The Woodruff School

SUMMARY

Functionally Graded Lattice (FGL) Structures have shown improved performance over uniform lattice structures in different fields. These structures contain unit-cells of varying porosity based on different functional requirements, which alters the properties of the structures. Another form of functional grading can be seen in materials in nature, where the cellular structure can vary in both porosity and size. Therefore, to distinguish between lattice structures that vary in porosity only and lattice structures that vary in both, we will refer to the latter in this research as Naturally Functionally Graded Lattice (NFGL) structures. However, research into NFGL structures' performance against FGL structures in the literature is lacking. Furthermore, the current methods in the literature to generate these structures are severely limited and suffer from multiple drawbacks, such as being computationally expensive, generate non-conformal lattice structure, stochastic in structure, limited in their ability to vary the unit-cell size ratios, and other drawbacks. To address these issues, this research aims to develop a framework, namely the NFGL Framework, to generate NFGL structures without the drawbacks that exist in current methods and to improve the performance of the generated structures using the NFGL framework against existing FGL structures. The NFGL Framework uses the nodes of a finite element mesh that conforms to the design domain and a density field input of the domain in a developed simplified sphere backing algorithm to generate the NFGL structure nodes, which are then connected using Delaunay Triangulations. Furthermore, the NFGL Framework can perform a similarity analysis using a modified Mean Structural Similarity (MSSIM) index to improve the performance of the generated NFGL structure. The generated structures using the NFGL Framework were tested against the existing methods and showed to overcome the drawbacks of these methods with improved performance and computational time. Furthermore, the generated NFGL structures were tested against FGL structures and the results showed a performance gain from the use of NFGL structures over FGL structures with a reduced computational cost. https://gatech.webex.com/gatech/j.php?MTID=mafa254e3774fe19a15872c3511b0b76f

SUBJECT:	Ph.D. Dissertation Defense

BY:	Mahmoud Alzahrani

TIME:	Friday, July 3, 2020, 2:00 p.m.

PLACE:	WebEx Virtual Presentation,

TITLE:	Framework for The Generation and Design of Naturally Functionally Graded Lattice Structures

COMMITTEE:	Dr. Seung-Kyum Choi, Chair (ME) Dr. David W. Rosen (ME) Dr. Steven Y. Liang (ME) Dr. Jarek Rossignac (CC) Dr. Graeme J. Kennedy (AE)

SUMMARY Functionally Graded Lattice (FGL) Structures have shown improved performance over uniform lattice structures in different fields. These structures contain unit-cells of varying porosity based on different functional requirements, which alters the properties of the structures. Another form of functional grading can be seen in materials in nature, where the cellular structure can vary in both porosity and size. Therefore, to distinguish between lattice structures that vary in porosity only and lattice structures that vary in both, we will refer to the latter in this research as Naturally Functionally Graded Lattice (NFGL) structures. However, research into NFGL structures' performance against FGL structures in the literature is lacking. Furthermore, the current methods in the literature to generate these structures are severely limited and suffer from multiple drawbacks, such as being computationally expensive, generate non-conformal lattice structure, stochastic in structure, limited in their ability to vary the unit-cell size ratios, and other drawbacks. To address these issues, this research aims to develop a framework, namely the NFGL Framework, to generate NFGL structures without the drawbacks that exist in current methods and to improve the performance of the generated structures using the NFGL framework against existing FGL structures. The NFGL Framework uses the nodes of a finite element mesh that conforms to the design domain and a density field input of the domain in a developed simplified sphere backing algorithm to generate the NFGL structure nodes, which are then connected using Delaunay Triangulations. Furthermore, the NFGL Framework can perform a similarity analysis using a modified Mean Structural Similarity (MSSIM) index to improve the performance of the generated NFGL structure. The generated structures using the NFGL Framework were tested against the existing methods and showed to overcome the drawbacks of these methods with improved performance and computational time. Furthermore, the generated NFGL structures were tested against FGL structures and the results showed a performance gain from the use of NFGL structures over FGL structures with a reduced computational cost. https://gatech.webex.com/gatech/j.php?MTID=mafa254e3774fe19a15872c3511b0b76f