Multi-Scale Modeling With Machine Learning and Uncertainty Quantification for Predicting Fatigue Crack Evolution in Titanium Alloys: Part II

2021 ◽  
Author(s):  
Somnath Ghosh
Keyword(s):  
Machine Learning ◽  
Fatigue Crack ◽  
Titanium Alloys ◽  
Uncertainty Quantification ◽  
Crack Evolution ◽  
Multi Scale ◽  
Scale Modeling ◽  
Multi Scale Modeling

Integrating an Analytical Uncertainty Quantification Approach to Multi-Scale Modeling of Nanocomposites

2019 ◽  
pp. 1-22
Author(s):  
Pinar Acar
Keyword(s):  
Uncertainty Quantification ◽  
Uncertainty Propagation ◽  
Epoxy Nanocomposites ◽  
Linear Solution ◽  
Analytical Uncertainty ◽  
Multi Scale ◽  
Scale Modeling ◽  
Macro Scale ◽  
Multi Scale Modeling ◽  
Scale Properties

Abstract The present study addresses the integration of an analytical uncertainty quantification approach to multi-scale modeling of single-walled carbon nanotube (SWNT)-epoxy nanocomposites. The main highlight is the investigation of the stochasticity of nanotube orientations, and its effects on the homogenized properties. Even though the properties of SWNT-epoxy nanocomposites are well-studied in the literature, the natural stochasticity that arises from the nanotube orientations has not been observed. To understand the effects of the variability in SWNT orientations to material properties of interest, an analytical uncertainty quantification algorithm is utilized. The analytical scheme computes the propagation of the orientational uncertainty to the volume-averaged properties with a linear solution and uses the transformation of random variables principle to obtain the variations in non-linear properties. The results indicate that the uncertainty propagation affects the macro-scale properties, including stiffness, thermal expansion, thermal conductivity, and natural frequencies.

scholarly journals Machine Learning Coupled Multi‐Scale Modeling for Redox Flow Batteries

2019 ◽  
Vol 3 (2) ◽  
pp. 1900167 ◽  
Author(s):  
Jie Bao ◽  
Vijayakumar Murugesan ◽  
Carl Justin Kamp ◽  
Yuyan Shao ◽  
Litao Yan ◽  
...  
Keyword(s):  
Machine Learning ◽  
Multi Scale ◽  
Redox Flow Batteries ◽  
Scale Modeling ◽  
Flow Batteries ◽  
Multi Scale Modeling

scholarly journals Redox Flow Batteries: Machine Learning Coupled Multi‐Scale Modeling for Redox Flow Batteries (Adv. Theory Simul. 2/2020)

2020 ◽  
Vol 3 (2) ◽  
pp. 2070004
Author(s):  
Jie Bao ◽  
Vijayakumar Murugesan ◽  
Carl Justin Kamp ◽  
Yuyan Shao ◽  
Litao Yan ◽  
...  
Keyword(s):  
Machine Learning ◽  
Multi Scale ◽  
Redox Flow Batteries ◽  
Scale Modeling ◽  
Flow Batteries ◽  
Multi Scale Modeling

Integrating an Analytical Uncertainty Quantification Approach to Multi-Scale Modeling of Nanocomposites

2018 ◽  
Author(s):  
Pınar Acar
Keyword(s):  
Uncertainty Quantification ◽  
Single Walled Carbon Nanotube ◽  
Analytical Uncertainty ◽  
Multi Scale ◽  
Scale Modeling ◽  
Macro Scale ◽  
Multi Scale Modeling ◽  
Scale Properties ◽  
Dynamics Simulations ◽  
Growth Approach

This work addresses the integration of an analytical uncertainty quantification approach to multi-scale modeling of single-walled carbon nanotube (SWNT)-epoxy nanocomposites consisting of pristine systems. The computational modeling starts with the dendrimer growth approach, which is used to build an epoxy-SWNT network. Next, the molecular dynamics simulations are performed to obtain thermal and mechanical properties. The SWNT orientations are assumed to have natural stochasticity which is modeled by an analytical uncertainty algorithm. Next, the propagation of the uncertainties to the volume-averaged properties of the SWNT and nanocomposite is obtained. The uncertainties are shown to affect the macro-scale properties such as stiffness, thermal expansion, thermal conductivity and natural frequencies.

Sign in / Sign up

Export Citation Format

Share Document