A MULTISCALE FRAMEWORK FOR THE STOCHASTIC ASSIMILATION AND MODELING OF UNCERTAINTY ASSOCIATED NCF COMPOSITE MATERIALS

Conference · Fri Jan 01 04:00:00 EST 2016

DOI:https://doi.org/10.7712/100016.2251.12207· OSTI ID:1430233

Mehrez, Loujaine ^[1]; Ghanem, Roger; Mcauliffe, Colin; Rodgers, William R.; Aitharaju, Venkat

---

1. General Motors

Multiscale framework to construct stochastic macroscopic constitutive material models is proposed. A spectral projection approach, specifically polynomial chaos expansion, has been used to construct explicit functional relationships between the homogenized properties and input parameters from finer scales. A homogenization engine embedded in Multiscale Designer, software for composite materials, has been used for the upscaling process. The framework is demonstrated using non-crimp fabric composite materials by constructing probabilistic models of the homogenized properties of a non-crimp fabric laminate in terms of the input parameters together with the homogenized properties from finer scales.

• Research Organization: General Motors
• Sponsoring Organization: USDOE Office of Energy Efficiency and Renewable Energy (EERE), Vehicle Technologies Office (EE-3V)
• DOE Contract Number: EE0006826
• OSTI ID: 1430233
• Report Number(s): DOE-GM-CarbonFiber-ICME
• Country of Publication: United States
• Language: English

Related Subjects

multiscale homogenization
micro- meso- and macro-scales
stochastic assimilation
uncertainty propagation
NCF composite materials
polynomial chaos expansion