Random polygon partition forming method applied to additive manufacturing

Abstract

The invention discloses a random polygon partition forming method applied to additive manufacturing, and the method comprises the steps: obtaining an envelope boundary of a geometric contour of a slice layer, and carrying out the outward offset to generate a grid coverage area; based on the position relationship between the geometric contour and each grid area in the grid coverage area, giving marks to the corresponding grid areas, the marks including excursion, intersection and inclusion; generating a seed point coordinate in each marked grid region; constructing a Thiessen polygon based on the coordinates of all seed points, and reserving the Thiessen polygon inside the geometric contour of the slice layer by using the geometric contour of the slice layer; after the reserved Thiessen polygon is biased inwards, all partitions in the geometric contour range of the application slice layer are obtained; the problem of layer-to-layer space overlapping at the partition boundary position in a traditional partition method is avoided, the situation that small-area Thiessen polygons appear on the edge part after cutting is reduced, and the uniformity of microstructure distribution of the part is improved in the laser forming process.

Description

Technical Field

[0001] The present invention relates to the technical field of additive manufacturing, and in particular to a random polygon partition forming method applied to additive manufacturing. Background Art

[0002] Laser powder bed fusion technology is a mainstream method in the field of additive manufacturing. The workflow of this technology first involves 3D modeling, then the model is sliced ​​to obtain layer data, and then the laser filling trajectory is generated according to the geometry of the sliced ​​layer. Finally, the slice and trajectory files are encoded into a format that the device can recognize for printing. In the manufacturing process, the laser scanning strategy adopted has a significant impact on the mechanical properties of the formed parts. This is mainly attributed to the high solidification rate of the molten pool formed during the laser scanning process, which leads to the accumulation of a large amount of thermal stress, which will increase with the increase of the number of scanned layers. If the laser scanning strategy is not suitable, printing may fail in the face of specific materials or precision requirements.

[0003] Commonly used laser scanning strategies include single-track scanning, zigzag scanning, checkerboard scanning, etc. The first two can be summarized as using a series of parallel scanning paths to fill the area to be printed. Although this scanning strategy is relatively easy to implement, as the vector length of the laser scanning line increases, the thermal stress generated after the molten pool solidifies is also greater, which is prone to printing defects and reduces the forming accuracy; checkerboard scanning is another commonly used scanning strategy, which shortens the laser vector length by dividing the sub-areas, thereby reducing the residual stress after the molten pool solidifies. However, the forming quality of checkerboard scanning often decreases at the sub-partition boundaries. On the one hand, this is because the molten pool in a single grid turns or stops at the boundary, and the laser energy density received by the molten pool near the boundary is different from that of the molten pool in the middle area. If the molten pool stops at the boundary, the front end of the molten pool will accumulate and solidify at the boundary position, eventually causing a difference between the microstructure at the boundary and the microstructure in the middle part. On the other hand, in the process of checkerboard division in the existing algorithm, the checkerboard area overlaps in the spatial position between layers, which undoubtedly aggravates the problem of uneven organization of the existing solidified layer at the boundary position. Therefore, a random polygonal partition forming method for additive manufacturing is urgently needed to solve the above problems. Summary of the invention

[0004] The present invention aims to provide a random polygonal partition forming method for additive manufacturing, which can effectively solve the problems existing in the above-mentioned prior art.

[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions: a random polygon partition forming method applied to additive manufacturing, by partitioning each slice layer and repeating the slice layer partitioning step, a partitioned laser scanning trajectory applied to the model is obtained, and the slice layer partitioning includes the following steps:

[0006] Obtain the envelope boundary of the slice layer geometry contour and offset it outward to generate a grid coverage area;

[0007] Based on the positional relationship between the geometric outline and each grid area in the grid coverage area, the corresponding grid area is marked, including outlying, intersecting and containing;

[0008] Generate seed point coordinates in all outer and inner grid regions;

[0009] In the intersecting grid area, obtaining the corresponding influence weight based on the seed point coordinates of the adjacent grid area; and

[0010] Based on each influence weight and constructing an error function, the seed point coordinates of the intersecting grid area are obtained;

[0011] Construct Thiessen polygons based on all seed point coordinates and use the slice layer geometry outline to preserve the Thiessen polygons inside it;

[0012] The retained Thiessen polygons are biased inward to obtain all partitions within the geometric outline of the applied slicing layer.

[0013] Preferably, the offset distance is d, where d=2a, and a is the grid unit size.

[0014] Preferably, each grid area marking constraint is: when the grid area g ij When included, mark m ij =1, when the grid area g ij When they intersect and the intersection area is greater than the design threshold Arealimit, mark m ij =0; when the grid area g ij When the outer separation or intersection area is less than or equal to the design threshold Arealimit, mark m ij =-1.

[0015] Preferably, the coordinates of the seed points generated in the outer and inner grid areas are:

[0016] p m＝1,-1 (x,y)=(xmin+rand·(xmax-xmin),ymin+rand·(ymax-ymin));

[0017] Among them, xmin, xmax, ymin, and ymax represent the grid area g respectively. ij boundary range.

[0018] Preferably, a seed point coordinate is randomly generated for each grid area.

[0019] Preferably, the influence weight is obtained as follows:

[0020] Calculate the centroid coordinates of the intersection area of ​​the intersecting grid area and the geometric contour [C x ,C y ]:

[0021]

[0022] Among them, (x i ,y i ) are the vertex coordinates of the intersection area, and A is the area of ​​the intersection area;

[0023] Calculate the Euclidean distance between the seed coordinates in all adjacent grid areas around the grid area and the centroid coordinates;

[0024] The corresponding influence weight is calculated based on the Euclidean geometric distance:

[0025]

[0026] Among them, L ij is the Euclidean geometric distance, σ represents the influencing parameter value, and w ij ′ is the normalized weight coefficient.

[0027] Preferably, a point coordinate is randomly generated in a grid area adjacent to the intersecting grid areas and currently without a seed coordinate as a temporary seed point coordinate for calculation, and the temporary seed point coordinate is reset after the calculation is completed.

[0028] Preferably, an error function is constructed and the minimum error function is solved. The error function is:

[0029]

[0030] Among them, d(X c ,y c ,C) represents the closest distance between the central grid seed point and the geometric contour C within the range, and λ represents the balance parameter between the distance between the grid point and the contour.

[0031] Preferably, the area of ​​the retained Thiessen polygon is calculated and compared with the threshold S area Compare, if it is less than the threshold S area , then merge the Thiessen polygon area with the surrounding adjacent areas, or extract the seed points of the corresponding Thiessen polygons, reset and generate new grid seed points.

[0032] Preferably, hatchspace / 2 is biased inwardly, where hatchspace is the scanning spacing in the laser scanning process.

[0033] Beneficial effects: The present invention records a random polygon partitioning forming method applied to additive manufacturing, which realizes a random partitioning scanning strategy based on Voronoi polygons. It not only optimizes the laser path through partitioning to reduce residual stress, but also ensures the consistency of microstructure forming at different positions of different slice layers during the forming process of the part through random partitioning, reduces the risk of failure caused by thermal cracks and residual stress during the forming process of the part, avoids the problem of spatial overlap between layers at the partition boundary position in the traditional partitioning method, and improves the uniformity of the microstructure distribution of the part during the laser forming process.

[0034] In addition, the present invention also controls the size of the Thiessen polygon by predefining the grid size, and optimizes the generation position of the initial Thiessen polygon seed point by judging the position relationship between the slice polygon and the grid; further, combined with the judgment of the area of ​​the cropped Thiessen polygon and iterative generation of the edge position seed point coordinates, the appearance of small-area Thiessen polygons in the edge part after cropping, that is, low-quality random partitioning, is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0036] In the attached picture:

[0037] Figure 1 is a flow chart of the partition forming method of the present invention;

[0038] Figure 2 It is a distribution diagram of the geometric contour C, the mesh generation area A and the label value m corresponding to the mesh generation area of ​​the present invention;

[0039] Figure 3 It is a schematic diagram of generating seed points in different grids of the present invention;

[0040] Figure 4 This is a schematic diagram of the voronoi polygons generated by different algorithms;

[0041] Figure 5 It is a comparison diagram of the random partitioning of Thiessen polygons and the rectangular partitioning of the present invention;

[0042] Figure 6 It is a schematic diagram of the filling laser scanning path after biasing of the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose and advantages of the present invention more clearly understood, the present invention is specifically described in conjunction with the following embodiments. It should be understood that the following text is only used to describe a random polygon partition forming method or several specific implementation methods of the present invention for additive manufacturing, and does not strictly limit the scope of protection specifically requested by the present invention.

[0044] In order to solve the problems of uneven microstructure distribution and overlapping of interlayer partitions in laser powder bed fusion technology, this paper proposes a random polygon partition forming method for additive manufacturing. This method significantly improves the performance of parts in the laser forming process by optimizing the processing of slice geometric boundaries, the generation of seed points, and the cutting and splicing of Thiessen polygons.

[0045] Example: Reference Figure 1 As shown, a random polygon partition forming method applied to additive manufacturing is shown, and the specific steps include:

[0046] S1. Obtain the envelope boundary of the slice layer geometry and offset it outward to generate a grid coverage area C. The offset distance can be d, d = 2a, a is the grid unit size; refer to Figure 2 As shown;

[0047] Based on the positional relationship between the geometric outline and each grid area in the grid coverage area, the corresponding grid area is marked, including outlying, intersecting and containing;

[0048] That is, the grid coverage area is divided into grid sub-areas g ij , according to the sub-region g ij The center point g ij (x,y) and the four vertex coordinates g ij (x n ,y n ) and the slice layer geometric contour C, and is labeled m ij , m ij The following constraints are met:

[0049]

[0050] refer to Figure 3 As shown: When the grid area g ij When it is included, the grid area gi j When completely inside the contour range C, mark m ij =1, when the grid area g ij When the grid area gi j There is an intersection with the contour range C, and the intersection area is greater than the design threshold Arealimit, marked m ij=0, in this embodiment, Arealimit=0.5, where the design threshold value is adjusted according to the actual model; when the grid area g ij When it is outward, that is, when the grid area g ij When it is completely outside the contour range C or the intersection area is less than or equal to the design threshold Arealimit, mark m ij = -1;

[0051] S2, based on the grid mark in step S1, preferentially generate seed point coordinates in the grid with grid mark m=1 or -1; randomly generate a seed point coordinate for each grid area; the seed point coordinate p m＝1,-1 (x, y) in the grid range g ij The coordinates of the generated seed points are randomly generated:

[0052] p m＝1,-1 (x,y)=(xmin+rand·(xmax-xmin),ymin+rand·(ymax-ymin));

[0053] Among them, xmin, xmax, ymin, and ymax represent the grid area g respectively. ij The boundary range;

[0054] S3. Construct a data structure unit based on the above construction:

[0055] Gi j ={m ij ,[center x ,center y ],[gx1,gy1],[gx2,gy2],

[0056] [gx3,gy3],[gx4,gy4],[C x ,C y ],[p x ,p y ]} As the basic data unit format, each structure unit format corresponds to a grid; the structure unit stores m ij , [center_x, center_y] are the coordinates of the center of the grid, [gx1, gy1], [gx2, gy2], [gx3, gy3], [gx4, gy4] are the coordinates of the four vertices of the grid, [C x ,C y ] is the centroid of the intersection of the grid and the slice contour C, [p x ,p y ] are the coordinates of the seed point in the grid.

[0057] S4, based on the seed point coordinates p generated in step S2 m＝1,-1(x, y), generate the seed coordinates q(x, y) in the grid with label m=0; specifically including the following sub-steps:

[0058] S41, select the grid area with label m=0, calculate the grid g ij The centroid coordinates of the intersection area with contour C [C x ,C y ]:

[0059]

[0060] Among them, (x i ,y i ) is the vertex coordinate of the intersection area, A is the area of ​​the intersection area; after the calculation is completed, the centroid coordinates are stored in the corresponding grid structure unit;

[0061] S42, read all adjacent grids g around the m=0 grid area ij The seed point coordinates p ij (x, y); if there is no seed point in the surrounding grid, a point coordinate is randomly generated within the grid range as a temporary seed point to participate in subsequent calculations. After the calculation is completed, the temporary seed point is reset; calculate the coordinates of various sub-points p respectively ij (x,y) and the centroid coordinates of the intersection area [C x ,C y ]'s Euclidean distance L ij ;

[0062] S43, based on the surrounding grid seed point p of S42 ij Coordinate position and its corresponding geometric distance L ij , find the influence weight w corresponding to the surrounding seed points ij ; The mathematical description of the generation process of the center grid seed point coordinate q(x,y) can be expressed as:

[0063]

[0064] S44, construct an error function, solve the minimum error function optimization problem, and determine the coordinate value of the seed point of the center grid (m=0);

[0065]

[0066] Among them, d(X c ,y c ,C) represents the closest distance between the central grid seed point and the contour C within the range, λ represents the balance parameter between the distance between the grid point and the contour, for example, λ=1;

[0067]

[0068] where wij (L ij ) indicates that the ij The calculated weight coefficient represents the influence of the seed points in the corresponding adjacent grids on the generation of the coordinates of the seed points in the central grid. In this example, it can be described as:

[0069]

[0070]

[0071] Among them, w ij The weight function in this example uses a Gaussian function, which is suitable for smoothing the influence of adjacent grids on the central grid. σ represents the influence parameter value. For example, σ = 0.5 is adjusted according to the actual situation, representing the influence effect of the adjacent grids. w ij ' is the normalized weight coefficient;

[0072] Step S45, traverse each grid area with label m=0, and loop step S4 until all grids with m=0 in the range generate new seed points;

[0073] S5. Generate Voronoi polygons based on the seed points in step S4, refer to Figure 4 As shown, find the intersection of the geometric contour C and the Voronoi polygon, retain the Voronoi polygon within the contour C; calculate the area S of the retained Voronoi polygon i And compare with the threshold S area , for example, take S area =0.5, if S i area , then determine to merge the area with the surrounding adjacent areas or extract the seed points corresponding to the Voronoi polygons, reset and repeat step S4 to generate new grid seed points;

[0074] refer to Figure 5 As shown, in the partition forming method between layers of the rectangular partition scanning strategy, there is spatial boundary overlap between layers, while in the Thiessen polygon partition scanning strategy of the present invention, the partition between layers is completely random and there is no spatial overlap problem on the boundaries.

[0075] Step S6: Based on the polygon clipped in step S5, offset hatchspace / 2 inward to obtain all partitions within the geometric contour C of the slice layer, refer to Figure 6 As shown, hatchspace is the scanning spacing in the laser scanning process, for example, hatchspace=0.07; an internal laser scanning trajectory is generated based on the partition after the internal bias;

[0076] ​Step S7: Repeat the above steps S1-S6 to obtain the partitioned laser scanning trajectory applied to the model, compile the obtained laser scanning trajectory into a machine code that can be recognized by the printing device and complete the printing.

[0077] The embodiments of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above embodiments. For ordinary technicians in this technical field, after knowing the contents recorded in the present invention, they can make several equivalent changes and substitutions without departing from the principle of the present invention. These equivalent changes and substitutions should also be regarded as belonging to the protection scope of the present invention.

Claims

1. A random polygon partitioning forming method for additive manufacturing, by partitioning each slice layer and repeating the slice layer partitioning step, to obtain a partitioned laser scanning trajectory applied to the model, characterized in that: Slice layer partitioning includes the following steps: Obtain the envelope boundary of the slice layer geometry contour and offset it outward to generate a grid coverage area; Based on the positional relationship between the geometric outline and each grid area in the grid coverage area, the corresponding grid area is marked, including outlying, intersecting and containing; Generate seed point coordinates in all outer and inner grid regions; In the intersecting grid area, the corresponding influence weight is obtained based on the seed point coordinates of the adjacent grid area; as well as Based on each influence weight and constructing an error function, the seed point coordinates of the intersecting grid area are obtained; Construct Thiessen polygons based on all seed point coordinates and use the slice layer geometry outline to preserve the Thiessen polygons inside it; The retained Thiessen polygons are biased inward to obtain all partitions within the geometric outline of the applied slicing layer.

2. A random polygonal partition forming method for additive manufacturing according to claim 1, characterized in that: The offset distance is d, where d=2a, and a is the grid unit size.

3. The random polygon partition forming method for additive manufacturing according to claim 1, characterized in that: The marking constraints of each grid area are: ij When included, mark m ij =1, when the grid area g ij When they intersect and the intersection area is greater than the design threshold Arealimit, mark m ij =0; when the grid area g ij If the outer separation or intersection area is less than the design threshold Arealimit, mark m ij =-1.

4. The random polygon partition forming method for additive manufacturing according to claim 1, characterized in that: The coordinates of the seed points generated in the outer and inner grid areas are: P m＝1，-1 (x,y)=(xmin+rand·(xmax-xmin),ymin+rand·(ymax-ymin)); Among them, xmin, xmax, ymin, and ymax represent the grid area g respectively. ij boundary range.

5. The random polygon partition forming method for additive manufacturing according to claim 4, characterized in that: A seed point coordinate is randomly generated for each grid area.

6. A random polygonal partition forming method for additive manufacturing according to any one of claims 1 to 5, characterized in that: The influence weight is obtained specifically as follows: Calculate the centroid coordinates of the intersection area of ​​the intersecting grid area and the geometric contour [C x ,C y ]: Among them, (x i ,y i ) are the vertex coordinates of the intersection area, and A is the area of ​​the intersection area; Calculate the Euclidean distance between the seed coordinates in all adjacent grid areas around the grid area and the centroid coordinates; The corresponding influence weight is calculated based on the Euclidean geometric distance: Among them, L ij is the Euclidean geometric distance, σ represents the influencing parameter value, and w ij ′ is the normalized weight coefficient.

7. The random polygon partition forming method for additive manufacturing according to claim 6, characterized in that: A point coordinate is randomly generated in the grid area adjacent to the intersecting grid area and without the current seed coordinate as the temporary seed point coordinate for calculation, and the temporary seed point coordinate is reset after the calculation is completed.

8. The random polygon partition forming method for additive manufacturing according to claim 6, characterized in that: Construct an error function and solve the minimum error function. The error function is: Among them, d(X c ,y c ,C) represents the closest distance between the central grid seed point and the geometric contour C within the range, and λ represents the balance parameter between the distance between the grid point and the contour.

9. The random polygon partition forming method for additive manufacturing according to claim 1, characterized in that: Calculate the area of ​​the retained Thiessen polygon and compare it with the threshold S area Compare, if it is less than the threshold S area , then merge the Thiessen polygon area with the surrounding adjacent areas, or extract the seed points of the corresponding Thiessen polygons, reset and generate new grid seed points.

10. The random polygon partition forming method for additive manufacturing according to claim 1, characterized in that: Bias inward by hatchspace / 2, where hatchspace is the scanning spacing in the laser scanning process.

Images