Laser melting deposition aluminum alloy thin-wall structure morphology defect monitoring system and method

Abstract

The present invention discloses a system and method for monitoring morphological defects of thin-walled structures of laser-melted aluminum alloys, comprising a coaxial monitoring system for laser-melted deposition, an online image processing unit, a morphological defect recognition unit, and an in-situ feedback control unit. The online image processing unit removes noise interference from the collected image signal and extracts key features required by the morphological defect recognition unit. The morphological defect recognition unit identifies and judges morphological defects based on the image data processed by the online image processing unit, determining which type of defect it belongs to and which stage it is in. The in-situ feedback control unit determines whether the above-mentioned defects are feasible to be suppressed through in-situ control based on the results obtained by the morphological defect recognition unit. For suppressible defects, in-situ control based on process path parameters is performed. This monitoring system and method are completed on the basis of existing monitoring systems, are low-cost, and have wide applicability. They can solve the industrial problem of the difficulty in suppressing morphological defects during laser-melted deposition of aluminum alloys.

Description

Technical Field

[0001] The present invention belongs to the field of online monitoring and control of laser additive manufacturing and aluminum alloy laser processing. Specifically, it is a system and method for monitoring the morphology defects of laser melting deposition aluminum alloy thin-wall structures, which can be used for long-cycle, large-scale laser melting deposition forming of aluminum alloy components. Background Art

[0002] Laser melting deposition (LMD) is a directed energy deposition technology utilizing a laser beam. It can be used in areas such as direct forming of complex metal components, surface coating cladding, and remanufacturing and repair. Aluminum alloy is a high-specific strength and high-specific stiffness material that holds a significant position in industry, particularly in aerospace. Thin-walled structures are the foundation of complex structural components and a key advantage of additive manufacturing technology. Therefore, the laser melting deposition of thin-walled aluminum alloy structures, focusing on the three key points mentioned above, has significant application value and practical significance.

[0003] At present, when using laser melting deposition technology to form aluminum alloy thin-walled structures, the main problems faced are as follows: first, aluminum alloy itself has physical properties such as low melting point, high thermal conductivity, and high reflectivity to lasers, which makes the forming quality very sensitive to process path parameters and the process path window is difficult to control; second, long-term multi-layer thermal cycle processing causes the aluminum alloy molten pool temperature to continue to increase, viscosity to decrease, surface tension to decrease, and instability to increase; third, the low heat dissipation capacity of the thin-walled structure and the unconstrained state on both sides make it difficult to maintain the stability of the overall molten pool. The above three problems together lead to the problem that laser melting deposition aluminum alloy thin-walled structures are prone to significant morphological defects, which in turn causes the deposition process to fail.

[0004] In-situ monitoring and control during the laser melting deposition process is an important means to solve the above-mentioned problems. It can effectively suppress the generation of morphological defects and improve processing quality. At present, there are some methods for monitoring and controlling the molten pool during laser melting deposition of materials such as titanium alloys and iron alloys. However, because the metallurgical characteristics of aluminum alloy molten pools are different from those of other alloys, it is difficult to meet the requirements of industrial applications by simply relying on traditional molten pool area, width, temperature, etc. to monitor morphological defects. At the same time, the existing single process parameter control method is also difficult to solve the morphological quality problem of aluminum alloys. Therefore, the current methods are still difficult to apply to the laser melting deposition of aluminum alloys.

[0005] In summary, based on the current urgent demand for laser melting deposition of aluminum alloy thin-walled structures in the industrial field and the insufficiency of existing morphology defect monitoring and control methods, it is necessary to propose a laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system and method, so as to make real-time judgment on the formation and development of morphology defects in the deposition process of aluminum alloy thin-walled structures, provide in-situ control methods, and thus achieve high-quality processing of complex aluminum alloy components. Summary of the Invention

[0006] In response to the urgent demand for aluminum alloy laser melting and deposition components in the industrial field, and the problem of difficulty in suppressing morphological defects when depositing aluminum alloy thin-walled structures, the present invention provides a laser melting and deposition aluminum alloy thin-walled structure morphology defect monitoring system and method. Based on the existing laser melting and deposition coaxial monitoring system and combined with the material metallurgical properties of the aluminum alloy itself, the morphology defects in the laser melting and deposition thin-walled structure process are monitored and regulated. The system has low cost, simple system, fast processing speed, and can be expanded to aluminum alloy overlap cladding, overhang deposition and other processing technologies. It has wide applicability and can solve the shortcomings of the existing technology.

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

[0008] On the one hand, a laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system is provided, including a laser melting deposition coaxial monitoring system, an image online processing unit, a morphology defect recognition unit, and an in-situ feedback control unit. The laser melting deposition coaxial monitoring system includes a laser, a laser head, a displacement device, a material feeding device, a coaxial monitoring unit and an industrial camera unit. The laser beam emitted by the laser is transmitted to the laser head, and then irradiated to the aluminum alloy substrate or the deposition layer, and aluminum alloy powder is fed synchronously to form a liquid aluminum alloy molten pool melt. The laser head is connected to the displacement device for displacement. During the deposition process, the coaxial monitoring unit and the The industrial camera unit collects images of the molten pool and finally transmits the data to the image online processing unit; the image online processing unit removes laser scattering or powder splashing interference from the image signal collected by the industrial camera unit and extracts the required key features. The morphology defect recognition unit identifies and judges the morphology defects based on the image data processed by the image online processing unit to determine which type of defect it belongs to and the stage it is in. The in-situ feedback control unit determines whether the above-mentioned defects are feasible to be suppressed through in-situ control based on the results obtained by the morphology defect recognition unit, and performs in-situ control based on process path parameters for the suppressible defects.

[0009] As described in the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system, 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 the interference caused by scattering between the laser beam and the powder, and simultaneously removes powder splashes with pixel values less than 3 in the powder splashes. The visual feature extraction module extracts the area characteristics of the internal area of the molten pool image, the overall clarity characteristics of the image, and the number characteristics of the molten pool.

[0010] For example, in the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system, the morphology defect recognition unit identifies whether the internal area area characteristics and related characteristics exceed a given threshold, identifies the overall clarity of the molten pool image, and identifies the number of molten pools. Finally, it comprehensively determines whether there is a morphology defect and the stage it is in.

[0011] As described in the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system, the process path parameters of the in-situ feedback control unit include laser power, scanning speed, powder feeding amount, residence time at both ends, interlayer interval time, and lifting amount.

[0012] For example, in the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system, the laser includes a semiconductor laser or an Nd:YAG laser, and the laser is connected to the laser head via an optical fiber connection.

[0013] As in the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system, the material feeding device includes a powder feeding device or a wire feeding device.

[0014] As in the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system, the displacement device includes a CNC machine tool or a robot.

[0015] As described in the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system, the industrial camera unit includes an ordinary industrial camera, a high dynamic industrial camera, an infrared industrial camera, and a hyperspectral industrial camera.

[0016] As in the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system, the coaxial monitoring unit transmits the laser processing molten pool image signal to the industrial camera unit, and the industrial camera unit is located at the unit light outlet of the coaxial monitoring unit.

[0017] As in the laser melting deposition aluminum alloy thin-wall structure morphology defect monitoring system, the displacement device performs 3-axis to 11-axis displacement.

[0018] On the other hand, a method for monitoring morphology defects of a laser melting deposition aluminum alloy thin-walled structure is provided, which is implemented based on any of the above-mentioned laser melting deposition aluminum alloy thin-walled structure morphology defects monitoring systems, comprising:

[0019] a. Dimension calibration: Place the industrial camera at the light outlet of the coaxial monitoring unit, adjust the distance l between the laser head and the substrate to the actual optimal processing distance l', and after focusing, calibrate the ratio of the image to the actual size. The ratio of the image pixel value to the actual size is k:1;

[0020] b. Establishing the definition-distance relationship: Adjust the distance between the laser head and the substrate to l', l'+1, l'+2, l'+3, and l'+4 mm, determine the definition characteristics at different distances between the laser head and the substrate, that is, the average grayscale gradient value h of the image, and establish a mapping relationship between the two: h = f(l);

[0021] c. Image acquisition: When the laser head is processing, the industrial camera collects images of the processing process in real time, with a frame rate range of 5 to 200 fps;

[0022] d. Preprocessing: The image online processing unit preprocesses the collected images, including grayscale processing and image filtering and noise reduction, with a processing speed of 5 to 200 ms;

[0023] e. Feature extraction: Extract the area features of the internal area of the melt pool, the overall image clarity features, and the number of melt pools. The internal area feature of the melt pool is the real-time pixel value s of the internal area of the melt pool, the overall image clarity feature is the average grayscale gradient value h of the image, and the number of melt pools is the number of liquid melt pools within the existing image range q.

[0024] f. Defect type and stage identification: Based on the calibration ratio k:1 in step a and the processing of the feature image by the online image processing unit in steps c-e, the actual area of the molten pool internal area S=s / k is used to determine the state of the laser melting deposition morphology defect; the types of thin-walled structure morphology defects include end tilting, deposition failure, local collapse, and spheroidization separation.

[0025] 1) The inclination at both ends is caused by the instability of the molten pool at both ends of the deposition layer: in the initial stage, the area inside the molten pool is small and cannot maintain stability, that is, the average deviation of S within 0.5s at the start or end of each layer is Among them, m is the number of melt pool images. When it is greater than a given threshold value α1, the initial error of the two-end tilt morphology defect begins to appear; in the development stage, the molten pool area of the single layer starting end gradually increases, and the molten pool area of the ending end gradually decreases. The linear function y=ax+b of the fitting position and the molten pool area, when |a| is greater than a given threshold value α2, the two-end tilt morphology defect occurs at the location and is in the development stage;

[0026] 2) The main reason for the failure of deposition is that the distance l between the laser head and the deposition layer gradually increases: in the initial stage, the average area of the internal area of each layer excluding the starting and ending 0.5s Gradually increase the area change rate of each layer To judge, when θ is greater than a given threshold β1, deposition failure defects begin to appear; in the development stage, the molten pool image gradually moves away from the focus of the industrial camera unit, and the average grayscale gradient value h of each layer image, excluding the 0.5s at the start and end ends, gradually increases. Given the optimal threshold range of h, when h of all images is less than β2, the distance between the laser head and the deposited layer is abnormal, and it has entered the development stage where deposition cannot be achieved;

[0027] 3) Local collapse is mainly caused by the instability of the molten pool morphology: In the initial stage, the internal area S of the single-layer molten pool fluctuates, and the average deviation of S When it is greater than γ1 and less than γ2, it is the initial stage of the morphological defect; when When it is greater than γ2, it is in the development stage;

[0028] 4) Spheroidization separation is mainly caused by the collapse of the sintered powder layer on both sides of the aluminum alloy thin-walled component: in the initial stage, the area inside the molten pool in the single layer will increase significantly, S ≥ δ; in the development stage, the number of molten pools q is greater than 1;

[0029] g. Defect in-situ control decision and execution: In-situ control of morphological defects is performed using an interlayer in-situ method, including:

[0030] 1) For the tilt defects at both ends, when in the initial stage, they can be controlled by increasing the dwell time at both ends in the next layer; in the development stage, the distance l between the laser head and the deposited layer is determined by the given threshold α3. The distance l is calculated based on the relationship in step b. When l ≥ α3, the processing is terminated and cannot be controlled. When l < α3, the first two dwell times are increased and the self-optimization negative feedback is used for adjustment;

[0031] 2) For defects that cannot be deposited, when in the initial stage, the lifting amount can be reduced in the next layer, the laser power can be increased, and the scanning speed can be reduced; when in the development stage, it is necessary to adjust l to a reasonable range in the next layer according to the distance l between the laser head and the deposited layer, and then reduce the lifting amount, increase the laser power or reduce the scanning speed in subsequent processing;

[0032] 3) For local collapse, when it is in the initial stage, it can be solved by reducing the lifting amount in the next layer. When the molten pool stabilizes, continue processing with the original lifting amount. When it is in the development stage, it is necessary to perform only a slight lifting in the next few layers to fill the collapse. When the filling returns to the initial stage, it is executed according to the initial stage control strategy. If it recurs, it is necessary to increase the interlayer interval time or reduce the laser power.

[0033] 4) For spheroidization separation, when it is in the initial stage, it can be regulated by reducing the laser power in the next layer or increasing the interlayer interval time; when it is in the development stage, the processing has failed and the processing can be terminated.

[0034] The beneficial effects of the technical solution of the present invention are:

[0035] 1. It can monitor the processing process of laser melting deposition aluminum alloy thin-walled structures. By identifying the metallurgical state of the aluminum alloy molten pool, it can monitor the initial and development stages of morphological defects. At the same time, it provides an in-situ control method to solve the problem of difficulty in suppressing morphological defects in laser melting deposition aluminum alloy thin-walled structures.

[0036] 2. It has a high degree of integration and strong applicability. It can utilize the coaxial monitoring capability of the laser head and does not require any new hardware equipment, so the implementation cost is low. It can monitor and control by relying only on the coaxial collected images and the image online processing unit, morphology defect recognition unit, in-situ feedback control unit, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] 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.

[0038] Figure 1 This is a schematic block diagram of the structure of a preferred embodiment of the present invention;

[0039] Figure 2a 、 Figure 2b They are respectively a side view and a top view of an aluminum alloy molten pool according to a preferred embodiment of the present invention;

[0040] Figure 3 This is a schematic diagram of the structure of a laser melting deposition aluminum alloy thin-wall structure morphology defect monitoring system according to a preferred embodiment of the present invention;

[0041] Figure 4 This is a flow chart of a method for monitoring and controlling morphology defects of thin-walled aluminum alloy structures deposited by laser melting according to a preferred embodiment of the present invention;

[0042] Figure 5 This is a data table related to the characteristics of the internal area of the molten pool;

[0043] Figure 6 This is a data table related to the overall clarity characteristics of the melt pool image;

[0044] Figure 7 This is a data table related to the quantity characteristics of the molten pool;

[0045] In the figure: 1. Laser melting deposition coaxial monitoring system; 101. Liquid aluminum alloy molten pool melt; 1011. Internal area of the molten pool; 1012. External area of the molten pool; 102. Coaxial monitoring unit; 103. Industrial camera unit; 2. Image online processing unit; 201. Grayscale processing module; 202. Image filtering and noise reduction module; 203. Visual feature extraction module; 3. Morphology defect recognition unit; 4. In-situ feedback control unit. DETAILED DESCRIPTION

[0046] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0047] See Figure 1 As shown, the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system of the present invention includes a laser melting deposition coaxial monitoring system 1, an image online processing unit 2, a morphology defect recognition unit 3, and an in-situ feedback control unit 4. The laser melting deposition coaxial monitoring system 1 includes a laser, a laser head, a displacement device, a material feeding device, a coaxial monitoring unit 102 and an industrial camera unit 103. The laser beam emitted by the laser is transmitted to the laser head and then irradiated to the aluminum alloy substrate or the deposition layer. The aluminum alloy powder is fed synchronously to form a liquid aluminum alloy molten pool melt 101. Figure 2a 、 Figure 2b As shown, 1011 is the inner area of the molten pool, and 1012 is the outer area of the molten pool. The laser head is connected to the displacement device for 3-axis to 11-axis displacement. During the deposition process, the coaxial monitoring unit 102 and the industrial camera unit 103 are used to collect images of the molten pool, and the data is finally transmitted to the image online processing unit 2. The image online processing unit 2 removes the laser scattering or powder splash interference of the image signal collected by the industrial camera unit 103 and extracts the required key features. The morphology defect recognition unit 3 identifies and judges the morphology defects based on the image data processed by the image online processing unit 2, determines the type of defect and the stage it is in. The in-situ feedback control unit 4 determines whether the above defects are feasible to be suppressed through in-situ control based on the results obtained by the morphology defect recognition unit 3, and performs in-situ control based on the process path parameters for the suppressible defects. Specifically, the laser is preferably a semiconductor 450nm blue laser, the laser head is preferably a ring-shaped powder feeding laser head, the displacement device is preferably a three-axis CNC machine tool, and the industrial camera unit 103 is preferably an ordinary color industrial camera.

[0048] In a preferred embodiment, the process path parameters of the in-situ feedback control unit 4 include laser power, scanning speed, powder feed rate, dwell time at both ends, interlayer pause time, and lift amount. The laser comprises a semiconductor laser or an Nd:YAG laser. The connection between the laser and the laser head is preferably an optical fiber connection, and the blue laser beam emitted by the blue laser is transmitted to the laser head via the optical fiber.

[0049] The material feeding device includes a powder feeder or a wire feeder, and the displacement device includes a CNC machine tool or a robot. The industrial camera unit 103 includes a standard industrial camera, a high-dynamic range industrial camera, an infrared industrial camera, and a hyperspectral industrial camera. The coaxial monitoring unit 102 transmits the laser processing melt pool image signal to the industrial camera unit 103, which is located at the unit light output port of the coaxial monitoring unit 102.

[0050] Further, see Figure 1 As shown, the image online processing unit 2 includes a grayscale processing module 201, an image filtering and denoising module 202, and a visual feature extraction module 203. The grayscale processing module 201 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 202 removes interference caused by scattering between the laser beam and the powder, and removes powder splashes with pixel values less than 3. The visual feature extraction module 203 extracts the area characteristics of the internal area of the molten pool, the overall image clarity characteristics, and the number of molten pools.

[0051] In the morphology defect recognition unit 3, it is identified whether the internal area and its associated features exceed a given threshold, the overall clarity of the molten pool image is identified, the number of molten pools is identified, and finally the current deposition state and defect morphology type are comprehensively judged.

[0052] In the in-situ feedback control unit 4, the selectable process path parameters include laser power, scanning speed, powder feeding amount, dwell time at both ends, inter-layer interval time, lifting amount, etc.

[0053] The parameters of the initial stage of deposition processing were set as follows: laser power 900 W, scanning speed 5 mm / s, powder feeding rate 1.1 r / min, dwell time at both ends 0 s, interlayer interval time 5 s, and lifting amount 0.2 mm.

[0054] See Figure 3 、 Figure 4 As shown, the aforementioned laser melting deposition aluminum alloy thin-wall structure morphology defect monitoring system is used, and the monitoring process of the embodiment is as follows:

[0055] a. Place the industrial camera at the light outlet of the coaxial monitoring unit 102, adjust the distance between the laser head and the substrate to l = 12 mm, focus, and calibrate the ratio of the image to the actual size. The obtained ratio of the image pixel value to the actual size is 45 pixels:1 mm;

[0056] b. Establishing the definition-distance relationship: The distance between the laser head and the substrate was adjusted to 12, 13, 14, 15, and 16 mm, and the definition characteristics at different distances between the laser head and the substrate were determined. That is, the average grayscale gradient values of the image were 30.46, 28.65, 25.14, 23.91, and 20.78, respectively. The fitting formula was h = -2.41l + 59.528;

[0057] c. Image acquisition: When the laser head is processing, the industrial camera collects images of the processing process in real time, with a frame rate of 20fps and a resolution of 400×400 pixels. 2 ;

[0058] d. Preprocessing: The image online processing unit 2 preprocesses the collected image, including grayscale processing: gamma transformation value 0.4; image filtering and noise reduction, median filter template 3×3;

[0059] e. Feature extraction: Extract the area feature of the inner area of the molten pool, the overall clarity feature of the image, and the number of molten pools. The inner area feature of the molten pool is the real-time pixel value s of the inner area of the molten pool, the overall clarity feature is the average grayscale gradient value h of the image, and the number of molten pools is the number q of liquid molten pools within the existing image range. The data in the embodiment is the molten pool data when the 100th layer is deposited. For specific data, see Figure 5 、 Figure 6 and Figure 7 form;

[0060] f. Identification of defect type and stage: Based on the calibration ratio of 45:1 in step a and the processing of the characteristic image by the image online processing unit 2 in steps c-e, the laser melting deposition state is judged. Based on the accumulation of previous experiments, the key threshold α1 = 0.5mm is given. 2 ;α2=2; α3=15mm; β1=8%; β2=23.91; γ1=0.3mm 2 ;γ2=1.2mm 2 ;δ=20mm 2 Because morphological defects may occur concurrently, the following judgments are made in sequence:

[0061] 1) Tilt at both ends: Calculate the average deviation of S within 0.5s between the start and end ends

[0062] 2.84 and 1.23mm respectively 2 , both ends have begun to show tilt defects; the fitting functions at both ends are y = 6.499x + 2.1278 and y = -1.3253x + 7.3321, respectively. Therefore, the tilt at the starting end is developing at this time, and the ending end is still in the initial stage.

[0063] 2) Unable to deposit: The area change rate of each layer is 3%, so the undepositability defect in the initial stage does not occur; the grayscale gradient value h is between 25 and 31, and the undepositability defect in the initial stage does not occur.

[0064] 3) Local collapse: The average deviation of S is 1.95mm 2 , which is greater than the γ2 value, and is therefore in the development stage of local collapse.

[0065] 4) Spheroidization separation: The internal area of the molten pool during the entire process is less than 20mm 2 , and the number of molten pools is always 1, so no spheroidization separation defects occur.

[0066] g. In-situ defect control decision-making and execution: Topographic defects were controlled in-situ using an interlayer in-situ method. In this implementation, both end tilt and localized collapse defects occurred. Regarding the end tilt, l was between 12mm and 14mm, less than α3, so it could be adjusted by increasing the dwell time at both ends. Regarding the localized collapse, since it was already in the development stage, it was only lifted by 0.1mm in the next three layers to compensate. When returning to the initial stage, the lift was increased to 0.2mm, and the defect disappeared immediately.

[0067] This embodiment is applicable to blue laser processing of AlSi7Mg substrates and AlSi7Mg-2%TiB2 powder, with the substrate being horizontal and perpendicular to the laser head. This embodiment provides a method for determining morphological defects based on different coaxial images and clarifies the threshold range for identifying features under these conditions.

[0068] The present invention can also provide important reference for aluminum alloy lap cladding, overhang deposition and other processes.

[0069] The above are only preferred embodiments of the present invention and do not limit the implementation mode and protection scope of the present invention. For those skilled in the art, it 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 monitoring morphological defects of laser melting deposition aluminum alloy thin-walled structures, characterized in that: The laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system is implemented based on the laser melting deposition aluminum alloy thin-walled structure morphology defect monitoring system, which includes a laser melting deposition coaxial monitoring system, an image online processing unit, a morphology defect recognition unit, and an in-situ feedback control unit. The laser melting deposition coaxial monitoring system includes a laser, a laser head, a displacement device, a material feeding device, a coaxial monitoring unit and an industrial camera unit. The laser beam emitted by the laser is transmitted to the laser head, and then irradiated to the aluminum alloy substrate or the deposition layer, and aluminum alloy powder is fed synchronously to form a liquid aluminum alloy molten pool melt. The laser head is connected to the displacement device for displacement. During the deposition process, the laser head is used to The coaxial monitoring unit and the industrial camera unit collect images of the molten pool and finally transmit the data to the image online processing unit; the image online processing unit removes laser scattering or powder splashing interference from the image signal collected by the industrial camera unit and extracts the required key features. The morphology defect recognition unit identifies and judges the morphology defects based on the image data processed by the image online processing unit to determine which type of defect it belongs to and the stage it is in. The in-situ feedback control unit determines whether the above-mentioned defects are feasible to be suppressed through in-situ control based on the results obtained by the morphology defect recognition unit, and performs in-situ control based on process path parameters for suppressible defects. The following steps are involved: a. Dimension calibration: Place the industrial camera at the light outlet of the coaxial monitoring unit, adjust the distance l between the laser head and the substrate to the actual optimal processing distance l', and after focusing, calibrate the ratio of the image to the actual size. The ratio of the image pixel value to the actual size is k:1; b. Establishing the definition-distance relationship: Adjust the distance between the laser head and the substrate to l', l'+1, l'+2, l'+3, and l'+4 mm, determine the definition characteristics at different distances between the laser head and the substrate, that is, the average grayscale gradient value h of the image, and establish a mapping relationship between the two: h = f(l); c. Image acquisition: When the laser head is processing, the industrial camera collects images of the processing process in real time, with a frame rate range of 5 to 200 fps; d. Preprocessing: The image online processing unit preprocesses the collected images, including grayscale processing and image filtering and noise reduction, with a processing speed of 5 to 200 ms; e. Feature extraction: Extract the area features of the internal area of the melt pool, the overall image clarity features, and the number of melt pools. The internal area feature of the melt pool is the real-time pixel value s of the internal area of the melt pool, the overall image clarity feature is the average grayscale gradient value h of the image, and the number of melt pools is the number of liquid melt pools within the existing image range q. f. Defect type and stage identification: Based on the calibration ratio k:1 in step a and the processing of the feature image by the online image processing unit in steps c-e, the actual area of the molten pool internal area S=s / k is used to determine the state of the laser melting deposition morphology defect; the types of thin-walled structure morphology defects include end tilting, deposition failure, local collapse, and spheroidization separation. 1) The inclination at both ends is caused by the instability of the molten pool at both ends of the deposition layer: in the initial stage, the area inside the molten pool is small and cannot maintain stability, that is, the average deviation of S within 0.5s at the start or end of each layer is Among them, m is the number of melt pool images. When it is greater than a given threshold value α1, the initial error of the two-end tilt morphology defect begins to appear; in the development stage, the molten pool area of the single layer starting end gradually increases, and the molten pool area of the ending end gradually decreases. The linear function y=ax+b of the fitting position and the molten pool area, when |a| is greater than a given threshold value α2, the two-end tilt morphology defect occurs at the location and is in the development stage; 2) The main reason for the failure of deposition is that the distance l between the laser head and the deposition layer gradually increases: in the initial stage, the average area S of the internal area of each layer, excluding the 0.5s between the starting and ending ends, gradually increases at the rate of area change of each layer. To judge, when θ is greater than a given threshold β1, deposition failure defects begin to appear; in the development stage, the molten pool image gradually moves away from the focus of the industrial camera unit, and the average grayscale gradient value h of each layer image, excluding the 0.5s at the start and end ends, gradually increases. Given the optimal threshold range of h, when h of all images is less than β2, the distance between the laser head and the deposited layer is abnormal, and it has entered the development stage where deposition cannot be achieved; 3) Local collapse is mainly caused by the instability of the molten pool morphology: In the initial stage, the internal area S of the single-layer molten pool fluctuates, and the average deviation of S When it is greater than γ1 and less than γ2, it is the initial stage of the morphological defect; when When it is greater than γ2, it is in the development stage; 4) Spheroidization separation is mainly caused by the collapse of the sintered powder layer on both sides of the aluminum alloy thin-walled component: In the initial stage, The internal area of the molten pool in a single layer will increase significantly, S≥δ; in the development stage, the number of molten pools q is greater than 1; g. Defect in-situ control decision and execution: In-situ control of morphological defects is carried out using an interlayer in-situ method, including: 1) For the tilt defects at both ends, when in the initial stage, they can be controlled by increasing the dwell time at both ends in the next layer; in the development stage, the distance l between the laser head and the deposited layer is determined by the given threshold α3. The distance l is calculated based on the relationship in step b. When l ≥ α3, the processing is terminated and cannot be controlled. When l < α3, the first two dwell times are increased and the self-optimization negative feedback is used for adjustment; 2) For defects that cannot be deposited, when in the initial stage, the lifting amount can be reduced in the next layer, the laser power can be increased, and the scanning speed can be reduced; when in the development stage, it is necessary to adjust l to a reasonable range in the next layer according to the distance l between the laser head and the deposited layer, and then reduce the lifting amount, increase the laser power or reduce the scanning speed in subsequent processing; 3) For local collapse, when it is in the initial stage, it can be solved by reducing the lifting amount in the next layer. When the molten pool stabilizes, continue processing with the original lifting amount. When it is in the development stage, it is necessary to perform only a slight lifting in the next few layers to fill the collapse. When the filling returns to the initial stage, it is executed according to the initial stage control strategy. If it recurs, it is necessary to increase the interlayer interval time or reduce the laser power. 4) For spheroidization separation, when it is in the initial stage, it can be regulated by reducing the laser power in the next layer or increasing the interlayer interval time; when it is in the development stage, the processing has failed and the processing can be terminated.

Images