Selecting a suitable time step for discrete element simulations that use the central difference time integration scheme

摘要

The distinct element method as proposed by Cundall and Strack uses the computationally efficient, explicit, central difference time integration scheme. A Emitation of this scheme is that it is only conditionally stable, so small time steps must be used. Some researchers have proposed using an implicit time integration scheme to avoid the stability issues arising from the explicit time integrator typically used in these simulations. However, these schemes are computationally expensive and can require a significant number of iterations to form the stiffness matrix that is compatible with the contact state at the end of each time step. In this paper, a new, simple approach for cakulating the critical time increment in explicit discrete ekment simulations is proposed. Using this approach, it is shown that the critical time increment is a function of the current contact conditions. Considering both two- and three-dimensional scenarios, the proposed refined estimates of the critical time step indicate that the earlier recommendations contained in the literature can be unconservative, in that they often overestimate the actual critical time step. A three-dimensional simulation of a problem with a known analytical solution illustrates the potential for erroneous results to be obtained from discrete ekment simulations, if the time-increment exceeds the critical time step for stable analysis.