Data Driven Computing
Author: Kirchdoerfer, Trenton Thomas
Year: 2018
Degree: Dissertation (Ph.D.)
Advisor: Ortiz, Michael
Committee Members: Lapusta, Nadia; Asimaki, Domniki; Kochmann, Dennis M.; Ortiz, Michael
Option: Aeronautics
DOI: 10.7907/Z9Z899MV
Abstract
Data Driven Computing is a new field of computational analysis which uses provided data to directly produce predictive outcomes. This thesis first establishes definitions of Data-Driven solvers and working examples of static mechanics problems to demonstrate efficacy. Significant extensions are then explored to both accommodate noisy data sets and apply the deveoloped methods to dynamic problems within mechanics. Possible method improvements discuss incorporation of data quality metrics and adaptive data sampling, while new applications focus on multi-scale analysis and the need for public databases to support constitutive data collaboration.
Files
- Kirchdoerfer_Trenton_2017_Thesis.pdf (application/pdf)