Wavelet analysis of DEM simulations of samples under biaxial compression

The results of a wavelet analysis of data from discrete element modelling (DEM) simulations of samples under biaxial compression are presented. We show how a wavelet technique may be used to find the strain scales on which critical events occur and to identify regions both in space and in strain when particles in the sample undergo significant activity. The wavelet analysis indicates that most activity occurs along a line, and this line coincides with a localization or shear band that develops in the specimen during compression. The location of this shear band can be visually identified by considering the cumulative particle rotation. Furthermore, using cross-correlation we show that the principal stress ratio is correlated with the porosity of the sample along this line. In order to investigate the robustness of the technique, the wavelet analysis is carried out on two different size specimens that both show the same general phenomena.