Premium
Fuzzy‐stochastic FEM–based homogenization framework for materials with polymorphic uncertainties in the microstructure
Author(s) -
Pivovarov Dmytro,
Oberleiter Thomas,
Willner Kai,
Steinmann Paul
Publication year - 2018
Publication title -
international journal for numerical methods in engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.421
H-Index - 168
eISSN - 1097-0207
pISSN - 0029-5981
DOI - 10.1002/nme.5947
Subject(s) - homogenization (climate) , finite element method , uncertainty quantification , fuzzy logic , probability density function , mathematics , probabilistic logic , propagation of uncertainty , representative elementary volume , mathematical optimization , stochastic process , computer science , algorithm , statistics , structural engineering , engineering , artificial intelligence , biology , biodiversity , ecology
Summary Uncertainties in the macroscopic response of heterogeneous materials result from two sources: the natural variability in the microstructure's geometry and the lack of sufficient knowledge regarding the microstructure. The first type of uncertainty is denoted aleatoric uncertainty and may be characterized by a known probability density function. The second type of uncertainty is denoted epistemic uncertainty. This kind of uncertainty cannot be described using probabilistic methods. Models considering both sources of uncertainties are called polymorphic. In the case of polymorphic uncertainties, some combination of stochastic methods and fuzzy arithmetic should be used. Thus, in the current work, we examine a fuzzy‐stochastic finite element method–based homogenization framework for materials with random inclusion sizes. We analyze an experimental radii distribution of inclusions and develop a stochastic representative volume element. The stochastic finite element method is used to obtain the material response in the case of random inclusion radii. Due to unavoidable noise in experimental data, an insufficient number of samples, and limited accuracy of the fitting procedure, the radii distribution density cannot be obtained exactly; thus, it is described in terms of fuzzy location and scale parameters. The influence of fuzzy input on the homogenized stress measures is analyzed.