A multiscale computational framework for ceramic composite structures with localized material nonlinearity

Mazurowski, Bryce Patrick

Permalink

https://hdl.handle.net/2142/127203 Copy

Description

Title

A multiscale computational framework for ceramic composite structures with localized material nonlinearity

Author(s)

Mazurowski, Bryce Patrick

Issue Date

2024-11-20

Director of Research (if dissertation) or Advisor (if thesis)

Duarte, Carlos A

Doctoral Committee Chair(s)

Duarte, Carlos A

Committee Member(s)

• O'Hara, Patrick
• Lambros, John
• Geubelle, Philippe
• Lopez-Pamies, Oscar

Department of Study

Civil & Environmental Eng

Discipline

Civil Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• GFEM
• CMC
• Composites
• Damage
• Contact
• Fastener
• Heterogeneous
• global-local

Abstract

Composite materials are gaining ground as the material of choice in many structural applications. These tailored materials allow engineers to meet more aggressive design goals, but they bring significant analysis challenges. Nonlinear material behavior greatly complicates homogenization methods and multiscale analysis techniques. In applications like high-speed aircraft, material nonlinearity is an expectation. The loading and environment are too extreme for materials to maintain linear behavior. This nonlinearity is, however, often localized to stress concentrations around features like connections, sharp gradient loads, or areas of high temperature. This dissertation develops a computational framework that allows the majority of an analysis domain to use linear elastic homogenized material properties and coarse, dimensionally-reduced structural models. In localized areas around stress concentrations, material heterogeneity and nonlinearity are introduced through high-fidelity models capable of capturing failure. These local areas can appear anywhere in the analysis domain, as needed. Communication across scales allows global behavior to influence local behavior and vice-versa. A Generalized Finite Element Method with global-local enrichment functions (GFEMgl) is the backbone of this method. Coarse, defeatured global models use linear elastic, homogenized material models. Auxiliary local problems capture heterogeneous, nonlinear material behavior and introduce local structural features, like fasteners. State-of-the-art material models are used to capture damage in ceramic matrix composites (CMCs) around local structural features. These material models are introduced via local problems and carried into the global analysis. Modifications to the GFEMgl allow these local problems to be introduced anywhere in the analysis domain. The GFEMgl developed in this work bridges three material length scales: global homogenized behavior, explicit laminate modeling, and intraply CMC-specific damage with the aforementioned material models. The GFEMgl framework is also included in an Iterative Global-Local method to create an IGL- GFEMgl solution procedure. Shell elements are the de facto choice for modeling aircraft at the global scale, but they do not have the resolution to capture CMC failure. The IGL- GFEMgl allows global shell models to be coupled with the proposed GFEMgl framework to capture failure of the CMC material. This creates a complete analysis loop from a global aircraft model down to the material level, where all scales provide insight and feedback to one another. This work presents a systematic study of the proposed multiscale analysis method. Relevant CMC material models are developed, calibrated, and validated on their own. Then the proposed GFEMgl is presented and verified. CMC behavior around open holes and fasteners is explored with the GFEMgl, including the use of contact models across scales. The method is consistently compared against industry-adopted multiscale methods and experiments. The IGL- GFEMgl is then used to couple global shell models, solved in a commercial software, with the GFEMgl framework to capture local CMC behavior. The computational framework is simple, efficient, and very capable when compared to competing methods.

Graduation Semester

2024-12

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/127203

Copyright and License Information

Copyright 2024 Bryce Patrick Mazurowski

Owning Collections