Error Analysis of Displacement Gradients via Finite Element Approximation of Digital Image Correlation System

In the field of experimental solid mechanics, conventional strain measurement devices such as LVDT and strain gauges provide mean values of strains and displacements at selected locations and gauge lengths; this result is inadequate for the evaluation of a non homogeneous material behaviour. Consequently, during the last decades various full-field non-contact measurement techniques have been proposed for the material characterization and have become more and more popular in the experimental mechanics community. In this work, the Digital Image Correlation (DIC) non-interferometric technique has been used to monitor experiments on aluminum flat bars and to measure displacement distribution on the surface of the specimen for further evaluation and calculation of strains. The results obtained by a Digital Image Correlation System are assessed and the error associated with the post processing of the experimental field data, obtained through the use of the Aramis software, is evaluated and analyzed with the aid of a least square approximation code. This code uses a finite element approximation of the displacement field in order to cover all the target points. A least square approximation of these data is performed and the best nodal displacement values are determined. Based on the nodal data, infinitesimal and finite strain distributions are determined over the surface image window of the specimen. It is observed that this post processing technique provides better results near perforations and edges that are not sensitive to the density of the captured displacement data.