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

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