Reliability confidence intervals for ceramic components as obtained from bootstrap methods and neural networks

Fracture of ceramic components is governed by the multiaxial Weibull theory. Due to the inherent scatter in fracture behaviour of ceramics, strength and lifetime predictions are prone to statistical uncertainty. In this paper we present a strategy for the assessment of uncertainty in the estimation of the failure probability of ceramic components due to the scatter of material data. Confidence intervals for the failure probability are obtained by means of stochastic resampling methods. A considerable computational effort can be saved when using neural networks for part of the simulation. The strategy presented in this paper may be applied for different purposes: estimation of the reliability of a ceramic component, setup of parameters for an experimental investigation, and interpretation of experimental results.