KR101956271B1 - Tree type support and optimum design method thereof - Google Patents

Abstract

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a tree-shaped support and a design method thereof, and more particularly to a tree-shaped support capable of minimizing material and output time in an output of a 3D printer and an optimization design method using an artificial neural network .

Description

[0001] DESCRIPTION [0002] TREE TYPE SUPPORT AND OPTIMUM DESIGN METHOD THEREOF [

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a tree-shaped support and a design method thereof, and more particularly to a tree-shaped support capable of minimizing material and output time in an output of a 3D printer and an optimization design method using an artificial neural network .

FDM (Fused Deposition Modeling), which has become popular among 3D printer methods, has a method of melting and stacking materials. The FDM method has advantages such as purchase cost, maintenance cost, and space utilization compared with other types of printing, but it has disadvantages such as monochromatic output, generation of pattern according to lamination, generation of support according to conditions of some shapes, and relatively slow speed Have.

The support supports the shape when the shape hangs from the object or when it outputs the shape floating from the ground. In the FDM 3D printing, problems due to the machined surface and support are generated.

Specifically, in the case of a single nozzle printer, after the output, since the support is discarded, it leads to waste of the material, and since the support and the substantially necessary shapes are laminated together with the same material, the more the contact surface with the support is, And a problem arises that requires additional post-processing.

The basic support shape of the slicer is formed to have a minimum output thickness according to an interval defined by the user when the angle of the shape is 45 degrees or less, and to support the vertical wall from the ground at regular intervals. Since the existing support shape has a narrow or wide gap between the structures, the completeness of the surface at the time of removing the support is lowered, and the support configuration is inefficient in comparison with the support area.

Korean Patent Publication No. 10-2017-0014619

SUMMARY OF THE INVENTION It is an object of the present invention to provide a tree-shaped support for preventing output time and material waste due to the inclusion of a support and an optimization design method thereof.

A triple support according to an exemplary embodiment of the present invention is a support for supporting a molding projected to be separated from a bed of a 3D printer, the support comprising: pillars provided at a predetermined height on an upper surface of the bed; And a branch portion provided between the column portion and the sculpture to support the sculpture.

In addition, the branches may include a plurality of upper branches arranged at regular intervals such that the upper ends thereof are in contact with the molding, and a plurality of lower branches extending downward from the respective upper branches.

In addition, the gap between the plurality of lower branches becomes narrower toward the lower side.

Further, the shape of the support depends on the mass of the molding to be laminated on the upper side

And is changed.

In addition, the support may be characterized in that the diameter of the column portion is changed according to the mass of the molding to be laminated on the upper portion.

Further, the cross-sectional area of the column portion may be smaller than the area of the molding to be laminated on the upper portion of the support.

Further, the number of the upper branches may be determined by the following formula (1).

[Equation 1]

(N _b : total number of upper branches, W _m : length of sculpture, B _side : number of branches in both directions)

A method of optimizing a tree-type support according to an exemplary embodiment of the present invention includes: a columnar portion having a predetermined height on an upper surface of a 3D printer bed; And a branch portion provided between the column portion and the sculpture to support the sculpture, the method comprising the steps of: (S100) setting a design condition of the triple support; Determining a shape design parameter of the triple support (S200); Extracting sample data using the determined design variables (S300); And learning the sample data through an artificial neural network (S400).

In addition, the step of extracting the sample data may be performed through an optimization analysis method after finite element analysis.

In addition, the design parameters include the width W _m , the length H _m , the height D _{m of} the molding, the total height H _t of the support, the radius R _c of the column, and the height H _b of the upper branch .

The triple support according to the preferred embodiment of the present invention has the effect of enabling the output of a highly completed sculpture by appropriately supporting a required portion while consuming a smaller amount of material than the shape of an existing support.

In addition, it is possible to design an optimized tree type support without the procedure of repeated finite element analysis and optimization through artificial neural network learning.

1 is a conceptual view of a tree-shaped support according to an embodiment of the present invention;
2 is a flow chart showing a method for optimizing design of a tree-like support according to an embodiment of the present invention.
FIG. 3 is a flow chart schematically showing a method of extracting sample data for artificial neural network learning; FIG.
4 is a flowchart schematically showing a method of optimizing design variables through finite element analysis.
FIG. 5 is a graph showing an error of an artificial neural network learning result. FIG.
FIG. 6 is a graph showing a result of regression analysis on an artificial neural network learning result. FIG.
7 is a diagram showing a support model actually printed based on a result of learning an artificial neural network;

It is to be understood that the words or words used in the present specification and claims are not to be construed in a conventional or dictionary sense and that the inventor can properly define the concept of a term to describe its invention in the best way And should be construed in accordance with the meaning and concept consistent with the technical idea of the present invention. Therefore, the embodiments described in the present specification and the configurations shown in the drawings are merely the most preferred embodiments of the present invention and are not intended to represent all of the technical ideas of the present invention. Therefore, various equivalents It should be understood that water and variations may be present. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a conceptual diagram of a tree-shaped support 200 according to an embodiment of the present invention. 1, a tree-type support 200 according to an embodiment of the present invention basically supports a molding 100 that is spaced apart from a bed (not shown) of a 3D printer, A columnar section 250 provided at a predetermined height on the upper surface of the columnar section 250; And a branch part 210 provided between the column part 250 and the molding 100 to support the molding 100.

The branch 210 includes a plurality of upper branches 211 directly contacting and supporting the surface of the molding 100 and a plurality of lower branches 212 extending downward from the respective upper branches 211 can do.

A plurality of upper branches 211 are formed to have a predetermined height so as to be perpendicular to the ground at regular intervals on a predetermined portion of a lower surface of the molding 100 requiring support and a lower branch 212 is formed from each of the upper branches 211 And extends downward.

The upper branches 211 are formed so as to be perpendicular to the paper surface while the lower branches 212 extending from the lower end surface of the upper branches 211 are formed to be inclined toward the upper surface of the pillars 250. Accordingly, in the preferred embodiment of the present invention, as shown in FIG. 1, the intervals of the plurality of lower branches 212 are narrowed downward to form a tree shape having an upward light-tight conformation.

Generally, as the distance between supporters becomes narrower, the consumed amount of the material increases and the post treatment after the output may be disadvantageous. On the contrary, the farther the interval is, the more advantageous is the opposite. However, There is a possibility that the quality of the display device 100 may be lowered.

Therefore, as will be described later, in the optimization design of the tree-shaped supports 200 according to the embodiment of the present invention, the intervals are modeled at intervals of 1.5 mm to 4.0 mm at intervals of 0.5 mm, . In the case of a support having a spacing of 4.0 mm, the post-treatment capability was excellent, but the lamination on the support was unstable. In the case of 1.5 mm, the post-treatment capability was severely deteriorated, Respectively. On the other hand, the post-treatment of 2.5 mm was not problematic, and the lamination was the closest to the design value. Therefore, in the preferred embodiment, the branch shape spacing of the tri-shaped supports 200 is set to 2.5 mm.

It is preferable that the pillars 250 are positioned at the central portion of the area required for supporting the molding 100 for the purpose of lamination stability and the angle at which the lower branches 212 extend from the top surface of the pillars 250 is It is preferable that the thickness is about 45 DEG in consideration of the lamination characteristics.

In addition, as will be described later, in the optimization design of the tree-shaped support 200 according to the embodiment of the present invention, the triple support 200 is supported by the supports and is shaped according to the shape of the molding 100, Can also be changed. In the Mendel type 3D printer, since the external force can be applied to the support in the process of conveying the Y-axis of the bed, the triple support 200 is provided so as to minimize the consumed amount of the material while having the optimum rigidity, To change the shape of the surface.

The number of total upper branches 211 required for a certain area of the molding 100 required to support is determined by the width of the supported molding 100 and the distance between the upper branches 211, And can be determined in detail by the following equation (1).

Here, referring to FIG. 1, Nb denotes the number of the upper branches 211, Wm denotes the width of the molding 100, and Bside denotes the number of both ends in the width direction.

In addition, the lower ends of the lower branches 212 are connected to the pillars 250 through the lower branch receiving portions 220, so that the stability can be maintained.

Since the magnitude of the external force applied to the column 250 in the process of transferring the bed can be determined by the sum of the mass of the molding 100 and the mass of the branch 210 stacked on the support, It is possible to minimize the movement of the column due to the mass of the sculpture 100 supported by the tongues 210 and the supports during the design process and minimize the volume of the entire support to minimize the consumption of the material. The radius, that is, the diameter, and the height of the upper branch 211 can be changed.

As described above, according to the embodiment of the present invention, the shape of a tree and the support pillars including the conventional support shape capable of appropriately supporting the surface of the molding object 100 to be outputted are narrowed. That is, by forming the cross section area of the columnar section 250 smaller than the area of the molding 100 stacked on the upper part of the support, it is possible to appropriately fill the necessary part while consuming a smaller amount of filament than the shape of the support recommended by the conventional slicer Thereby enabling a high-quality printout.

FIG. 2 is a flow chart illustrating a method for optimizing a tree-type support 200 according to an embodiment of the present invention. Next, an optimization design method of the triple support 200 proposed by the present invention will be described by way of examples.

In the embodiment of the present invention, a Menden Type-Prusa i3 3D printer was used, and PLA (Poly Lactic Acid) was used as a filament material.

Table 1 below shows the hardware and software used to derive the preferred embodiment of the present invention and activates the Rid function to minimize the seating problem between the desired molding 100 and the bed of the printer.

Hardware Model Prusa i3 Type Mendel Nozzle Size 0.4mm Filament Material PLA + Filament diameter 1.75mm Software Host Program Repetier V1.6.2 Slicer CuarEngine v3 Quality 0.1mm Print Speed 30-40mm / s Extruder
Temperature 210 ℃ Bed Temperature 55 ° C

A method of optimizing a tree-type support 200 according to an embodiment of the present invention includes a column 250 provided at a predetermined height on a top surface of a 3D printer bed; A step (S100) of setting a design condition of the tree-like support (200) including a column portion (250) and a branch portion (210) provided between the columnar portion (250) and the molding material (100) to support the molding material (100); Determining a shape design parameter (S200); Extracting sample data using the determined design variables (S300); And learning the sample data through an artificial neural network (ANN) (S400).

First, a step (S100) of setting a design condition of the triple support 200 proposed by the present invention may be included. Specifically, the tri-shaped support 200 proposed in the embodiment of the present invention maintains a conventional support shape capable of appropriately supporting the surface of the molding 100, This shape allows the required part to be adequately supported by consuming a smaller amount of filament than the shape of the support recommended in the existing slicer, thereby enabling a high-quality output. The design of the following triple support 200 Conditions were set.

Condition 1. The area of the support that touches the shape is minimized.

Condition 2. The lamination on the support should be correct.

Condition 3. The distance between the branches of the material in the tree structure is maximized.

Condition 4. Minimize the shaking of the support due to the movement of the base during output.

Conditions 1, 2, and 3 are for saving material and optimizing the completeness of shape, and condition 4 is due to the characteristics of the Mendelian method.

The Mendel type 3D printer used in the present invention is an open source and is applied to a large number of products currently available in the market. The head is connected to the X-axis conveying part and the bed located thereunder is connected to the Y-axis conveying device The head moves in the X-axis direction while the bed moves in the Y-axis direction to set the coordinates. Such a Mendel type printer should have no problems in the output of the support branch shape according to the principle of output in the Y axis transportation process including the bed, which is determined by the lower column constituting the support. As the diameter of the column is increased, a relatively large number of filaments are consumed, and the branch shape and the lamination of the molding 100 are output without any problem in various feeding conditions of the bed.

In the present invention, a specific shape design parameter (hereinafter, referred to as a design parameter) is set in order to model the shape of the triangle support 200 meeting the above conditions, and a model for determining the shape design of the triangle support 200 using the artificial neural network It suggests.

The triangular support 200 proposed in the present invention is composed of a branch shape and a column based on the object 100 to be output, and the upper end of the branch shape, that is, the upper branch 211, Basic linear pattern support is used. At this time, as the spacing of each branch becomes denser, the consumed amount of material increases, and it is disadvantageous to the post-treatment after the output, and as the interval becomes longer, the opposite advantage is obtained. However, Can be obtained.

Therefore, it is preferable to set the intervals between 2.0 mm and 3.5 mm as a result of modeling the intervals at intervals of 0.5 mm to 1.5 mm and outputting them and comparing the quality. In the case of a support with a spacing of 4.0 mm, the post-treatment capability was excellent, but the lamination on the support was unstable. In the case of 1.5 mm, the post-treatment capacity was severely reduced, and the thickness of the main shape became thicker. On the other hand, in the case of 2.5 mm, relatively post-treatment was not a problem, and the lamination layer closest to the design value was formed. Therefore, in the preferred embodiment of the present invention, the branch shape spacing of the triple support 200 is set to 2.5 mm.

The optimization design method of the tree-type support 200 according to the embodiment of the present invention may include a step S200 of determining a design parameter for the tree shape set as described above, and as shown in Fig. 1, The height H _b of the molding 100 and the upper branch 211 designed with the width W _m , the length H _m and the height D _m , the radius R _c of the lower column 250 and the total height of the triangle support 200 H _{t, and} is designed so that the entire tree shape can be changed by changing the shape of the molding 100 to be laminated.

The starting point of the column shape was designed to be the center of the area required for the support for the stability of the lamination, and was set at 45 ° in consideration of the lamination characteristics in which the branches of the supports were extended. In addition, the uppermost support pattern was modeled as a wall thickness of 0.45 mm according to the minimum slicing size of the slicer, and the gap was generated on the basis of 2.5 mm as described above. In this case, if the total number of branches to be generated is N _b , and the number of branches at both ends is defined as B _side , N _b is determined according to the above-mentioned expression (1). Also, the height H _c of the columnar shape Can be expressed by the following equation (2).

Referring to FIG. 1, H _c denotes the height of the column, H _t denotes the total height of the triangular support 200, and W _m denotes the width of the molding 100.

The horizontal, vertical, thickness and height of the molding 100 to be stacked on the support are selected as learning independent variables so that the shape of the support can be changed according to the weight of the molding 100 stacked on the support. By setting the radius and the height of the upper branch 211 as the dependent variable according to the independent variable, the shape is designed to be deformable in real time according to the 163 arbitrary independent variables. Here, the arbitrary independent variables are set such that the upper limit and the lower limit are extracted so as to be within the range of [Table 2] considering the smoothness of the actual output and the finite element operation.

Parameter Inf. Range Sup. Range

5 mm 50 mm

5 mm 70 mm

1 mm 10 mm

10 mm 50 mm

Next, the method may include extracting sample data for artificial neural network learning (S300) using the determined design variables as described below. In order to extract the sample data, sufficient data can be obtained by repeating the finite element method and the optimization analysis.

Specifically, in the preferred embodiment of the present invention, the optimal columnar radius R _c and the height H of the support upper branch 211, which minimizes the volume of the column and the column support due to the mass of the molding 100, _b .

In order to find the optimum shape according to the analysis result, finite element and optimization analysis should be repeated for each sample data type. Therefore, in order to reduce the complexity of the tree shape and to minimize the error of the result value and to increase the efficiency of the experiment, the following conditions are assumed in the embodiment.

First, in a Mendel-type printer in which a Y-axis bed feed is carried out, the bending stiffness of the lower column is greatest when the Y-axis bed is stopped after the conveyance, so that the reference setting for limiting the degree of change of the support column The reference of the maximum displacement value was set to 0.005 mm.

For this purpose, the criterion for limiting the degree of change of the support column must be clarified and the optimization design must be made accordingly. Therefore, the behavior of the column is numerically confirmed. A cylindrical shape with a radius of 2.0 mm and a height of 10 mm was output and the force required for 1.0 mm deformation was measured 10 times by fixing the shape.

As a result of measuring the force, the average value of 1.72 N was obtained and the behavior was confirmed by applying it to the same finite element model as the output. The result of the finite element model shows that the displacement value changes up to 0.207 mm. In order to reduce external factors such as external temperature of actual output, friction occurring during lamination, shrinkage, etc., and to obtain a safety result, the standard of maximum displacement value is determined as 0.005 mm.

It is assumed that the force applied to the upper end of the column at the time of transferring the bed has a value similar to the product of the sum of the mass of the tree excluding the column and the stacked mass thereon and the acceleration, mm / s, and the time of one step until stopping was set to 0.05 Sec. At this time, it was assumed that the settlement problem of the bed, the support, and all the stacked shapes did not occur.

Under the condition described above, the condition of the external force was set by the product of the acceleration and the mass of 0.8 m / s ^{2 in} the conveying direction Y axis while the lower part of the support was fixed to the bed. By mass to determine the value of the external force here can be calculated as the sum of the weight (m _m) of the mass of the shape of the support other than the support pillar (m _b) and sculptures 100. The physical property values of the support shape were applied as shown in Table 3 with reference to the specifications of the PLA + product used in the examples.

Material

)Density
(

) Young's
Modulus (

) Poisson's
Ratio PLA + 1240 3.5e + 009 0.36

The process of optimizing the result of the finite element analysis may be performed as follows, as shown in FIG. First, each support model is determined by the above-described finite element analysis, and each support model is sequentially selected, and optimization data can be found through CAE (computer aided engineering). At this time, sample data can be secured by storing only the data of the model in which the displacement of the element satisfies 5e-04 mm or less according to the result of the finite element analysis.

In the embodiment of the present invention, CATIA's Product Engineering Optimizer is used as an optimization solution tool, and a parabolic approximation for calculating the slope of an objective function is applied using an algorithm of the Gradient Algorithm With Constraints type to find an optimal value .

Specifically, in order to derive the optimal column radius value and the height value of the upper branch 211 according to 163 support shapes composed of arbitrary independent variables, the volume of the entire shape among the optimization tool conditions is set as the optimization minimum value And the additional condition is set so that the displacement of the element is less than 0.005 mm by referring to the result of the above-described finite element analysis. The range of the dependent variable R _c and H _b is set as shown in Table 4 below.

Object Range Optimized
parameter Volume Minimum Independent
Variable R _c 0.25 - 3.0 mm Independent
Variable H _b 2.0 - 5.0 mm Constraint Distance ≤ 5e-04 mm

At this time, R _c And the minimum value of H _b was determined to be 2.0 mm considering the ease of post-processing.

The obtained sample data includes a learning step (S400) using an artificial neural network, which is one of the learning machines, so that it is possible to design the tree-shaped support 200 optimally rotated without the procedure of optimization and finite element analysis will be.

In the embodiment of the present invention, the design parameters W _m , H _m , D _m , and H _t of the tri-shaped support 200 and the molding product 100 can minimize the volume of the support and the shake of the shape of the printer bed during transportation And the design variables H _b and R _c are used as output variables. Where the variables H _b is the time to have the minimum of all 2.0 mm in the 163 pieces of data as it is possible to configure the optimum shape according to a H _b with the exception, and the input data W _m, H _m, D _m, H _t target value R _c Artificial neural network learning was carried out. Table 5 below shows some of the sample data for arbitrary input variables and objective values.

Part Number W _m
(mm) H _m
(mm) D _m
(mm) H _t
(mm) R _c
(mm) # 001 20 20 2 20 0.581 # 002 43 17 5 48 1.722 # 003 23 17 9 42 1.79 # 004 25 47 3 45 2.002 # 005 25 7 3 22 0.497 # 006 20 9 7 25 0.826 # 007 40 20 2 50 1.682 # 008 23 34 5 39 1.737 # 009 11 26 4 42 1.556 # 010 42 57 8 23 0.624

# 163 46 16 8 41 1.434

Specifically, in the embodiment of the present invention, the software for learning an artificial neural network is constructed using Matlab and a gradient descent with momentum and adaptive learning rate (GDX), which can set learning rate and momentum of learning progress based on feed- back-propagation type functions. The learning rate was 0.01 and the momentum value was 0.7. The results were more accurate than the fast learning time and the error was based on the Mean Squared Error (MSE). In addition, the completion of the learning model is enhanced by integrating the continuous verification process during learning. In the learning of the algorithm, 10% of the total data is used as the criterion for evaluating the reliability of the algorithm.

As a result of learning the sample data using the artificial neural network, as shown in FIG. 5, the MSE value reaches 0.0083 through 7128 learning, and the value of the determination coefficient R is 0.98977 as shown in FIG. 6 We confirmed that the algorithm that is learned is appropriate.

After learning through the artificial neural network, step S500 of verifying the learning result may be further included. In the embodiment of the present invention, the verification step is proceeded as follows.

7 is a diagram illustrating a support model actually printed on the basis of a learning result using an artificial neural network, and Tables 6 to 7 below show how the shape of a support predicted through learning is versatile and consistent output is possible under the same output condition 7 is a table comparing the actual printing possibility and the reduction of the material compared to the support shape of the existing slicer by comparing the displacements of the respective lattice nodes through the finite element analysis using the support model of Fig.

The model used in this test is to model five arbitrary shapes and extract the same design variables W _m , H _m , D _m , and H _t as the experiment performed in the paper, and set them as input variables, To form a support shape. These shapes were analyzed by finite element analysis on the condition of the movement of the actual printer bed. As a result, as shown in Table 6, the maximum error with 0.0005 mm of the average displacement of all the nodes of the column was 0.00041 mm in <Model # 1> shape, and the average error rate of 5 shapes was 4.2% It is possible to confirm that the shape satisfies the theoretical output condition at the same time as the prediction of the optimized design condition.

In addition, the filament consumption was compared through the G-code conversion in the slicer software, the actual printing was performed, and the expected material consumption of each shape was compared and shown in Table 7. As a result, a maximum saving of about 1.9 times was achieved in <Model # 5> among the entire models of FIG. 7, and the reduction of material use was observed as a whole, and the reliability of the shape was confirmed by showing a normal output as shown in FIG. .

As described above, in the preferred embodiment of the present invention, a support shape for Mendel type low cost FDM 3D printer and its optimization design method are proposed. A suitable tree type support (200) design method is presented, and a sample shape is optimized to optimize volume and strength with optimization tools and finite element analysis. We constructed a method for designing the optimized tree type support (200) without the procedure of optimization and finite element analysis by constructing learning sample data by extracting arbitrary design variables from this shape and learning by artificial neural network. It was confirmed that an average of 40% of the material reduction effect was obtained by using one of five shapes using the tree-type support 200. The surface of the molding 100, which is in contact with the support while using such a small amount of material, It is possible to have an effect.

Predicted Models Model # 1 Model # 2 Model # 3 Model # 4 Model # 5 Optimum
Radius
(mm) 0.382 0.649 0.812 0.561 1.074 Predicted
Radius
(mm) 0.366 0.659 0.723 0.577 1.086 Displacement
Magnitude
(mm) 8.66
E-005 4.57
E-004 7.90
E-004 4.20
E-005 4.81
E-004 Gap
of Value
(mm) -4.13
E-004 -4.27
E-005 2.90
E-004 -8.03
E-005 -1.89E-005

Filament needed Models
Basic Support
(mm) Optimum
Support (mm) Decrease
Rate (%) #One 618 581 6.37 #2 2335 2184 6.91 # 3 1449 986 46.96 #4 2083 1347 54.64 # 5 832 441 88.66

Although the present invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, The present invention is not limited thereto. It will be apparent to those skilled in the art that various substitutions, modifications and variations are possible within the scope of the present invention, and it is obvious that those parts easily changeable by those skilled in the art are included in the scope of the present invention .

100: Sculpture
200: Tree type support
210:
211: Upper branch
212:
220: lower branch receiving portion
250:

Claims (10)

A support for supporting a sculpture output from a bed of a 3D printer,
A column having a predetermined height on the upper surface of the bed;
A branch portion provided between the column portion and the sculpture to support the sculpture;
Lt; / RTI &gt;
In the branching portion,
A plurality of upper branches arranged at regular intervals such that an upper end thereof contacts the molding;
A plurality of lower branches extending downward from the respective upper branches and spaced apart from each other, the lower branches becoming narrower toward the lower side;
/ RTI &gt;
A lower branch receiving portion for receiving the lower ends of the plurality of lower branches;
And the number of upper branches is determined by Equation (1).
[Equation 1]

(N _b : total number of upper branches, W _m : length of sculpture, B _side : number of branches in both directions)
delete delete The method according to claim 1,
Wherein the shape of the support is changed according to the mass of the molding to be laminated on the upper part.
5. The method of claim 4,
Wherein the support has a diameter varying in accordance with a mass of a molding to be laminated on the upper portion.
5. The method of claim 4,
And the cross-sectional area of the column portion is smaller than the area of the molding to be laminated on the upper portion of the support.
delete A column portion provided at a predetermined height on the upper surface of the 3D printer bed;
A plurality of upper branches provided between the pillars and the sculpture to support the sculptures and provided at regular intervals such that the upper ends of the sculptures come in contact with the sculptures; 1. A method of optimizing a tree-type support including a plurality of lower branches that gradually become narrower,
Setting a design condition of the tree-shaped support (S100);
(S200) determining shape design parameters of the triple support through Equation (1);
Extracting sample data using the determined design variables (S300); and
Learning the sample data through an artificial neural network (S400);
A method for optimizing design of a triangular support for manufacturing the triangular support of claim 1,
9. The method of claim 8,
Wherein the step of extracting the sample data is performed through an optimization analysis method after finite element analysis.
9. The method of claim 8,
Wherein the design parameter includes a width W _m , a length H _m , a height D _m of the molding, and a total height H _t of the support, a radius R _c of the column, and a height H _b of the upper branch. Optimized design method of tree type support.

Images