TY - JOUR

T1 - Stochastic predictions of interfacial characteristic of polymeric nanocomposites (PNCs)

AU - Vu-Bac, N.

AU - Lahmer, T.

AU - Zhang, Y.

AU - Zhuang, Xiaoying

AU - Rabczuk, T.

N1 - Funding information: We gratefully acknowledge the support by the Deutscher Akademischer Austausch Dienst (DAAD), IRSES-MULTIFRAC and the German Research Foundation (DFG) through the Research Training Group 1462. X. Zhuang acknowledges the supports from the NSFC (41130751), National Basic Research Program of China (973 Program: 2011CB013800), Changjiang Scholars and Innovative Research Team (PCSIRT, IRT1029). and Shanghai Pujiang Program (12PJ1409100).

PY - 2013/11/23

Y1 - 2013/11/23

N2 - The effect of the single-walled carbon nanotube (SWCNT) radius, the temperature and the pulling velocity on interfacial shear stress (ISS) is studied by using the molecular dynamics (MD) simulations. Based on our MD results, the mechanical output (ISS) is best characterized by the statistical Weibull distribution. Further, we also quantify the influence of the uncertain input parameters on the predicted ISS via sensitivity analysis (SA). First, partial derivatives in the context of averaged local SA are computed. For computational efficiency, the SA is based on surrogate models (polynomial regression, moving least squares (MLS) and hybrid of quadratic polynomial and MLS regressions). Next, the elementary effects are determined on the mechanical model to identify the important parameters in the context of averaged local SA. Finally, the approaches for ranking of variables (SA based on coefficients of determination) and variance-based methods are carried out based on the surrogate model in order to quantify the global SA. All stochastic methods predict that the key parameters influencing the ISS is the SWCNT radius followed by the temperature and pulling velocity, respectively.

AB - The effect of the single-walled carbon nanotube (SWCNT) radius, the temperature and the pulling velocity on interfacial shear stress (ISS) is studied by using the molecular dynamics (MD) simulations. Based on our MD results, the mechanical output (ISS) is best characterized by the statistical Weibull distribution. Further, we also quantify the influence of the uncertain input parameters on the predicted ISS via sensitivity analysis (SA). First, partial derivatives in the context of averaged local SA are computed. For computational efficiency, the SA is based on surrogate models (polynomial regression, moving least squares (MLS) and hybrid of quadratic polynomial and MLS regressions). Next, the elementary effects are determined on the mechanical model to identify the important parameters in the context of averaged local SA. Finally, the approaches for ranking of variables (SA based on coefficients of determination) and variance-based methods are carried out based on the surrogate model in order to quantify the global SA. All stochastic methods predict that the key parameters influencing the ISS is the SWCNT radius followed by the temperature and pulling velocity, respectively.

KW - A. Nano-structures

KW - A. Polymer-matrix composites (PMCs)

KW - B. Interface/interphase

KW - C. Computational modeling

KW - Stochastic prediction

UR - http://www.scopus.com/inward/record.url?scp=84890447496&partnerID=8YFLogxK

U2 - 10.1016/j.compositesb.2013.11.014

DO - 10.1016/j.compositesb.2013.11.014

M3 - Article

AN - SCOPUS:84890447496

VL - 59

SP - 80

EP - 95

JO - Composites Part B: Engineering

JF - Composites Part B: Engineering

SN - 1359-8368

ER -