John P. Morrissey
Institute for Infrastructure and Environment,
University of Edinburgh
Supervisors: J. Y. Ooi, J.F. Chen
Introduction & Motivation
Cohesive Contact Model Details
Numerical Results
Powders and bulk solids are stored and handled in large quantities in
many industries. The materials can differ greatly in size, shape and
type but are commonly affected by handling and storage difficulties,
such as the blockage of silo outlets & transfer points or rat-holing in
silos, which all occur as a result of the development of cohesion.
The contact model is a linear/non-linear spring model for elastic-plastic
deformation.
Simple verification tests show that on a particle-particle level, the
model performs as expected (Fig. 6). As well as demonstrating a
stress history dependence, a new model must also capture the correct
failure mode.
Adhesive forces between particles in a granular material are known to
be dependent on the external forces previously exerted on the bulk
solid. As a result of this, the previous stress states of a bulk solid need
to be considered when evaluating the adhesive strength of any bulk
material.
Research Project Aims
The model includes
adhesion as a function
of plastic deformation.
Model parameters:
• f0, k1, k2, Fmin, n, x
f hys
The project started in late 2009 with DEM Solutions Ltd. and LKAB
(Sweden) as industrial collaborators for the project.
Loading Stiffness
Parameter
Constant
Pull-off Force
k1δ n
if k 2 (δ n − δ pn ) ≥ k1δ n

= k 2 (δ n − δ pn ) if k1δ n > k 2 (δ n − δ pn ) > − k adhδ n

n
if − k adhδ n ≥ k 2 (δ n − δ pn )
− k adhδ
Fig. 6 Verification of Force-Overlap
Fig. 2 Force-Overlap
The aim of this project is to develop an improved contact model that
accounts for the cohesive, stress history dependent behaviour found
in granular materials.
The material properties that affect
the cohesive strength are investigated
through experiments and are used
for calibration of the newly developed
contact model.
The effect of ingress of moisture and
fines on the handle-ability of iron ore
fines will be investigated using the newly
developed model.
Experimental
Calibration
Contact
Model
Verification
Unloading/
Reloading
Stiffness
Parameter
Load Dependant
Adhesion Parameter
Relationship for new model
Relationship in Contact Model
Fig. 7 Shear Band Prediction From Early
Planar Model
The numerical simulation shows a typical shear failure with a
conjugate pair of shear bands, which is similar to the experiment, as
can be seen in figure 3.
Calibration From Experiment
The Edinburgh Powder Tester (EPT) is a semi-automated uniaxial
tester, in which the cohesive strength of a bulk solid is evaluated from
an unconfined compression test following a period of consolidation at
a pre-defined stress.
A 3D numerical simulation of mono-sized spherical particles with a
mean diameter of 2.5mm is compared with the unconfined
compression results from the EPT.
Validation of
Model
Quantitative Prediction
Fig. 3 EPT Test: (1)Confined Consolidation (2)Unconfined Sample (3)Crushing to Failure
Developing a Frictional-Cohesive Contact Model
For cohesive solids, it is crucial that the stress history dependent
behaviour is adequately captured. A new contact model that accounts
for this the has been implemented through the API capabilities of
EDEM® in order to capture this. The model parameters for the
simulations are phenomenologically based with the target in capturing
the key bulk characteristics exhibited by the solid.
(4-5)Failed Samples
The unconfined stress-strain relationship and stress history
dependence of a wide range of bulk solids has been evaluated using
the EPT. The stress-strain response and bulk density variation during
confined consolidation is also evaluated by the EPT.
Fig. 8 EPT Simulation in EDEM: (1) 3D Model (2) Filled Sample (3) Confined
Consolidation (4) Unconfined Sample (5) Crushing to Failure
Particle
Size &
Shape
Fig. 9 Force Chain Network During
Consolidation
Particle
Interactions
Fig. 4 Bulk Density Variation For LKAB Iron Fig. 5 Flow Function For LKAB Iron Ore Fines
Ore Fines at Various Moisture Contents
Computational
Considerations
at Various Moisture Contents
Contact me…
Factors Affecting
Modelling Strategy
Fig. 1 Model Scale Relationship – From Meso to If you have any comments about the research or want some more
Macro (Top Left: Iron Ore Agglomerate Top Right:
Industrial Silo Bottom Right Experiment)
EDEM®
Fig. 10 Flow Function Comparison
features such as multiple-processors and the ability to import
and use complex geometries in simulations are also exploited in this
project.
information feel free to contact me. [email protected]
References:
Bell, et al. (2007). Evaluation of the Edinburgh Powder Tester. PARTEC 2007
Härtl & Ooi. (2011). Numerical investigation of particle shape and particle friction on limiting bulk friction in direct shear tests and
comparison with experiments.
Morrissey, et al. (2011). An Experimental and DEM Study of the Behaviour of Iron Ore Fines. CHoPS 2012
Walton & Braun (1986). “Viscosity, granular-temperature, and stress calculations for shearing assemblies of inelastic, frictional disks.”
Luding, S. (2008). “Cohesive, frictional powders: contact models for tension.”
Key Results
Model captures the correct failure modes
Higher consolidation stress produces a higher unconfined strength,
capturing the stress-history dependence.
What Next?
Calibration of experimental data to contact model parameters
Study of the effects of cohesion on silo flow discharge
• DEM simulations of (part of) LKAB Narvik silos