Computational design of multiaxial tests for anisotropic material characterization

Although anisotropic materials provide more capabilities for mission- and application-tailored design and functional flexibility to final structures than regular isotropic materials, the directional behavior of the anisotropic materials further complicates their inelastic and damage behavior. Such a non-linear behavior can be effectively observed and characterized by multiaxial testing, but how to design a multiaxial test for material characterization given a specimen remains an untouched issue. This paper presents a methodology that numerically designs the loading path of a multiaxial testing machine to characterize anisotropic materials. The multiaxial test must be able to exhibit quantities used to characterize materials as distinctly as possible. The proposed methodology formulates distinguishability and uniqueness as such quantities by first analyzing the specimen on a continuum basis with finite element method and then applying singular value decomposition. Associating the distinguishability and uniqueness with the informativeness of the loading path, the design problem is formulated such that an effective loading path can be found efficiently by a standard optimization method. Numerical examples first investigate the validity of the distinguishability and the uniqueness as performance measures to evaluate loading paths. The efficacy of the proposed methodology has been then confirmed by analyzing it with and applying it to design problems. Copyright (c) 2007 John Wiley & Sons, Ltd.