RT Journal Article ID 3d99629959fcf021 A1 Bogdanor, Michael J. A1 Mahadevan, Sankaran A1 Oskay, Caglar T1 UNCERTAINTY QUANTIFICATION IN DAMAGE MODELING OF HETEROGENEOUS MATERIALS JF International Journal for Multiscale Computational Engineering JO JMC YR 2013 FD 2013-05-02 VO 11 IS 3 SP 289 OP 307 K1 composite materials K1 Bayesian calibration K1 rate-dependent damage K1 Gaussian processsurrogate model AB This manuscript investigates the use of Bayesian statistical methods for calibration and uncertainty quantification in rate-dependent damage modeling of composite materials. The epistemic and aleatory uncertainties inherent in the model prediction due to model parameter uncertainty, model form error, solution approximations, and measurement errors are investigated. Gaussian process surrogate models are developed to replace expensive finite element models in the analysis. A viscous damage model is employed with a solution algorithm designed for implementation within a commercial finite element software package (Abaqus). Experimental results from a suite of monotonic load tests conducted on unidirectional glass fiber reinforced epoxy composite samples at multiple strain rates and strain orientations are used to quantify the uncertainty in the prediction of the composite response within a Bayesian framework. PB Begell House LK https://www.dl.begellhouse.com/journals/61fd1b191cf7e96f,4c5b137e22f4b830,3d99629959fcf021.html