Ph.D. in Mechanical Engineering
Yong Hoon Lee received the Degree of Doctor of Philosophy in Mechanical Engineering.
- Date: August 10, 2020
- Dissertation: Methods for the integrated design of viscoelastic materials and structural geometry
- Doctoral Committee:
- Dr. Randy H. Ewoldt, Chair
- Dr. James T. Allison, Director of Research
- Dr. Pingfeng Wang
- Dr. Mattia Gazzola
Table of Contents
PART I. Methodologies for Surface Texture Design
Arbitrary Surface Texture Design Enhances Full-Film Lubrication Performance
Efficient Design of Textured Surface Lubrication Using Linearization
Efficient Sampling and Implicit Constraint Management Methods for Multiobjective Adaptive Surrogate-Based Optimization
PART II. Integrated Design With Viscoelasticity
Numerical Design With Viscoelasticity: Survey, Efficient Strategies, and Outlook
Continuous Relaxation Spectra Description for Design With Linear Viscoelasticity
Simultaneous Design of Non-Newtonian Lubricant and Surface Texture Using Surrogate-Based Multiobjective Optimization
- Conclusions and Outlook
Design engineers have traditionally selected simple materials, including linear elastic solids and Newtonian fluids, for their designs, and optimize designed systems within the limitations of preselected materials. However, allowing complex materials (e.g., viscoelastic materials and soft solids) as options for system materials can eliminate some associated limitations on design freedom. Furthermore, designing material characteristics may open up avenues to unprecedented designs with optimal materials. This dissertation research presents integrated design methodologies for the system-level performance-driven design of structural geometry and viscoelastic material properties. By relaxing unnecessary assumptions and optimizing system-level performance by searching the rich spectrum of possible design options throughout the combined structural and material design space, the studies in this dissertation seek to obtain design improvements, which previously were inaccessible using conventional design paradigms and assumptions.
Achieving this integrated design of structural geometry and viscoelastic material properties has many underlying challenges. One of the main challenges is how to formulate design representations that can efficiently and effectively model the system behavior within the limits of physics, but are not restrictive with respect to other factors to allow full design exploration freedom. Another challenge comes from the numerical implementations and limits on computational resources available to design engineers. Higher-fidelity models or models with complex behavior (e.g., viscoelastic stress relaxation) are generally more expensive to solve. However, the computational cost of numerical models is a crucial factor in the success of integrated and multidisciplinary design optimization. Not all models are design-appropriate. Specifically, quantities defining viscoelastic constitutive relations are interrelated to each other, and not all are independently controllable. Such challenges obstruct design with and of complex materials and integrated design associated with these material models. Thus, identifying underlying challenges and overcoming them are the main purposes of this dissertation research.
This dissertation is organized into two parts. The chapters in Part I present methodologies for designing surface texture geometries and other challenging design problems. Studies in this part present surface texture parameterizations for enhancing the lubricated frictional performance indices by eliminating unnecessary design constraints (Chapter 2), high-dimensional design optimization of surfaces texture using linearization, sequential linear programming, and trust-region methods (Chapter 3), and novel efficient sampling and implicit constraint generation methods for multiobjective surrogate-based optimization for challenging problems (Chapter 4). The chapters in Part II present integrated design studies and methodologies that apply to the design problems of viscoelastic material systems. Studies in this part present numerical methods for efficiently solving viscoelastic stress relaxation with convolution integrals (Chapter 5), a design-appropriate continuous stress relaxation spectra design representation and its use in the linear viscoelastic design problem (Chapter 6), and method for a simultaneous design of surface texture and non-Newtonian lubricant properties using nonlinear viscoelastic models to achieve better tribological performance beyond the results of the study presented in the earlier chapter that demonstrated surface texture design (Chapter 7).
The methodologies and design studies presented in this dissertation have an impact on exploring previously-unexplored design spaces from an integrated structural and material systems design perspective. The contributions presented in this dissertation open up new possibilities for design engineers to better use complex materials in the system-level performance-driven design studies incorporating a new material design paradigm.