Probing DEM specimen heterogeneity by simulated CPT

Abstract

A standard procedure for specimen formation in DEM simulations is the radius expansion method. Initialization of stress and density is performed by means of a servo-controlled procedure in which the wall boundaries of the specimen are actuated to attain the specified target values. The limitations of this procedure for creating large specimens are here explored using simulated CPT tests. Penetration resistance at the chamber axis fails as the chamber size increases above a certain limit and the actuated radial boundary increases its distance from the axis. Detailed examination of the microscopic characteristics of the specimen shows a large localized fall on coordination number and local stress at the chamber axis that is not sensed at the external walls. Alternatives to remediate this problem are discussed.