Reliability analysis in an unloading model of silo storage by means of the Discrete Element Method DEM

Abstract

Modelling based upon the discrete element method has opened the possibility of simulating different processes, which under the paradigm of traditional continuum mechanics, had not been duly modelled due to their granular nature. Many fields have benefited from the headway made in recent years in the modelling of granular materials with discrete elements: agricultural engineering that requires to design grain storages; geotechnical engineering that needs to deepen the understanding of geomechanical processes; the mining industry which needs to optimise the design of cut slopes for open pits; and the pharmaceutic industry among others. Despite the recent developments in the discrete element modelling, the method itself has associated an epistemic and a random uncertainty. In this case, the random uncertainty within DEM can be linked to the constitutive models. The review of specialised literature reveals few attempts to address the issue of assessing the uncertainty within simulations using DEM. On the other hand, many works have been published since the 80's about uncertainty within finite element modelling (finite element stochastics). This article presents a first approach to assess uncertainty in DEM modelling. A procedure to assess the random uncertainty in the DEM simulation of silos storing corn is proposed. FOSM (First Order Second Moment) method is used to approximate the statistical moments of the performance function (velocity of discharge), using the central finite difference technique. The obtained results allow to foresee the feasibility of using the proposed procedure to assess the probability of a non-clogged operation of the hopper system under different hopper geometries.