Two methods are used in this work to estimate the porous shape memory alloy
(SMA) thermomechanical behavior. The porous SMA is assumed to be made of t
wo components, the dense SMA matrix and the pores. An existing rate-indepen
dent type constitutive model is employed to describe the matrix behavior. T
wo contrasting strategies are used to estimate the overall thermomechanical
behavior: (1) the unit cell finite element method (UCFEM) to account for p
eriodic distribution of pores in the SMA matrix, and (2) an averaging micro
mechanics method based on the incremental formulation of the Mori-Tanaka me
thod to account for random distribution of pores in the matrix. Cylindrical
and spherical shapes are considered as approximations of open and closed p
ores, respectively, in both methods. Results are presented for both types o
f pores and comparisons are made between the two methods under various load
ing conditions. Both methods compare well in predicting the isothermal elas
tic material properties and pseudoelastic response under axial and out-of-p
lane shear loading. However, the transformation results differ under transv
erse and in-plane shear loading. This difference is found to be due to the
use of an average value of stress for the SMA matrix in the micromechanics
averaging method, which diminishes the effect of local stress concentration
thereby delaying the onset of phase transformation caused by an applied lo
ad. On the other hand, the actual values of stress at all material points i
n the SMA matrix are used in the UCFEM causing phase transformation in regi
ons near the pores at smaller applied load values than what is calculated b
y the micromechanics averaging method. (C) 2001 Published by Elsevier Scien
ce Ltd.