A powder suction path planning method for multi-metal additive manufacturing process

Abstract

The application discloses a powder suction path planning method for a multi-metal additive manufacturing process, which can accurately obtain a residual powder area by high-resolution imaging, preprocessing and detection and recognition of a forming area, and fill a scanning path according to the residual powder area, so that effective planning of the powder suction path is realized. The method further comprises calculating a moving path and sequence of a powder suction port according to a residual powder area image obtained by the scanning path filling, and controlling the powder suction port to move at corresponding speed, acceleration and direction parameters, so that effective cleaning of the residual powder is realized. The application can effectively solve the quality and efficiency problems in the multi-metal additive manufacturing process caused by the residual powder in the prior art, and can realize accurate control in the multi-metal additive manufacturing process, thereby providing strong technical support for the development of the multi-metal additive manufacturing technology.

Description

TECHNICAL FIELD

[0001] The present application relates to a powder suction path planning method for multi-metal additive manufacturing process, belonging to the field of multi-metal additive manufacturing. BACKGROUND

[0002] Multi-metal additive manufacturing technology (MMAM) is an advanced manufacturing technology that can realize the free combination and distribution of different metal materials in the same part, has the advantages of improving part performance, reducing cost, realizing functional integration, etc., and has wide application prospects in the fields of aerospace, medical treatment, energy, etc. In the multi-metal additive manufacturing process, base structure and different structure are two main structure types. The base structure refers to the structure formed by the base material in the multi-metal material structure, the base material usually has good mechanical properties and thermal stability, and is used to bear most of the load of the part, and the powder laying method is mainly realized by a scraper. The different structure refers to the structure formed by the different material in the multi-metal material structure, the different material usually has good functionality, and is used to meet the specific needs of the part, and the powder laying is mainly completed by a powder feeding device.

[0003] In the existing multi-metal additive manufacturing technology, powder bed fusion technology (PBF) is one of the most widely used technologies. PBF technology uses high-energy density laser beams or electron beams to perform selective melting or selective sintering on the powder bed, and accumulates the required parts layer by layer. Multi-metal PBF technology can realize the gradient transition or interface connection between different metal materials, thereby obtaining parts with composite performance or function. However, multi-metal PBF technology also faces many challenges and difficulties, one of which is powder management in the multi-metal powder bed additive manufacturing process, which involves powder transportation, laying, suction, recycling and other links, directly affecting the forming quality, efficiency and safety. At present, the multi-metal powder bed additive manufacturing process mainly adopts the method of multiple powder laying and multiple scanning in the layer, that is, the base material powder is laid by a scraper first, the base structure is formed by using laser selective melting technology (SLM), then the different material powder is laid by a powder feeding device, and then the different material powder is laid flat by a scraper, and then the different material powder is formed by laser selective melting. This method needs to suction and recycle the residual powder after each layer to facilitate the powder laying and scanning of the next layer.

[0004] The above existing PBF additive manufacturing process has obvious shortcomings, for example: (1) multiple switching of laser parameters and scanning strategies is required, which not only increases the process complexity, but also significantly lengthens the manufacturing time; (2) multiple uses of scrapers for powder leveling may cause mixing and contamination of the powder, affecting the forming quality; (3) residual powder needs to be sucked and recycled after each layer, which requires the use of a special powder suction device and recycling system, increasing equipment cost and maintenance difficulty. Therefore, when sucking residual powder, a reasonable powder suction path needs to be planned to avoid damage to the formed structure and interference with the unformed area, while minimizing the moving distance and time of the suction port. Therefore, how to design a powder suction path planning method for a multi-metal additive manufacturing process to improve the powder suction efficiency, reduce the moving path of the suction port during the powder suction process, and quickly achieve the suction and cleaning of residual powder to avoid contamination of the base powder is a technical problem that needs to be solved. SUMMARY

[0005] (I) Invention purposes

[0006] In view of the above defects and deficiencies of the prior art, the present application aims to provide a powder suction path planning method for a multi-metal additive manufacturing process, which accurately obtains the area with residual powder by high-resolution imaging, preprocessing and detection and identification of the forming area in the multi-metal additive manufacturing process, and then fills the scanning path according to the residual powder area, thereby effectively planning the powder suction path in the multi-metal additive manufacturing process, improving the multi-metal additive manufacturing efficiency and quality, and providing strong technical support for the development of multi-metal additive manufacturing technology. At the same time, by introducing Z-shaped filling and reasonable filling vector direction selection, the efficiency and stability of the powder suction process are ensured, providing an effective method for precise control in the multi-metal additive manufacturing process.

[0007] (II) Technical solutions

[0008] To achieve the purpose of the application, the application adopts the following technical solutions:

[0009] A powder suction path planning method for a multi-metal additive manufacturing process, characterized in that the powder suction path planning method comprises at least the following steps when implemented:

[0010] SS1. In a multi-metal additive manufacturing process based on a laser powder bed, after laser selective melting of different materials, a high-resolution camera is used to image the forming area to obtain image data I=F(C img ,R frm ) of the forming area, wherein I represents the obtained image data of the forming area, R frm is the forming area, and C imgwherein, F is a function of imaging parameters and at least includes resolution, exposure time, aperture size and focal length of the camera, I is the image data of the formed area, and N is the number of slice layers of the formed area, C img_i (R frm_i ) is the image data of the i-th slice layer of the formed area R frm_i .

[0011] SS2. Pre-process the image data I of the formed area obtained in step SS1 using the pre-known slice contour information S slc to segment the base material area B, dissimilar material area D and all the support fused areas S, and mark each area with different color or gray value for distinction, wherein, I = B∪D∪S, (B, D, S) = P(I, S slc ), P is a pre-processing function of the image data of the formed area and satisfies:

[0012]

[0013] wherein, S slc_B , S slc_D and S slc_S represent the slice contour information of the base material, dissimilar material and support structure, respectively.

[0014] SS3. Detect and identify the fused area image S segmented in step SS2 to obtain the area R with residual powder, and mark it with a specific color or gray value for distinction from other areas, wherein, R = G(S, T dct ), T dct is a detection and identification threshold value determined according to the reflectivity, scattering and / or absorption optical properties of the residual powder, and G is a detection and identification function and satisfies:

[0015]

[0016] SS4. Scan the residual powder area R with a scan path P s using zigzag filling, the filling vector direction is perpendicular or parallel to the doctor blade powder laying direction, and the filling vector width is less than the powder suction width, wherein, and satisfies P s = U(R, D Z , D vec , D width ), DZ is a zigzag filling parameter, D vec is a filling vector direction parameter, and D width is a filling vector width parameter, and U is a filling function and satisfies:

[0017]

[0018] In the formula, P s To fill according to the zigzag parameter D Z Fill vector direction parameter D vec and fill vector width parameter D width The scanning path obtained after filling the residual powder region R. An empty set is defined as the region R containing the residual powder being an empty set. At that time, scan path P s It is also an empty set to avoid ineffective powder suction operations when there is no residual powder;

[0019] SS5. According to the scan path P s The image of the residual powder area is filled, the moving path M and sequence of the powder suction port are calculated, and the powder suction port is controlled to move according to the corresponding speed, acceleration and direction parameters to achieve effective cleaning of residual powder. And M = W(P) s V spd A acc D dir V spd For the velocity parameter, A acc For acceleration parameters, D dir Let W be the direction parameter, and let W be the function that calculates the movement path and sequence of the powder suction port and satisfies the following:

[0020]

[0021] In the formula, M represents the velocity parameter V. spd Acceleration parameter A acc and direction parameter D dir For scan path P s The optimized movement path of the powder suction port. If the set is empty, when scanning path P s When the scan path is empty, the movement path M of the suction port is also empty to avoid invalid suction operations without a scan path.

[0022] Preferably, in step SS1 above, the imaging function F further includes at least a denoising operation on the image data I to eliminate the effects of laser, thermal radiation, and / or dust interference, wherein the denoising operation is based on the following formula:

[0023] I denoised =I-μ(σ)·N(I)

[0024] This formula eliminates random noise in an image by subtracting the product of the image data I and the denoising factor of its noise model, where σ is the noise standard deviation, N is the noise model function, and μ is the denoising factor adjusted according to the noise standard deviation σ.

[0025] N(I) = φ(I) + ω,

[0026] wherein φ is a Gaussian white noise function, ω is a uniform distribution noise function, and in the formula of the de-noising factor μ, the de-noising factor μ is made to decay rapidly with the increase of the noise standard deviation σ by equating the de-noising factor μ to the negative exponential of the square of the noise standard deviation σ, so as to enhance the de-noising effect.

[0027] Preferably, in the step SS2, the pre-processing function P further comprises at least an edge detection operation and a contour matching operation on the image data I, so as to improve the correspondence between the slice contour information S slc and the image data I and the segmentation accuracy, wherein the edge detection operation and the contour matching operation are respectively performed by the following formulas:

[0028] E = ED(I, κ), (B, D, S) = CM(E, S slc , λ)

[0029] wherein ED is an edge detection function, E is the edge information of the image data I, κ is an edge detection parameter, CM is a contour matching function, and λ is a contour matching parameter.

[0030] Preferably, in the step SS2, the pre-processing function P further comprises at least a region growing operation and a region merging operation on the image data I, so as to improve the coverage and segmentation continuity between the slice contour information S slc and the image data I, wherein the region growing operation and the region merging operation are respectively performed by the following formulas:

[0031] R = RG(I, θ), (B, D, S) = RM(R', S slc , δ)

[0032] wherein RG is a region growing function, R' is the region information of the image data I, θ is a region growing threshold, RM is a region merging function, and δ is a region merging threshold.

[0033] Preferably, in the step SS2, the pre-processing function P further comprises at least a superpixel segmentation operation and a superpixel labeling operation on the image data I, so as to improve the similarity and segmentation efficiency between the slice contour information S slc and the image data I, wherein the superpixel segmentation operation and the superpixel labeling operation are respectively performed by the following formulas:

[0034] SP = SPS(I, α), (B, D, S) = SPL(SP, S slc , β)

[0035] Wherein, SPS is a superpixel segmentation function, SP is superpixel information of image data I, a is a superpixel segmentation parameter, SPL is a superpixel labeling function, and β is a superpixel labeling parameter.

[0036] Preferably, in the step SS3, the detection identifies the threshold T dct determined according to the optical properties of reflectivity, scattering and / or absorption of the residual powder, wherein the optical properties are obtained by spectral analysis and spectral classification operations on the solidified region image S, which are performed by the following formulas:

[0037] SP = SA(S, ω), R = SC(SP, T spt )

[0038] Wherein, SA is a spectral analysis function, SP is spectral information of the solidified region image S, ω is a spectral analysis parameter, SC is a spectral classification function, T spt is a spectral classification threshold, and the detection identification threshold Tdct is based on the following formula:

[0039] T dct = TC(SP, τ)

[0040] Wherein, TC is a detection identification threshold calculation function, and τ is a detection identification threshold calculation parameter.

[0041] Preferably, in the step SS4, the filling function U further includes at least an optimization, smoothing and / or constraint operation on the scan path P s Preferably, in the step SS4, the filling function U further includes at least an optimization, smoothing and / or constraint operation on the scan path P

[0042] P s,opt = O(Ps, β)

[0043] Wherein, O is a scan path optimization function, β is a scan path optimization parameter and includes at least path length, direction and / or coverage factors.

[0044] In this process, path optimization is to reduce the length and / or intersection of the scan path while ensuring coverage efficiency, in order to reduce time and resource consumption in the manufacturing process. Path smoothing is to eliminate or reduce sharp corners and discontinuities in the scan path, so as to reduce vibration and deceleration when the machine moves, and improve manufacturing efficiency and quality. The smoothing operation can be achieved by various mathematical methods, such as Bezier curve, B-spline curve, etc. Path constraint is to ensure that the scan path meets specific process and mechanical constraint conditions, such as not exceeding the maximum speed and acceleration of mechanical movement, avoiding collision, etc.

[0045] Preferably, in the step SS4, the zigzag filling is performed by optimizing the scan path Ps The Z-shaped transformation and Z-shaped optimization operations are implemented, wherein the Z-shaped transformation operation and the Z-shaped optimization operation are based on the following formulas:

[0046] Ps Z = ZT(Ps, a), Ps ZO = ZO(Ps Z , b)

[0047] wherein ZT is a Z-shaped transformation function, Ps Z is a transformed scanning path, a is a Z-shaped transformation parameter, ZO is a Z-shaped optimization function, Ps ZO is an optimized scanning path, b is an optimization parameter. The scanning path is converted into a series of Z-shaped curves by using the Z-shaped transformation function, and the curves are adjusted and connected by using the Z-shaped optimization function, so as to improve the continuity and uniformity of the scanning path.

[0048] Preferably, in the above step SS5, the movement process of the powder suction port is optimized by introducing a multi-stage speed and acceleration control strategy, and the movement speed V spd and the acceleration A acc of the powder suction port are dynamically adjusted according to the distribution and density of the residual powder region R, so as to achieve more efficient and accurate powder cleaning, and the dynamic speed and acceleration parameters of the powder suction port are determined by using the following formulas:

[0049] Vspd = V(R, a), A acc = A(R, b)

[0050] wherein V is a speed control function, A is an acceleration control function, R is residual powder region information, a is a speed control parameter, and b is an acceleration control parameter. The movement process of the powder suction port is controlled in multiple stages by using the speed control function and the acceleration control function, and is dynamically adjusted according to the distribution and density of the residual powder region, so as to improve the movement efficiency and accuracy of the powder suction port.

[0051] Preferably, in the above step SS5, an adaptive control strategy is adopted, and the movement speed V spd , the acceleration A acc and the direction parameter D dir of the powder suction port are dynamically adjusted according to the real-time monitored residual powder condition, so as to achieve more efficient and accurate residual powder cleaning:

[0052] V spd_adj ,A acc_adj ,D dir_adj = AC(V spd ,A acc ,D dir ,F fb )

[0053] wherein AC represents an adaptive control function, F fb represents real-time feedback information.

[0054] (Three) Technical effects

[0055] Compared with the prior art, the powder suction path planning method for the multi-metal additive manufacturing process has the following beneficial and significant technical effects:

[0056] (1) The present application accurately obtains the area with residual powder by high-resolution imaging, pre-processing and detection and recognition of the forming area, and fills the scanning path according to the residual powder area, thereby effectively planning the powder suction path in the multi-metal additive manufacturing process and improving the efficiency and quality of multi-metal additive manufacturing.

[0057] (2) The present application calculates the moving path and sequence of the powder suction port according to the residual powder area image obtained by scanning path filling, and controls the powder suction port to move according to the corresponding speed, acceleration and direction parameters, thereby effectively planning and adjusting the moving path of the powder suction port and effectively cleaning the residual powder, avoiding the pollution of the residual powder to the base powder.

[0058] (3) The present application can effectively avoid invalid powder suction operation in the absence of residual powder or scanning path, thereby saving time and resources and improving the efficiency and quality of the multi-metal additive manufacturing process.

[0059] (4) The present application can effectively solve the quality and efficiency problems in the multi-metal additive manufacturing process caused by the presence of residual powder in the prior art, such as the influence of residual powder on the forming quality of the next layer, the generation of thermal stress and thermal cracks, and the increase of subsequent cleaning and polishing workload. BRIEF DESCRIPTION OF DRAWINGS

[0060] Figure 1 The figure shows the implementation flowchart of the powder suction path planning method for the multi-metal additive manufacturing process of the present application.

[0061] Figure 2 The figure shows the schematic diagram of each region obtained by segmenting the pre-processed forming area image.

[0062] Figure 3 The figure shows the schematic diagram of the area with residual powder obtained after detection and recognition.

[0063] Figure 4 The figure shows a schematic diagram of the transverse path of the powder suction scanning path.

[0064] Figure 5 The figure shows a schematic diagram of the longitudinal path of the powder suction scanning path.

[0065] BRIEF DESCRIPTION OF DRAWINGS

[0066] 1 - base material region, 2 - dissimilar material region, 3 - outer side supported solidified region, 4 - inner side supported solidified region, 5 - residual powder region. DETAILED DESCRIPTION

[0067] For better understanding of the present application, the following further illustrates the content of the present application in combination with examples. In the drawings, the same or similar notations represent the same or similar elements or elements with the same or similar functions throughout. The described examples are part of the embodiments of the present application, but not all the embodiments. The examples described below by referring to the drawings are exemplary and are intended to explain the present application, but cannot be understood as a limitation of the present application.

[0068] Example 1

[0069] As shown in the following, the powder suction path planning method for multi-metal additive manufacturing process of the present application at least includes the following steps in implementation: Figure 1

[0070] SS1. In a laser powder bed based multi-metal additive manufacturing process, after the dissimilar material laser selective melting forming, the forming region is imaged by using a high resolution camera to obtain the image data I = F(C img ,R frm ) of the forming region, wherein I represents the obtained image data of the forming region, R frm is the forming region, C img is the high resolution camera parameter and at least includes the resolution, exposure time, aperture size and focal length of the camera, F is the imaging function and N is the number of slice layers of the forming region, C img_i (R frm_i ) is the image data of the i-th layer slice region R frm_i ;

[0071] SS2. The forming region image data I obtained in step SS1 is pre-processed, and the previously known slice contour information S slc is used to segment out the base material region B1, dissimilar material region D2 and all supported solidified regions S3, 4 (as shown in the following), and each region is marked with different colors or gray values for differentiation, wherein I = B U D U S, (B, D, S) = P(I, S slc ), P is the pre-processing function of the forming region image data and satisfies: Figure 2

[0072]

[0073] In the formula, S​​slc_B S slc_D and S slc_S These represent the slice outline information of the matrix material, dissimilar material, and supporting structure, respectively.

[0074] SS3. Detect and identify the fused region image S obtained from step SS2 to obtain the region R5 with residual powder (e.g., Figure 3 As shown), it is marked with a specific color or grayscale value to distinguish it from other areas, where R = G(S,T) dct ), T dct Let G be the detection and identification threshold determined based on the optical properties of the residual powder, including its reflectivity, scattering, and / or absorptivity, and let G be the detection and identification function satisfying:

[0075]

[0076] SS4. Scan the residual powder region R along path P. s Fill using a zigzag pattern, with the fill vector direction parallel (e.g., ...). Figure 4 (as shown) or perpendicular to the direction of powder spreading by the scraper (e.g.) Figure 5 As shown), the fill vector width is smaller than the powder absorption width, where, And satisfy P s =U(R,D Z D vec D width ), D Z It is the Z-shaped fill parameter, D vec To fill the vector direction parameter, D width The fill vector width parameter is U, which is the fill function and satisfies:

[0077]

[0078] In the formula, P s To fill according to the zigzag parameter D Z Fill vector direction parameter D vec and fill vector width parameter D width The scanning path obtained after filling the residual powder region R. An empty set is defined as the region R containing the residual powder being an empty set. At that time, scan path P s It is also an empty set to avoid ineffective powder suction operations when there is no residual powder;

[0079] SS5. According to the scan path P s The image of the residual powder area is filled, the moving path M and sequence of the powder suction port are calculated, and the powder suction port is controlled to move according to the corresponding speed, acceleration and direction parameters to achieve effective cleaning of residual powder. and M = W(P s , V spd , A acc , D dir ), V spd is a velocity parameter, A acc is an acceleration parameter, D dir is a direction parameter, and W is a function for calculating the moving path and sequence of the suction port and satisfies:

[0080]

[0081] wherein M is a moving path of the suction port after optimizing the scanning path P spd according to the velocity parameter V acc , the acceleration parameter A dir , and the direction parameter D s , and is an empty set, and when the scanning path P s is an empty set, the moving path M of the suction port is also an empty set, so as to avoid invalid suction operation in the absence of a scanning path.

[0082] Embodiment 2

[0083] This embodiment is a further optimized technical solution based on Embodiment 1, for example:

[0084] In a preferred example of the present application, in the above step SS1, the imaging function F further includes at least a denoising operation on the image data I to eliminate the influence of laser, heat radiation and / or dust interference, wherein the denoising operation is based on the following formula:

[0085] I denoised = I - μ(σ) · N(I)

[0086] This formula eliminates random noise in the image by subtracting the product of the image data I and the denoising factor of its noise model, wherein σ is the noise standard deviation, N is the noise model function, μ is the denoising factor adjusted according to the noise standard deviation σ, and wherein

[0087] N(I) = φ(I) + ω,

[0088] wherein φ is a Gaussian white noise function, ω is a uniformly distributed noise function, and in the formula of the denoising factor μ, the denoising factor μ is made to decay rapidly with the increase of the noise standard deviation σ by equating the denoising factor μ to the negative exponential of the square of the noise standard deviation σ, thereby enhancing the denoising effect.

[0089] In the preferred embodiment of the present application, in the step SS2, the pre-processing function P further comprises at least an edge detection operation and a contour matching operation on the image data I, so as to improve the slice contour information S slc and the segmentation accuracy with the image data I, wherein the edge detection operation and the contour matching operation are respectively performed by the following formulas:

[0090] E = ED(I, K), (B, D, S) = CM(E, S slc , λ)

[0091] wherein ED is an edge detection function, E is the edge information of the image data I, K is an edge detection parameter, CM is a contour matching function, and λ is a contour matching parameter.

[0092] In the preferred embodiment of the present application, in the step SS2, the pre-processing function P further comprises at least a region growing operation and a region merging operation on the image data I, so as to improve the coverage and the segmentation continuity of the slice contour information S slc with the image data I, wherein the region growing operation and the region merging operation are respectively performed by the following formulas:

[0093] R = RG(I, θ), (B, D, S) = RM(R', S slc , δ)

[0094] wherein RG is a region growing function, R' is the region information of the image data I, θ is a region growing threshold, RM is a region merging function, and δ is a region merging threshold.

[0095] In the preferred embodiment of the present application, in the step SS2, the pre-processing function P further comprises at least a superpixel segmentation operation and a superpixel labeling operation on the image data I, so as to improve the similarity and the segmentation efficiency of the slice contour information S slc with the image data I, wherein the superpixel segmentation operation and the superpixel labeling operation are respectively performed by the following formulas:

[0096] SP = SPS(I, α), (B, D, S) = SPL(SP, S slc , β)

[0097] wherein SPS is a superpixel segmentation function, SP is the superpixel information of the image data I, α is a superpixel segmentation parameter, SPL is a superpixel labeling function, and β is a superpixel labeling parameter.

[0098] In the preferred embodiment of the present application, in the step SS3, the detection and recognition threshold T dctdetermined according to optical properties of reflectivity, scattering and / or absorption of the residual powder, wherein the optical properties are obtained by spectral analysis and spectral classification operations on the solidified region image S, the spectral analysis and spectral classification operations are performed by following equations:

[0099] SP = SA(S, ω), R = SC(SP, T spt )

[0100] wherein SA is a spectral analysis function, SP is spectral information of the solidified region image S, ω is a spectral analysis parameter, SC is a spectral classification function, T spt is a spectral classification threshold, and the detection recognition threshold Tdct is determined according to following equation:

[0101] T dct = TC(SP, τ)

[0102] wherein TC is a detection recognition threshold calculation function, and τ is a detection recognition threshold calculation parameter.

[0103] In preferred embodiments of the present application, in the step SS4, the filling function U further comprises at least an optimization, smoothing and / or constraint operation on the scan path P s , wherein the optimization operation is performed according to following equation:

[0104] P s,opt = O(Ps, β)

[0105] wherein O is a scan path optimization function, and β is a scan path optimization parameter and comprises at least path length, direction and / or coverage factors.

[0106] In this process, the path optimization is to reduce the length and / or intersection of the scan path while ensuring the coverage efficiency, so as to reduce the time and resource consumption in the manufacturing process. The path smoothing is to eliminate or reduce the sharp corners and discontinuous points in the scan path, so as to reduce the vibration and deceleration when the machine moves, and improve the manufacturing efficiency and quality. The smoothing operation can be realized by various mathematical methods, such as Bezier curve, B-spline curve, etc. The path constraint is to ensure that the scan path meets specific process and mechanical constraint conditions, such as not exceeding the maximum speed and acceleration of mechanical movement, avoiding collision, etc.

[0107] In preferred embodiments of the present application, in the step SS4, the zigzag filling is realized by zigzag transformation and zigzag optimization operations on the scan path P s , wherein the zigzag transformation operation and the zigzag optimization operation are performed according to following equations:

[0108] Ps Z= ZT(Ps, a), Ps ZO = ZO(Ps Z , b)

[0109] wherein ZT is a zigzag transformation function, Ps Z is the transformed scan path, a is a zigzag transformation parameter, ZO is a zigzag optimization function, Ps ZO is the optimized scan path, b is an optimization parameter. The zigzag transformation function is used to convert the scan path into a series of zigzag curves, and the zigzag optimization function is used to adjust and connect the curves to improve the continuity and uniformity of the scan path.

[0110] In the preferred embodiment of the present application, in the above step SS5, a multi-level speed and acceleration control strategy is introduced to optimize the movement process of the powder suction port. The movement speed V spd and acceleration A acc of the powder suction port are dynamically adjusted according to the distribution and density of the residual powder region R to achieve more efficient and accurate powder cleaning. The dynamic speed and acceleration parameters of the powder suction port are determined by the following formula:

[0111] V spd = V(R, a), A acc = A(R, b)

[0112] wherein V is a speed control function, A is an acceleration control function, R is residual powder region information, a is a speed control parameter, and b is an acceleration control parameter. The movement process of the powder suction port is controlled by the speed control function and the acceleration control function, and is dynamically adjusted according to the distribution and density of the residual powder region, thereby improving the movement efficiency and accuracy of the powder suction port.

[0113] In the preferred embodiment of the present application, in the above step SS5, an adaptive control strategy is adopted. According to the real-time monitoring of the residual powder situation, the movement speed V spd , acceleration A acc and direction parameter D dir of the powder suction port are dynamically adjusted to achieve more efficient and accurate residual powder cleaning:

[0114] V spd_adj , A acc_adj , D dir_adj = AC(V spd , A acc , D dir , F fb )

[0115] wherein AC represents an adaptive control function, and F fb represents real-time feedback information.

[0116] The objects of the present application are completely achieved by the above embodiments. Those skilled in the art can understand that the present application includes but is not limited to the content described in the drawings and the above specific embodiments. Although the present application has been described with respect to the presently preferred and most practical embodiments, it will be understood that the present application is not limited to the disclosed embodiments, and any modification not deviating from the functional and structural principles of the present application will be included in the scope of the claims.

Claims

1. A powder suction path planning method for multi-metal additive manufacturing processes, characterized in that, The fan acquisition path planning method includes at least the following steps when implemented: SS1. In the multi-metal additive manufacturing process based on laser powder bed, after the dissimilar materials are selectively melted and shaped by laser, a high-resolution camera is used to image the shaped area, obtaining image data of the shaped area I = F(C img ,R frm ), where I represents the image data of the obtained shaped region, R frm For the forming area, C img For high-resolution camera parameters, include at least the camera's resolution, exposure time, aperture size, and focal length, where F is the imaging function and N is the number of slice layers in the forming region, C img_i (R frm_i R is the slice region of the i-th layer. frm_i Image data; SS2. Preprocess the shaped region image data I obtained in step SS1, using the pre-known slice contour information S slc The matrix material region B, the dissimilar material region D, and the fused regions S of all supports are segmented, and each region is marked with a different color or grayscale value for distinction. Where I = B∪D∪S, (B,D,S) = P(I,S) slc P is the preprocessing function for the image data of the shaped region and satisfies: In the formula, S slc_B S slc_D and S slc_S These represent the slice outline information of the matrix material, dissimilar material, and supporting structure, respectively. SS3. Detect and identify the fused region image S obtained from step SS2 to obtain the region R with residual powder, and mark it with a specific color or grayscale value to distinguish it from other regions, where R = G(S,T) dct ), T dct Let G be the detection and identification threshold determined based on the optical properties of the residual powder, including its reflectivity, scattering, and / or absorptivity, and let G be the detection and identification function satisfying: SS4. Scan the residual powder region R along path P. s During filling, a zigzag filling pattern is used. The filling vector direction is perpendicular or parallel to the powder spreading direction of the scraper, and the filling vector width is smaller than the powder absorption width. And satisfy P s =U(R,D Z D vec D width ), D Z It is the Z-shaped fill parameter, D vec To fill the vector direction parameter, D width The fill vector width parameter is U, which is the fill function and satisfies: In the formula, P s To fill according to the zigzag parameter D Z Fill vector direction parameter D vec and fill vector width parameter D width The scanning path obtained after filling the residual powder region R. The residual powder region R is an empty set. At that time, scan path P s It is also an empty set to avoid ineffective powder suction operations when there is no residual powder; SS5. According to the scan path P s The image of the residual powder area is filled, the moving path M and sequence of the powder suction port are calculated, and the powder suction port is controlled to move according to the corresponding speed, acceleration and direction parameters to achieve effective cleaning of residual powder. And M = W(P) s V spd A acc D dir V spd For the velocity parameter, A acc For acceleration parameters, D dir Let W be the direction parameter, and let W be the function that calculates the movement path and sequence of the powder suction port and satisfies the following: In the formula, M represents the velocity parameter V. spd Acceleration parameter A acc and direction parameter D dir For scan path P s The optimized movement path of the powder suction port. If the set is empty, when scanning path P s When the scan path is empty, the movement path M of the suction port is also empty to avoid invalid suction operations without a scan path.

2. The powder suction path planning method for multi-metal additive manufacturing process according to claim 1, characterized in that, In step SS1 above, the imaging function F further includes at least a denoising operation on the image data I to eliminate the effects of laser, thermal radiation, and / or dust interference, wherein the denoising operation is based on the following formula: I denoised =I-μ(σ)·N(I) This formula eliminates random noise in an image by subtracting the product of the image data I and the denoising factor of its noise model, where σ is the noise standard deviation, N is the noise model function, and μ is the denoising factor adjusted according to the noise standard deviation σ. N(I)=φ(I)+ω, In the formula, φ is the Gaussian white noise function, ω is the uniformly distributed noise function, and the denoising factor μ is calculated by making the negative exponent of the square of the noise standard deviation σ equal to the denoising factor μ, so that the denoising factor μ decreases rapidly as the noise standard deviation σ increases.

3. The powder suction path planning method for multi-metal additive manufacturing process according to claim 1, characterized in that, In step SS2 above, the preprocessing function P further includes at least edge detection and contour matching operations on the image data I to improve the slice contour information S. slc The correspondence and segmentation accuracy between the image data I and the edge detection operation and the contour matching operation are respectively performed by the following formulas: E=ED(I,κ),(B,D,S)=CM(E,S slc ,λ) In the formula, ED is the edge detection function, E is the edge information of image data I, κ is the edge detection parameter, CM is the contour matching function, and λ is the contour matching parameter.

4. The powder suction path planning method for multi-metal additive manufacturing process according to claim 1, characterized in that, In step SS2 above, the preprocessing function P further includes at least region growing and region merging operations on the image data I, wherein the region growing and region merging operations are performed using the following formulas: R=RG(I,θ),(B,D,S)=RM(R′,S slc ,δ) Where RG is the region growing function, R′ is the region information of image data I, θ is the region growing threshold, RM is the region merging function, and δ is the region merging threshold.

5. The powder suction path planning method for multi-metal additive manufacturing process according to claim 1, characterized in that, In step SS2 above, the preprocessing function P further includes at least superpixel segmentation and superpixel labeling operations on the image data I to improve the slice contour information S. slc The similarity and segmentation efficiency between the superpixel segmentation operation and the superpixel labeling operation are calculated using the following formulas: SP=SPS(I,α),(B,D,S)=SPL(SP,S slc ,β) Where SPS is the superpixel segmentation function, SP is the superpixel information of image data I, α is the superpixel segmentation parameter, SPL is the superpixel labeling function, and β is the superpixel labeling parameter.

6. The powder suction path planning method for multi-metal additive manufacturing process according to claim 1, characterized in that, In step SS3 above, the detection and identification threshold T dct The optical properties of the residual powder are determined based on its reflectance, scattering, and / or absorptivity, wherein these optical properties are obtained through spectral analysis and classification of the image S of the fused region, and the spectral analysis and classification are performed using the following formula: SP=SA(S,ω), R=SC(SP,Tspt) Where SA is the spectral analysis function, SP is the spectral information of the melting and solidification region image S, ω is the spectral analysis parameter, SC is the spectral classification function, and T... spt The spectral classification threshold is Tdct, and the detection and recognition threshold Tdct is based on the following formula: T dct =TC(SP,τ) Where TC is the detection and recognition threshold calculation function, and τ is the detection and recognition threshold calculation parameter.

7. The powder suction path planning method for multi-metal additive manufacturing process according to claim 1, characterized in that, In step SS4 above, the fill function U further includes at least the scan path P. s Perform optimization, smoothing, and / or constraint operations, wherein the optimization operations are based on P. s,opt =O(Ps,β) is used for optimization, where O is the scan path optimization function and β is the scan path optimization parameter, which includes at least path length, direction and / or coverage factors.

8. The powder suction path planning method for multi-metal additive manufacturing process according to claim 1, characterized in that, In step SS4 above, the zigzag filling is achieved by adjusting the scan path P. s This is achieved by performing Z-shaped transformation and Z-shaped optimization operations, wherein the Z-shaped transformation and Z-shaped optimization operations are based on the following formula: Ps Z =ZT(Ps,α),=Ps ZO =ZO(Ps Z ,β) Where ZT is the Z-shaped transformation function, and Ps Z The transformed scan path is α, the zigzag transformation parameter is ZO, the zigzag optimization function is ZO, and Ps is Ps. ZO β is the optimization parameter for the optimized scan path.

9. The powder suction path planning method for multi-metal additive manufacturing process according to claim 1, characterized in that, In step SS5 above, a multi-stage speed and acceleration control strategy is introduced to optimize the movement of the powder suction port. The moving speed V of the powder suction port is dynamically adjusted according to the distribution and density of the residual powder region R. spd and acceleration A acc To achieve more efficient and accurate powder cleaning, the dynamic velocity and acceleration parameters of the powder suction port are determined using the following formula: V spd =V(R,α),A acc =A(R,β) Where V is the velocity control function, A is the acceleration control function, R is the residual powder region information, α is the velocity control parameter, and β is the acceleration control parameter.

10. The powder suction path planning method for multi-metal additive manufacturing process according to claim 1, characterized in that, In step SS5 above, an adaptive control strategy is adopted to dynamically adjust the moving speed V of the powder suction port based on the real-time monitoring of the residual powder situation. spd Acceleration A acc and directional parameter D dir To achieve more efficient and accurate removal of residual powder: V spd_adj ,A acc_adj ,D dir_adj =AC(V spd ,A acc ,D dir , F fb ) Where AC represents the adaptive control function, F fb This indicates real-time feedback information.

Images