Chair of Engineering Mechanics Ifm-publication 2-426 the Need for Linking Micromechanics of Materials with Stochastic Finite Elements: a Challenge for Materials Science

A basic assumption of Stochastic Finite Element Methods (SFEM) is that the spatial variation of uncertain structural material properties can be described by random fields. Due to the lack of measured data, SFEM researchers and users are usually forced to make assumptions regarding qualitative and quantitative characteristics for the random field representation of various material parameters involved in simulation runs, which are either not at all or insufficiently justified. According to a promising approach towards the alleviation of this problem, information about macroscopic material randomness can be deduced by examining the micromechanical characteristics of materials. The major aim of this work is to stimulate a systematic and multidisciplinary effort towards establishing a link between microstructural material models and stochastic material properties at the macroscale. In this respect, the macroscopic random field material properties requiring micromechanically derived stochastic information are pointed out and specific questions are posed regarding the validity of assumed stochastic characteristics and parameter Preprint submitted to Computational Materials Science 6 December 2006 values. Moreover, the relative influence of various stochastic material properties on SFEM results is numerically evaluated, in order to identify the most critical of these properties needing closer and more urgent attention. The SFEM community would largely benefit from micromechanical information gathered for the spatial variation characteristics of various structural material types. This need for experimental data poses a challenge to material scientists, who can make a valuable contribution in the understanding of unexplored issues affecting the application of sophisticated SFEM formulations that are being developed for decades.