A Multiscale Analysis on the Superelasticity Behavior of Architected Shape Memory Alloy Materials

Abstract

In this paper, the superelasticity effects of architected shape memory alloys (SMAs) are focused on by using a multiscale approach. Firstly, a parametric analysis at the cellular level with a series of representative volume elements (RVEs) is carried out to predict the relations between the void fraction, the total stiffness, the hysteresis effect and the mass of the SMAs. The superelasticity effects of the architected SMAs are modeled by the thermomechanical constitutive model proposed by Chemisky et al. 2011. Secondly, the structural responses of the architected SMAs are studied by the multilevel finite element method (FE2), which uses the effective constitutive behavior of the RVE to represent the behavior of the macroscopic structure. This approach can truly couple the responses of both the RVE level and structural level by the real-time information interactions between two levels. Through a three point bending test, it is observed that the structure inherits the strong nonlinear responses—both the hysteresis effect and the superelasticity—of the architected SMAs at the cellular level. Furthermore, the influence of the void fraction at the RVE level to the materials’ structural responses can be more specifically and directly described, instead of using an RVE to predict at the microscopic level. Thus, this work could be referred to for optimizing the stiffness, the hysteresis effect and the mass of architected SMA structures and extended for possible advanced applications.

Multiscale Analysis of Composite Structures with Artificial Neural Network Support for Micromodel Stress Determination

“…Modern heterogeneous materials like fiber-reinforced [ 2 ] or multimetallic composites [ 3 ] are gaining more and more popularity in engineering structures due to their application-specific tailored properties. To accurately describe their microstructural behavior, numerical methods such as the multiscale finite element method [ 4 , 5 , 6 , 7 ], asymptotic homogenization method [ 8 , 9 , 10 ], coarse-graining technique [ 11 ], and finite element and Fast Fourier Transforms method [ 12 ] have been developed. High-fidelity microstructure material modeling is essential for macroscale structural analysis prediction.…”