Uncertainty Quantification for Buckypaper Polymer Composite Computer Simulation

• Other Authors: Chen, Li-Jen (authoraut)
• Format: Others
• Language: English; English
• Published: Florida State University
• Subjects: Engineering; Manufacturing processes
• Online Access: http://purl.flvc.org/fsu/fd/FSU_migr_etd-3838

Since the discovery of carbon nanotubes in 1991 alongside their superior performance in mechanical and electrical properties, carbon nanotubes have been widely considered to be one of the most promising next generation materials. They have been frequently used in polymer composites due to their high strength-to-weight and modulus-to-weight ratios. Yet despite their promising qualities in manufacturing, carbon nanotube based composites still have many issues that need to be resolved before they can be used for industrial applications. In order to more cost effectively produce nanocomposites and improve their quality, it is necessary to accurately observe and understand the variations in their raw material properties. The variability of the raw material in nanotube based composites usually has a large impact on the properties of the eventual product. However, physical experimentation for the purpose of quantifying variability in nanomaterial properties is usually expensive and sometimes not feasible or accurate enough. This paper presents a constrained nonlinear programming approach for the quantification of raw material variability while also examining the impact of raw material variability on the properties of buckypaper polymer (BPP) composites. The proposed approach suggests conducting small physical experiments to collect data on raw material properties and final composite part properties before employing an inverse uncertainty propagation approach to estimate the parameters of the probability distribution of the material properties. Both univariate and multivariate probability distributions are considered. A case study based on data from a real buckypaper manufacturing process is used to illustrate the approach. It is shown that simultaneously modeling the material properties with a multivariate distribution improves the quality of the identified model. === A Thesis submitted to the Department of Industrial and Manufacturing Engineering in partial fulfillment of the requirements for the degree of Master of Science. === Summer Semester, 2010. === June 11, 2010. === Uncertainty Quantification, Monte Carlo Simulation, Nonlinear Programming, Nanocomposites === Includes bibliographical references. === Arda Vanli, Professor Directing Thesis; Joseph Pignatiello, Committee Member; Chad Zeng, Committee Member; Ben Wang, Committee Member; Chuck Zhang, Committee Member.