powder feeding status and whether there is a powder blockage problem.

[0015] As in the real-time monitoring system for the processing status of the laser cladding head, the laser head unit adopts a coaxial or paraxial method to perform powder feeding or wire feeding processing, and the processing method includes surface cladding or direct forming.

[0016] As described in the real-time monitoring system for the processing status of a laser cladding head, the coaxial monitoring unit includes an arm structure, a 45° beam splitter, a 45° reflector and an arm integrated lens, the 45° beam splitter is located in the laser head, the 45° beam splitter supports light in the band where the laser beam is located to pass forward and light in the non-laser beam band to pass backward, the arm structure is used to transmit the optical path for coaxial monitoring, the 45° reflector is located in the arm structure, the 45° reflector reflects light in the non-laser beam band to the light outlet of the laser head, and the arm integrated lens is integrated on the arm structure to amplify the image for collection by the industrial camera unit.

[0017] As in the real-time monitoring system for the processing status of a laser cladding head, the industrial camera unit may be a common industrial camera, a high-dynamic industrial camera, an infrared industrial camera, or the like.

[0018] As described in the real-time monitoring system for the processing status of a laser cladding head, the image online processing unit includes a grayscale processing module, an image filtering and denoising module, and a visual feature extraction module. The grayscale processing module compresses the grayscale distribution range in the grayscale histogram of the original image to at least 1 / 2 of the original image. The image filtering and denoising module removes interference caused by scattering between the laser beam and the powder, and removes powder splashes with pixel values ​​less than 3 in the powder splashes. The visual feature extraction module extracts image clarity features and laser light outlet contour features.

[0019] For example, in the real-time monitoring system for the processing state of a laser cladding head, the image clarity feature is calculated using a grayscale gradient algorithm in the spatial domain or a grayscale gradient algorithm in the frequency domain.

[0020] As in the real-time monitoring system for the processing status of a laser cladding head, the contour features of the laser light outlet are extracted and calculated using an edge detection algorithm.

[0021] On the other hand, a method for real-time monitoring of the processing state of a laser cladding head is provided, which is implemented based on the real-time monitoring system for the processing state of a laser cladding head as described above, and comprises:

[0022] a. Place the industrial camera on the arm structure of the coaxial monitoring unit, adjust the arm integrated lens in the coaxial monitoring unit for focusing. The distance between the laser head and the substrate during focusing is the actual processing distance. After focusing, calibrate the ratio of the image to the actual size. The ratio of the image pixel value to the actual size is n:1.

[0023] b. During laser cladding head processing, the industrial camera collects images of the processing process in real time, with a frame rate range of 5 to 200fps;

[0024] c. The image online processing unit pre-processes the collected image, including grayscale processing, image filtering and noise reduction, clarity feature extraction, and laser light outlet contour feature extraction, with a processing speed of 5 to 200 ms;

[0025] The clarity feature extraction obtains the real-time clarity feature value of the captured image, that is, the average grayscale gradient value t;

[0026] The contour feature of the laser light outlet is extracted to obtain the contour length pixel value l of the laser light outlet;

[0027] d. According to the calibration ratio n:1 in step a and the processing of the characteristic image by the image online processing unit in step c, the image clarity characteristic value t during laser processing is collected, and a function fitting y=ax+b is performed with m consecutive points to collect the actual contour length L=l / n during laser processing, and the actual contour length change ratio β=L / L is calculated. 1 , where L 1 is the actual contour length of the laser light outlet;

[0028] e. According to the value of a in the function y=ax+b in step d, determine the change of the curvature a of the fitting function when the laser head lens is contaminated, and give the range of a value when the lens is normal (a 1 ,a 2 ), thereby realizing the protection lens status recognition;

[0029] f. According to step d, the ratio of contour length change during laser processing is β = L / L 1 , determine the change of the profile length change ratio β when the laser head powder feeding nozzle is blocked, and give the β value range (β 1 ,β 2 ), thereby realizing powder feeding status identification.

[0030] The beneficial effects of the technical solution of the present invention are:

[0031] 1. It can monitor the contamination status of the protective lens of the laser cladding head in real time, accurately identify the lens contamination caused by long-term processing and accidents, and then adjust the protective lens in time to avoid affecting the processing quality or even processing failure;

[0032] 2. It can monitor the powder feeding status of the laser cladding head nozzle in real time. When processing continuously for a long time, especially when processing high-reflectivity materials continuously, it can accurately identify whether the nozzle is blocked and remind the operator to perform technical processing to avoid affecting the processing quality or even processing failure;

[0033] 3. Strong applicability. The protective lens status recognition unit is suitable for almost any laser beam processing technology such as laser cladding, laser marking, laser welding, laser cutting, etc. At the same time, it is not limited by the properties, size, surface state and other issues of the feed material or base material;

[0034] 4. It has a high degree of integration and can utilize the coaxial monitoring capability of the laser head. It does not require any new hardware equipment and has a low implementation cost. It can monitor the status of the protective lens and the powder feeding status in real time by relying only on the coaxial images collected in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] To further illustrate the above-mentioned objectives, structural features and effects of the present invention, the present invention will be described in detail below with reference to the accompanying drawings.

[0036] Figure 1 This is a schematic diagram of the system structure of a preferred embodiment of the present invention;

[0037] Figure 2 This is a schematic block diagram of the system structure of a preferred embodiment of the present invention;

[0038] Figure 3 This is a schematic block diagram of the system flow of a preferred embodiment of the present invention;

[0039] In the figure: 1. Laser head unit; 101. Laser head; 102. Light outlet; 103. Laser head protection lens; 104. Focusing lens; 105. Laser head powder delivery nozzle; 2. Coaxial monitoring unit; 201. 45° beam splitter; 202. Arm structure; 203. 45° reflector; 204. Arm integrated lens; 3. Industrial camera unit; 4. Image online processing unit; 401. Grayscale processing module; 402. Image filtering and denoising module; 403. Visual feature extraction module; 5. Protection lens status recognition unit; 6. Powder delivery status recognition unit; 7. Computer. DETAILED DESCRIPTION

[0040] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0041] See also Figure 1 , Figure 2 , Figure 3As shown, the real-time monitoring system for the processing state of the laser cladding head of the present invention includes a laser head unit 1, a coaxial monitoring unit 2, an industrial camera unit 3, an image online processing unit 4, a protective lens state recognition unit 5 and a powder feeding state recognition unit 6. The laser head unit 1 includes a laser head 101, and laser cladding processing is performed in the form of powder feeding or wire feeding. The coaxial monitoring unit 2 transmits the image signal emitted by the molten pool, and the industrial camera unit 3 collects the molten pool image transmitted by the coaxial monitoring unit 2. The image online processing unit 4 removes the noise interference in the molten pool image collected by the industrial camera unit 3, and extracts the required key image features. The protective lens state recognition unit 5 identifies and judges the degree of contamination of the laser head protective lens according to the image data processed by the image online processing unit 4, and the powder feeding state recognition unit 6 identifies the state of the laser head powder feeding nozzle according to the data of the image online processing unit 4, and judges the powder feeding state and whether there is a powder blocking problem. It can be understood that the above-mentioned image online processing unit 4, the protective lens state recognition unit 5 and the powder feeding state recognition unit 6 are all implemented on the computer 7.

[0042] Specifically, the laser head unit 1 uses a circular powder feeding form to directly form a thin-walled structure by laser cladding to generate a liquid metal molten pool. The coaxial monitoring unit 2 transmits the 650-750nm band image signal emitted by the above molten pool to the unit light outlet via the light outlet 102 of the laser head 101, the laser head protection lens 103 located near the light outlet 102 (located on its side is the focusing lens 104), the spectroscopic lens, the reflector, the lens, etc., so that the signal during laser processing can be collected in real time. The industrial camera unit 3 is a black and white industrial camera that can collect the signal transmitted by the coaxial monitoring unit 2 and convert it into image information. The image online processing unit 4 can remove the noise interference such as powder scattering and splashing in the molten pool image collected by the industrial camera unit 3, and extract the key image features required by the protective lens state recognition unit 5 and the powder feeding state recognition unit 6, so that the subsequent recognition has a more reliable feature source. The protective lens state recognition unit 5 can identify and judge the degree of contamination of the laser head protective lens 103 according to the image data processed by the image online processing unit 4, so that the degree of contamination of the laser head protective lens 103 can be confirmed during the continuous processing without stopping the process for observation. The powder feeding state recognition unit 6 can identify the state of the laser head powder feeding nozzle 105 according to the data of the image online processing unit 4, and judge the powder feeding state at this time, whether there is a powder blockage problem, so as to make the processing process more reliable and reduce accidents caused by powder blockage.

[0043] Continue to view Figure 1As shown, the coaxial monitoring unit 2 includes a 45° beam splitter 201 inside the laser head 1, a crank arm structure 202, a 45° reflector 203 inside the crank arm, and a crank arm integrated lens 204. The 45° beam splitter 201 inside the laser head 1 can pass light in the wavelength band of 980-1020nm laser beam in the forward direction, and pass light in the non-laser beam wavelength band of 650-750nm in the reverse direction. The crank arm structure 202 can be used to transmit the optical path for coaxial monitoring, and the 45° reflector 203 inside the crank arm can reflect non-laser beam wavelength band light to the light outlet of the coaxial monitoring unit. The crank arm integrated lens 204 is integrated on the crank arm structure 202, and can magnify the image by 0.49 times to facilitate the acquisition of the industrial camera unit 3.

[0044] The image online processing unit 4 includes a grayscale processing module 401, an image filtering and denoising module 402 and a visual feature extraction module 403. The grayscale processing module 401 can compress the grayscale distribution range in the grayscale histogram of the original image to 1 / 3 of the original image, making subsequent images easier to process. The image filtering and denoising module 402 can remove the interference caused by the scattering effect between the laser beam and the powder, and at the same time remove the powder splash with a pixel value less than 3 in the powder splash, making subsequent identification more convenient. The visual feature extraction module 403 can extract the image clarity features and the laser light outlet contour features.

[0045] Generally, the image clarity feature can be calculated using the Tenengrad gradient algorithm, and the laser light outlet contour feature can be obtained using the Canny edge detection algorithm, but the present invention is not limited thereto.

[0046] See also Figure 3 As shown, the specific steps of the real-time monitoring method of the laser cladding head processing state include:

[0047] a. Place the black and white industrial camera on the arm structure of the coaxial monitoring unit. The distance between the powder flow convergence point of the equipment and the laser head is 12mm. Adjust the integrated lens of the arm in the coaxial monitoring unit to focus at this distance. When focusing, reduce the aperture as much as possible to increase the depth of field and ensure the imaging effect. Then calibrate the ratio of the image to the actual size. The ratio of the image pixel value to the actual size 1mm is 35:1.

[0048] b. During laser cladding head processing, the black and white industrial camera collects images of the processing process in real time, with a frame rate of 100fps;

[0049] c. The image online processing unit pre-processes the collected images, including grayscale processing, image filtering and noise reduction, clarity feature extraction, and laser light outlet contour feature extraction, with a processing speed of 100ms;

[0050] The clarity feature extraction obtains the real-time clarity feature value of the captured image, that is, the average grayscale gradient value t;

[0051] The contour feature of the laser light outlet is extracted to obtain the contour length pixel value l of the laser light outlet;

[0052] d. According to the calibration ratio of 35:1 in step a and the processing of the characteristic image by the image online processing unit in step c, the image clarity characteristic value t during laser processing is collected, and a function fitting y=ax+b is performed with 10 consecutive points to collect the actual contour length L=l / 35 during laser processing, and the actual contour length change ratio β=L / L is calculated. 1 , where L 1 The actual outline length of the laser head light outlet is about 16mm;

[0053] The powder and matrix materials used in this embodiment are both AlSi7Mg, the powder diameter is 50-105um, and the matrix size is 120×30×10mm.

[0054] The data results in this example are shown in the following table:

[0055] Table 1

[0056]

[0057]

[0058] e. According to the value a in the function y=ax+b in step d, determine the change of the curvature a of the fitting function when the laser head lens is contaminated. The value a range is (-0.3, 0.3) when the protective lens in the embodiment is normal, thereby realizing the state identification of the protective lens. In this embodiment, the lens is not contaminated;

[0059] f. According to the contour length change ratio β=L / 16 during laser processing in step d, the change of the contour length change ratio β when the laser head powder feeding nozzle is blocked is determined, and the β value range during normal powder output is given (0.93, 1.07), thereby realizing powder feeding status identification. In this embodiment, the nozzle does not get blocked.

[0060] This embodiment is applicable to AlSi7Mg substrate and AlSi7Mg powder, and the substrate is a horizontal plane, perpendicular to the laser head. The embodiment obtains a calculation method for judging the contamination state of the protective lens and the powder feeding state of the nozzle according to different coaxial images, and clarifies the recognition feature threshold range under this working condition.

[0061] The present invention does not require an additional dedicated module and can be combined with the current coaxial monitoring technology to monitor the contamination state of the protective lens of the laser cladding head and the powder feeding state in real time according to the coaxial image characteristics. It has low cost, simple system, fast processing speed, and the protective lens state recognition unit is not only suitable for laser cladding, but also can be used in laser beam processing technologies such as laser marking, laser welding, and laser cutting, with wider applicability; the powder feeding state recognition unit is particularly suitable for laser processing of high-reflective materials such as aluminum alloys and copper alloys, accurately monitoring whether the powder feeding port is blocked, ensuring processing quality, and avoiding waste of resources. The present invention can also provide important references for the fields of laser welding, laser cutting, etc.

[0062] The above are only preferred embodiments of the present invention, and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for real-time monitoring of the processing status of a laser cladding head, characterized in that: include: a. Place the industrial camera on the arm structure of the coaxial monitoring unit, adjust the arm integrated lens in the coaxial monitoring unit for focusing. The distance between the laser head and the substrate during focusing is the actual processing distance. After focusing, calibrate the ratio of the image to the actual size. The ratio of the image pixel value to the actual size is n:

1. b. During laser cladding head processing, the industrial camera collects images of the processing process in real time, with a frame rate range of 5 to 200fps; c. The image online processing unit pre-processes the collected image, including grayscale processing, image filtering and noise reduction, clarity feature extraction, and laser light outlet contour feature extraction, with a processing speed of 5 to 200 ms; The clarity feature extraction obtains the real-time clarity feature value of the captured image, that is, the average grayscale gradient value t; The contour feature of the laser light outlet is extracted to obtain the contour length pixel value l of the laser light outlet; d. According to the calibration ratio n:1 in step a and the processing of the characteristic image by the image online processing unit in step c, the image clarity characteristic value t during laser processing is collected, and a function fitting y=ax+b is performed with m consecutive points to collect the actual contour length L=l / n during laser processing, and the actual contour length change ratio β=L / L1 is calculated, where L1 is the actual contour length of the laser outlet; e. According to the value a in the function y=ax+b in step d, determine the change of the slope a of the fitting function when the laser head lens is contaminated, and give the range of a value (a1, a2) when the lens is normal, thereby realizing the protection lens status identification; f. According to the contour length change ratio β=L / L1 during laser processing in step d, determine the change of the contour length change ratio β when the laser head powder feeding nozzle is blocked, give the β value range (β1, β2) during normal powder output, and then realize the powder feeding status identification.

Images