application of discrete element modelling for the ... · properties virtual experiments, bulk...
TRANSCRIPT
School of Process, Environmental
& Materials Engineering
Ali Hassanpour
Institute of Particle Science and Engineering, School of Process, Environmental and Materials Engineering, University of Leeds
Application of Discrete Element
Modelling for the Development of Particulate Processes:
linking materials properties to performance
1
Modelling Approach
(discrete/continuum)
Sensitivity
Analysis
Insufficient
materials/not
easily accessible
Parameters cannot
be measured
experimentally
Single Particle
Properties
Virtual
experiments, minimising trial and error
Bulk Behaviour
Characterisation of the bulk behaviour based on single particle properties
is of strategic importance in many processes involving particulate solids:
e.g. transportation, filling, mixing, compaction, milling and granulation.
Linking materials properties (single
particles) to performance (bulk
behaviour)
2
The term Discrete Element Method (DEM) is referred to a family of
numerical methods for computing the motion of a large number of
particles based on Newtonian laws of motion (Cundall and Strack,
1979).
MI
Fma
Solving Newtonian
equations of motion
particles
Solving Newtonian
equations of motion for a
large number of particles
Discrete Element Method (DEM)
3
MI
Fma
Fn
Fs
Normal stiffness
Normal force
Kn Fnd
Dashpot Ks
Ftd
Slider
Tangential force
mf
w Tangential stiffness
Cundall, P.A. and Strack, O.D.L.; Geotechnique (1979).
Modelling of Bulk Behaviour using
Distinct Element Method (DEM)
4
Effect of single particle properties and process variables on bulk
behaviour in particular processes needs to be understood......
sensitivity analysis
In a number of applications there is insufficient material for testing
or material is not easily accessible, e.g. pharmaceutical and
nuclear industries
Some parameters can not be measured or quantified in the -
experiments, e.g. internal particle flow and stresses
Scale-up: moving from lab scale to pilot plant and industrial scales
requires extensive trial and error….
Modelling is a mean to interpret experimental results
How DEM Modelling Can be Useful..
5
Both beads free-flowing
6
Analysis of Segregation of Mixtures: Sensitivity analysis
How to avoid segregation of light fine particles from dense course particles?
Dense Coarse Beads Cohesive
Surface Energy (G) = 0.5 J/m2
Understanding the cause of segregation during heap formation
7
Segregation during Heap formation • Interpretation of Experimental Results
• Sensitivity Analysis
Initial packing of sample: a randomly mixed system
Colours represent volume (related to size) of particles
5 mm
4 mm
3mm
2 mm
8
DEM Modelling of Segregation during Heap
Formation: effect of cohesion on
segregation
Colours represent average size of particles in each bin
5 mm
4 mm
3mm
2 mm
Fine particles are cohesive
All particles are
cohesive
All particles are free
flowing
9
DEM Modelling of Segregation during Heap Formation: effect of cohesion on segregation
10
Formation of Seeded Granules in
Cyclomix High Shear Mixer: Interpretation of Experimental Results
Seeded granules occur in Cyclomix high shear mixer under certain
operating conditions
11
DEM Simulation of Granulation in
Cyclomix
Seeded granules are quickly formed in the high shear region (middle part) and
break as soon as they approach the top part 12
DEM Simulation of Granulation in
Cyclomix: formation of seeded
granules
Modelling the agglomerate and compact breakage
0
0.5
1
1.5
2
2.5
3
3.5
4
0 0.02 0.04 0.06 0.08 0.1
Displacement (mm)
Pre
ssu
re (
kP
a)
Pre-consolidaition
pressure= 6 KPa
Load
Stainless
Steel
Stainless
Steel
Predicting Breakage of Granules,
Compacts
13
Particle motion analysis in a
paddle mixer
14
DEM of Mixing of Particulate Solids: Minimising Trial and Errors; Insufficient Available
Materials; Parameters Difficult to Measure
6l paddle mixer was seeded with a positron
charged tracer particle.
The mixer was run at various condition over
a period of time (usually 20-30 min).
Position of the tracer is continuously
recorded against time.
Bulk flow properties per trial is analysed
from the temporal velocity and occurrence
frequency of the tracer.
Position of particle
generation
Impellers
Experimental Measurement using
(Position Emission Particle Tracing (PEPT)
15
0
0.02
0.04
0.06
0.08
0.1
0.12
0.00 0.50 1.00 1.50 2.00 2.50
Fre
qu
en
cy
Velocity (m/s)
DEM modeliing
PEPT measurements
DEM: data on all particles
PEPT: the time averaged data for one tracer but over a long period of running
time (excess of 10 minutes)
Average normalised velocity from PEPT: 0.43
Average normalised velocity from DEM: 0.41
0.36 0.72 1.08 1.44
Normalised Velocity (-)
16
Quantitative Comparison of Powder
Flow between DEM and PEPT
5 l, 5 Hz Tip Speed 3.52 m/s
1 l, 7 Hz Tip Speed 3.17 m/s
Scale-up w2/w1=(d1/d2)n n = 1 Tip speed constant
n = 0.5 Froude number constant
m/s 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 50 l, 2 Hz
Tip Speed 4.13 m/s
DEM Modelling of Cyclomix High Shear
Mixer Granulators: Scale-up
The coefficient of variation of shear stress decreases as the impeller tip speed is increased.
DEM Modelling of Cyclomix High Shear
Mixer Granulators: Scale-up
0
0.2
0.4
0.6
0.8
1
1.2
50 5 1
Co
eff
icie
nt
of
Va
ria
tio
n o
f S
he
ar
Str
es
s (-
)
Granulator Size (l)
Tip Speed (n=1)
n=0.8
Froude (n=0.5)
Tip speed
increases
Tip speed
increases
Modelling of Particle Milling: • Minimising Trial and Errors
• Insufficient Available Materials
19
1- Modelling particle
breakage: particle made
of agglomerates, clusters
of smaller elements bond
Computationally very
expensive, difficult to
model full scale mill
2- Predicting the mill
performance: collisional
energy, stress magnitude
and distribution
DEMV Simulations
20
Modelling of Ball Milling
DEM simulation at 25 Hz of milling frequency in the single ball mill
n
j
jn vmE1
2
2
1 *
• Milling power (Pn ) is deduced from:
• Milling energy (En ) is deduced from the relative velocity (v ) and
reduced mass (m* ) of the two objects in contact by:
t
EP n
n
K p = 0.1218 P n H / K c2
R2 = 0.9826
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 0.1 0.2 0.3 0.4 0.5 0.6
P n H / K c2 (m
2 s
-1)
Mil
lin
g R
ate
Co
nst
an
t, K
p (s
-1)
MCC-18 HzMCC-25 Hz
αLM-25 HzStarch-25 Hz
Sucrose-12 HzSucrose+Aerosil-25 Hz
21
Unification of Results
Material property group
Milling Power
Solid-fluid flow modelling
Fluidization Sedimentation/
re-suspension
DEM + Continuum Method (CFD); Full Fluid-Solid coupling
Solid/fluid interaction
22 22
Dispersion
Pneumatic conveying,
powder dispersion and
fluid diffusion
Solid-fluid Flow Modelling:
DEM-CFD coupling
23
Agglomerate dispersion
With an increase in bonding interface energy it becomes increasingly difficult
to disintegrate particle clusters.
0.2 J/m2
0.6 J/m2
1.0 J/m2
t = 1×10-4 s t = 2×10-4 s t = 3×10-4 s t = 4×10-4 s
Solid-fluid Flow Modelling:
agglomerate dispersion
The dispersion ratio (DR); i.e. ratio of the number of broken bonds to the initial number of bonds, (DR = 1 means all bonds are broken)
0.0
0.2
0.4
0.6
0.8
1.0
0 50 100 150
DR
(-)
Relative velocity (m/s)
0.1 J/m²
0.2 J/m²
0.3 J/m²
0.4 J/m²
0.5 J/m²
1.0 J/m²
0
t0
N
N-NDR
Solid-fluid Flow Modelling:
agglomerate dispersion
DR shown as a function of relative velocity between the fluid and particles.
26
DEM provides useful information in understanding particulate
processes and obtaining parameters difficult to measure by
experiment.
DEM analysis shows good capabilities of interpretation of
experimental data
Numerical modelling capabilities can enable virtual experiments
instead of extensive trial and errors:
Particulate Process Development
Process Optimisation
Process Scale-up
Challenges and Opportunities
Realistic and Complex Models
High Performance Computing (CPU&GPU)
Model Calibration
Concluding Remarks
2002 (5,000) 2005 (40,000) 2008 (500,000) 2012 (10 million)
Development of Modelling Capabilities (Desktop Workstation)
28
Prof. Mojtaba Ghadiri
Dr Colin Hare
Dr Graham Calvert
Dr Hongsing Tan
Dr Andrew Bayly
Dr Boonho Ng
Dr Chi Kwan
Dr Hossein Ahmadian
Mr Massih Pasha
Mr Umair Zafar
Mr Mohammad Afkhami And colleagues in Ghadiri research group
Acknowledgements
Thank you for your attention ! 29