Identification of Model Parameters for the Simulation of SMA Structures Using Full Field Measurements

With the design of new complex devices in shape memory alloys (SMAs) and to take advantage of their large recoverable strains, SMA components are increasingly subjected to multiaxial loadings. To analyse the behaviour of such components, numerical models should include the dependence of the loading path, e.g tension-compression asymmetry. The problem of the identification of such parameters remains crucial, since standard characterization of these alloys is obtained using uniaxial tensile tests. Additional tests are required to determine the parameters characteristic of the loading path dependence.

In this work, the model parameters are determined from multiaxial and heterogeneous tests carried out on specimens with the same thermomechanical loading history. Digital Image Correlation technique is employed to obtain the strain fields at the surface of the specimen. Finite Element Analysis provides the numerically evaluated strain fields using a thermodynamical constitutive model developed in a previous work, which takes into account the tension-compression asymmetry. The strain fields computed numerically are compared with experimental ones obtained by DIC to find the model parameters which best matches experimental measurements. The procedure to obtain these material parameters is based on an inverse identification procedure, which consists on the minimization of the cost function built in terms of the strain components on the surface of the specimen. Model parameters for the simulation of SMA structures are thus obtained based on a few multiaxial and heterogeneous tests at different temperatures. Validation of the identification procedure is based on multiaxial tests in different configurations.