Fuzzy modeling of imprecise material parameters in the analysis of metal matrix composite structures

Engineering information is often uncertain, deviated, imperfect, erroneous, imprecise, etc. Fuzzy methods, based on possibility theory are well suited for describing epistemic uncertainty. Included in epistemic uncertainty are data sets where only few samples are available and are appropriately modeled with fuzzy methods. The resulting fuzzy parameters can then be used in structural analysis, allowing for nondeterministic consideration and, therefore, uncertainty assessment. In this study, different approaches to construct fuzzy membership functions are discussed and compared. Further, the influence of fuzzy modeling, or fuzzification, is evaluated within the framework of structural analysis and then in the safety factor assessment. The mechanical response was calculated using finite element analysis with uncertain material, evaluating stress and buckling load factor. Finally, the fuzzy safety factors were quantified for the different fuzzification methods, showing the importance of uncertainty consideration.