The calibration of material properties for use in discrete element models

[摘要] ENGLISH ABSTRACT: One of the main challenges in using the Discrete Element Method (DEM) is tospecify the correct input parameter values. In general, the models are sensitive tothe input parameter values and accurate results can only be achieved if the correctvalues are specified. For the linear contact model, micro parameters such as theparticle density, stiffness, coefficient of friction, as well as the particle size andshape distributions are required. Thus, there is a need for a procedure toaccurately calibrate these parameters before any attempt can be made toaccurately model a complete bulk materials handling system.Since the DEM is often used to model applications in the mining and quarryingindustries, a calibration procedure was developed for materials that consist ofrelatively large (up to 40 mm in size) particles. A coarse crushed aggregate wasused as the test material. Using a specially designed large scale shear box, theconfined Young's Modulus and internal friction angle of the material weremeasured by means of the confined compression test and the direct shear testrespectively. The bulk (macro) density and porosity were also measured. Theparticle size distribution was measured while visual inspection was used toidentify the different particle shapes.DEM models of the experimental set-up were developed and the input parametervalues were varied iteratively until a close correlation between the experimentaland numerical results was achieved. The resulting set of input parameter valueswere then verified through a series of anchor pull-out and angle of reposeexperiments and simulations. A good correlation between the experimental andnumerical results was observed.In a study, independent of the calibration process, a half fraction factorial designwas implemented to quantify the effect of the input parameter values on the bulkproperties and to construct multiple linear regression models that relate theparameters to the bulk properties. The results were found to be in accordance withexpected bulk behaviour, and can be used to develop advanced DEM calibrationstrategies.Based on the project outcomes, it was concluded that the developed calibrationprocedure performed satisfactorily and that the calibrated input parameters allowfor the accurate modelling of the coarse aggregate.

[发布机构] Stellenbosch University