One of the design theories and methodologies, design for additive manufacturing (DFAM), is of great importance in making additive manufacturing (AM) feasible for the industry. Because AM is not currently applicable to every sector in the industry, it is best to take full advantage of AM. In this work, a lattice structure generation method is presented to facilitate the implementation of DFAM concepts in AM part design.
This method is notable because, not requiring deep understanding and knowledge of FEA and TO, it is easily accessible to the general public. 40 Figure 3.22 (a) General stress state of an element with principal stresses σmax and σmin. b) Mohr's Circle representation of the general stress status.
Background
DFAM was developed because parts designed using conventional design methodology are not suitable when manufactured with AM; parts specifically designed for AM through DFAM have shown great potential for improvement [9]. For example, the fabrication of an overhang requires a support structure, except when using the selective laser sintering method, as shown in Figure 1.2 (a). Support structures must be installed carefully, as they affect the construction time, cost and removal process.
In the case of the powder bed fusion method, closed voids should be avoided as excess material can be trapped inside, as illustrated in Figure 1.2 (b). Although there are many cases that show improvements made with AM, many companies are still not using DFAM methods in the production of final parts.
Objectives
Outline
Feasibility of Additive Manufacturing
13] demonstrated the economic feasibility of the AM process compared to the high pressure die casting (HPDC) process. HPDC and AM process cost evaluation models were developed and 1:5 model aircraft landing gear production was analyzed. For the analysis, the chassis was redesigned to take advantage of the AM process and validated using ABAQUS finite element analysis.
The cost per assembly with increasing production volume showed that AM has a competitive advantage in terms of costs under the production volume of 42 pieces. Because HPDC requires expensive mold production costs, but these decrease as production volume increases, the cost per assembly decreases as production volume grows.
Design for Additive Manufacturing
Methodologies of DFAM cover the entire AM process, such as product requirements, redesign or part integration, structural or shape optimization, determination of AM process and parameters, and orientation of parts within the build volume [14]. In the first step of the method, functional integration of parts from the initial CAD model is performed. The goal of the optimization can be to achieve a lighter weight, a stiffer structure, dynamic properties or better heat dissipation.
Compared to the partial approach of DFAM that starts with initial CAD design that is generally designed with conventional manufacturing in mind, the authors global design approach proposed that a designer starts with constraints or requirements of the part design that AM as a considered means of manufacture. A new global DFAM methodology was proposed with a case study performed on a redesign of the robotic system.
Topology Optimization (TO) for DFAM
To realize the flexible wing that can be swept, the heterogeneous mechanical property is necessary for the skin of the wing. By using SLM process with TO that enabled the generation of the structure with specific stiffness, the authors were able to produce parts with desired properties with consistency. The authors implemented TO to maximize the permeability of the structure while limiting the stiffness to the level of bones.
They were able to manufacture the scaffold model using the 3DP process and validate the performance of the structure. The second phase of the optimization is performed only for the remaining stiffeners after deleting infeasible stiffeners so that the calculation amount could be reduced.
Example of DFAM – Case Study
Considering the analysis results, a designer can choose between modifying the optimization problem and manually modifying the design. Often this is called data-driven design process because the design of the structure is driven from the requirements and constraints of the design problem with the optimization model. Together with the global stiffness matrix K, the KU=F represents the governing equation of the entire structure.
In this study, one of the powder bed fusion technologies, selective laser sintering (SLS), was used to fabricate the frame structure. Designation of the design room for the design of the electric bicycle followed the requirements described below. Third, the frame should be sectioned later for the SLS manufacturing process, because the bed volume of the SLS system was mm, which is smaller than the total size of the bicycle frame.
A groove was made to firmly hold the battery in the hollow space of the frame. With the SLS machine used to make the frame parts, the maximum size of the part that can be printed is determined by the powder size. The DFAM concept was implemented in the design process of the bicycle frame and the entire manufacturing process was carried out.
By applying TO in the design process and FEA for the evaluation of the prototype model, the design and modification cycle was completed in a short period of time. Since the battery pack and motor compartment make up the majority of the total weight, it is expected that the weight of the frame part will be further reduced. It is difficult to conclude that the commercialization of the electric bicycle produced with the SLS process is feasible.
Drawbacks of Topology Optimization Method
Lattice Structure Generation Algorithm
In order to simulate the stress conditions that will be applied to the cell when it is stacked, normal and shear stress was applied to the surfaces (right, left, top and bottom) of the cell as illustrated in Figure 3.2.b. In the building block stacking phase, a finite element analysis was performed to the design space of the given problem using ABAQUS. Second, the interior space of the block must be adaptable to various stress conditions.
To alleviate the concentration and the resulting problem caused by the area, internal corners of the building block were rounded. These stress ratios were determined to cover the stress responses of the actual structures. The purpose of the building block stacking phase is to fill the grids with proper building blocks that can withstand stress states induced by loads and constraints of the problem.
When the element size is infinitely reduced, the convergence of the stress value with the analytical solution is predicted. The second condition of Equation 3.7, 𝝈𝑮𝒔, 𝑪⋅ 𝝈𝑩𝒊 Using the above conditions, building blocks that can withstand the stress state of the design problem are selected to populate the meshes.
As can be seen in the figure, the actual generation process starts with the structural analysis of the design space. The building blocks optimized with the larger stress state (component-by-component, with the same sign) are selected for the generation of the lattice structure. The assembly of the building blocks (which are .stl files) was also implemented in MATLAB code.
The Result of the Structure Generation
To generate a grid structure with the presented method, the design space must be gridded by the size of the building block. As described in the previous section, the structural analysis data was used to obtain the stress data in the grids, as illustrated in Figure 3.10. The structure was voxelized using MATLAB code, and static analysis was performed using ABAQUS software, as illustrated in Figure 3.12.
To compare the result with the general TO procedure, the design space was optimized using the compliance minimization method with the option of 50% volume restriction under the same supports and load conditions using the TO solidThinking® Inspire 2018 software (Figure 3.13). The estimated structure weight was calculated based on the voxels that make up the structures. As in the previous example, the structural analysis data of the design space (Figure 3.16) were used to calculate the average stress in each grid and select the appropriate building block.
To compare the result with a general TO process, the design space was optimized with the same conditions used previously (Figure 3.19). However, the time spent generating the structure can again be dramatically reduced. In Figure 3.3 (also in the appendix for larger images), the structures are grouped with the same x-direction normal stresses.
In Figure 3.21, there are two groups of structures with x-direction normal stresses that are +2 and -2. General state of stress with stress represented in Figure 3.22.a can be redrawn with the Mohr's circle representation shown in Figure 3.22.b. In this case, two Mohr's circles equidistant from the vertical axis according to the stress conditions are drawn as the right side, and the left side of the horizontal axis represents tensile and compressive stress, respectively, as illustrated in Figure 3.24.
Conclusion and Contribution
Future Research
In the method, the correct building blocks are determined by simple comparison of the stress state of the TO that the building block has passed through and the average stress on the grid for the design problem. The resulting structures show decent performance on the outcome, but do not guarantee that the chosen building blocks are the best combination of the blocks for the design problem. Further research requires research into the search for the best combination of the building blocks.
Since the method directly uses the stress distribution data from the structural analysis of the design space that varies with the structure, the optimization of the structure in the stress distribution is not accurate. The multi-step method in structure generation can be applied to further improve the method. Campbell, T., et al., Can 3D printing change the world?: technologies, potential and implications of additive manufacturing.
Conner, B.P., et al., Making sense of 3-D printing: Creating a map of additive manufacturing products and services. Wang, C., et al., Concurrent topology optimization design of structures and non-uniform parametrized mesh microstructures. Moon, S.K., et al., Application of 3D printing technology to design lightweight unmanned aircraft wing structures.
Lin, C.Y., et al., Structural and mechanical evaluations of a topologically optimized titanium fusion cage fabricated by a selective laser melting process. Wang, B., et al., Mechanical behavior of carbon fiber reinforced pyramid lattice core sandwich structures. In the loads panel in the inspire software, you can export the load cases to get the load cases in .csv format.
The details of the load cases can be modified by replacing the values in the file. At the prompt, users can use the odbAccess library to access the output database file (.odb file) and retrieve the structural analysis result from ABAQUS.
