Date

2020-9

Author

Özkapıcı Helvacı, Damla

Metadata

Show full item record

Item Usage Stats

489
views

493
downloads

Cite This

In this thesis, a method is proposed to enhance the performance of the parts optimized by Solid Isotropic Material with Penalization (SIMP) method. SIMP is a density-based topology optimization method which basically aims to distribute the material through the part subjected to various loads, boundary conditions and constraints in the optimum way. Thus, the part satisfies an optimization goal without violating the predefined constraints. The most prominent feature of this method is that the densities of the finite elements composing the part are either 1 or 0. The metIn this thesis, a method is proposed to enhance the performance of the parts optimized by Solid Isotropic Material with Penalization (SIMP) method. SIMP is a density-based topology optimization method which basically aims to distribute the material through the part subjected to various loads, boundary conditions and constraints in the optimum way. Thus, the part satisfies an optimization goal without violating the predefined constraints. The most prominent feature of this method is that the densities of the finite elements composing the part are either 1 or 0. The method penalizes the intermediate densities to converge to 1 or 0. One of the reasons behind the penalization is to provide a manufacturable geometry since intermediate density regions are difficult to fabricate. However, inclusion of these regions in the final topology may enhance the performance of the part. Based on this idea, a method is developed to use intermediate densities and it is applied on both 2D and 3D geometries. The proposed method uses the density data obtained for each finite element as a result of the SIMP optimization. Then, the part is remodelled with rectangular cellular structures which have user defined dimensions. The area or volume of the cellular structures are proportional to the average density of covered elements. The main focus of the method is to generate lightweight, simple-shaped, and manufacturable geometries with satisfying performances. Besides, it proposes a novel technique to generate fully connected geometries, and also a technique to remove excess powder remaining inside the part after fabrication for 3D geometries in powder or resin based additive manufacturing machineries. The performance of the method is compared with the SIMP method for all geometries and other optimization techniques, such as homogenization and genetic algorithms through analysis and tests. hod penalizes the intermediate densities to converge to 1 or 0. One of the reasons behind the penalization is to provide a manufacturable geometry since intermediate density regions are difficult to fabricate. However, inclusion of these regions in the final topology may enhance the performance of the part. Based on this idea, a method is developed to use intermediate densities and it is applied on both 2D and 3D geometries. The proposed method uses the density data obtained for each finite element as a result of the SIMP optimization. Then, the part is remodelled with rectangular cellular structures which have user defined dimensions. The area or volume of the cellular structures are proportional to the average density of covered elements. The main focus of the method is to generate lightweight, simple-shaped, and manufacturable geometries with satisfying performances. Besides, it proposes a novel technique to generate fully connected geometries, and also a technique to remove excess powder remaining inside the part after fabrication for 3D geometries in powder or resin based additive manufacturing machineries. The performance of the method is compared with the SIMP method for all geometries and other optimization techniques, such as homogenization and genetic algorithms through analysis and tests.

Subject Keywords

Additive Manufacturing, Topology Optimization, Simp, Cellular Structures, Connectivity Analysis

URI

https://hdl.handle.net/11511/69063

Collections

Graduate School of Natural and Applied Sciences, Thesis