Lecture DEM
19.01.2011
Discrete Elemente Method (DEM)
Matthias Börner – NaWiTec, Institute of Process Engineering
Matthias.Boerner@ovgu.de
1
Overview
Content
Introduction DEM
- What is DEM?
- Basics
- Equations and models
Software solutions
Example 1
- Screw auger
Coupling to CFD
Example 2
- Fluidized bed
2
Einführung
What is DEM
Problems in process engineering:
particle processes and particle
handling
3
Introduction
What is DEM
Working examples of DEM
4
Introduction
What is DEM
DEM – operational areas
Industry of bulk materials
- Mining
- Metallurgy
- Agriculture
Processes of particle formulation
- Pharmaceuticals
- Fertilizer industry
- Food industry
Mechanical + thermal process engineering
gas
- Drying
- Breakage events
- Behavior of bulk materials and transport solid
properties liquid
- Phase-coupled systems 5
Introduction
What is DEM
Diskrete Elemente Method x,vx F
Observation and mathematical description of
single objects / particles / granules (discrete
calculation)
Investigation of mechanical attributes of discrete
separated structures and allocations z,vz
y,vy
MeshIess method compared to FEM or FVM
Explicit numerical finding of solution (ODE
x,vx
systems)
Interactions between objects only at contact points
First scientific reference 1978 and further
developments by Cundal et al. using the program
BALL z,vz 6
y,vy
Introduction
What is DEM
Advantages
Few equations which have to be solved for each object
Forces and motions can be investigated for each single object
in the middle of system, which are usually not measureable
Limits in usage
Real particle systems consists of billions of elements
All-embracing description not feasible
Compromises by system simplifications – partially difficult to get
conclusions about the system behavior of real, large production
plants
7
Introduction
What is DEM
Current particle–paricle and
particle–wall position as well as
estabished connections
Newton„s Force displacement law
law of motion (applied on every contact)
(applied on every particle) Relative motion
force + momentum Constitutive laws
Contact forces
8
Introduction
What is DEM
Newton‘s laws of motion
I. law
v const wenn Fi 0
i
II. law
dI dmv Discription of particle
F= = motion
dt dt
III. law
FA-B = -FA-B
Superposition of forces
Fres =F1 +F2 +...+Fn
9
Introduction
What is DEM
Force displacement law
Finding contact forces during interaction with resulting
displacement of elements
contact models
Fn
Ft
10
Introduction
Basics
DEM - organization
Neighborhood search
- Determination of interacting particle pairs
- Wall dependencies
Force calculation
- Calculation of forces acting on particles
Integration
- transient dependencies of particle dynamics
11
Introduction
Basics
Neighbourhood search
Finding surfaces and volumes in contact with each other
If no contact exists – exceptionally Newton„s second law
At contact – solution of Newton and force displacement law
Basis: an efficient algorithm for contact search of polygonal
constructed particles – for identical particles “spatial
hashing”
Naive neighborhood search is time consuming due to complex
system structure
12
Introduction
Basics
Particle interactions
Description of single particles interactions with each other and with
their vicinity/wall
Description by so-called contact models
Basis is the force displacement law
Hard-Sphere and Soft-Sphere
(momentum based exchange forces) (contact forces)
13
I ab I n nab I t nab Fcontact ,a Fab,n Fab,t
Introduction
Basics
Contact models
At the contact of particles an overlap at the contact area is allowed
(Soft-sphere)
Choice of approach depends on particle velocities –
- low velocities: soft sphere
- high velocities: hard sphere
Hard-sphere approach fails in prediction at low particle velocities and
long contact times δ
At low contact times (high particle
velocities) the time step resolution
to handle particle contacts increases F F
for the Soft-sphere approach
- Correspondingly more time steps A 14
and longer calculation time
Introduction
Basics
Hertz-Mindlin contact modell
(with and without slip)
- Hertz (Journal of Applied Mechanics 1949)
- Approximation of particle impact like a dash pot, non-linear
- Interconection of springs and dash pots in series
- only for sheres in contact
15
Introduction
Basics
Linear-Spring contact modell
- Extention of Hertz-Model by Cundal 1979
- Simplified solution method – faster in calculation
- Partially based on non-physical models (spring stiffness)
16
Introduction
Basics
Further contact models
For considering various system characteristics
Bond model – strong connection/bond between particle, breakage is
possible, extension of Hertz-Mindlin
Ductile model – Extension of bond model to consider brittle breakage
Cohesion model – Additional inter-particle forces (e.g. van der Waals)
Electrostatic model – Impact on particles in surrounding without direct
contact
Burger‘s model – Combination of Maxwell and Kelvin to investigate
creep processes
etc.
17
Introduction
Basics
Time integration
Rayleigh time step
- transient description of particle impact
- tsim < 0.25 tR for sufficient accuracy
2
1
δ Shear waves
through particle
2
1 3
3
18
tR
Simulation
Parameter
Necessary material properties
Particulate properties to simulate the system of materials
Density ρ in [kg/m³] Volume
Mass
Partikulär
Particle diameter d in [m]
Moment of inertia
Shear modul in [Pa]
Poisson„s ratio
Coefficient of restitution
partikulär
Inter-
Coefficent of static friction
Coefficent of rolling friction
19
Simulation
Parameter
Values of a simulation
Calculated values as result of simulation
x-coordinate Total energy
Position
Potential energy
Kräfte / Energie
y-coordinate
z-coordinate Kinetic energy
Rotational energie
Total force
Abhängigkeit
Velocity (vx,vy,vz)
Zeitliche
Angular velocity Compression force
Mass Momentum
Volume
partikulär
Collisionfrequence
Inter-
Forces in bonds 20
Simulation
Software
Commercial Software
- EDEM by DEM-Solutions
- PFC by Itasca
- Chute Maven by Hustrulid
Technologies
- Elven by Rockfield Software
- SimPARTIK by Frauenhofer Institut
Open Source
- LAMMPS
- LIGGGHTS (CFDEM)
21
Simulation
Software
EDEM – Software for applied particle processes
Short presentation about DEM-applications using the commerical
software EDEM
Product of DEM-Solutions Ltd.
Company founded in 2002
First EDEM version in 2005
Contineous developments in the software
Simple structure with graphical user interface
22
Simulation
Software
Scheme of a DEM-simulation
Geometry Creator Simulator Analyzer
Globals Time step Analyzes and graphical
CAD-Program
Particles Cell size output of results, export to
Usual formats can be imported
Geometry Progress of external programs (e.g.
Simple geometries can be
Factory simulation EnSight)
created within
23
Example 1
Screw auger
Transport of bulk materials
Simulation of transport phenomena of spherical particles in a
rotating screw auger
24
Example 1
Screw auger
Results of simulation
Elongated particle form
(contained of 6 particles)
Rotation of particles
+ transport of particles
25
Coupling with CFD
Basics
Coupled simulations between CFD and DEM
Pneumatic transport
Dry powder inhaler
Particle motions through constrictions 26
Coupling with CFD
Basics
Eulerian-Lagrange coupling between CFD and
DEM
Gas / fluid phase
- Calculation of fluid flow with the continuums approach,
Navier-Stokes equations
- FVM, mesh based
Solid phase
- Discrete calculation of single particles
- solid concentration below <10%
- e.g. DPM, Ansys
UDF
CFD DEM
Momentum exchange
27
Coupling with CFD
Basics
Eulerian-Eulerian coupling between CFD and DEM
Gas / fluid phase
- Calculation of fluid flow with the continuums approach,
Navier-Stokes equations
- FVM, mesh based
Solid phase
- Discrete calculation of single particles
- facilitates the handling of high solid concentrations
- in EDEM
UDF
CFD DEM
Momentum exchange
+ conservation of mass-
28
and momentum
Coupling with CFD
Basics
Field coupling between DEM und CFD
Gas / fluid phase
- Stationary calculation of fluid flow with the continuums
approach, Navier-Stokes equations
- FVM, mesh based
- Export of field data (CGNS-Format)
Solid phase
- Discrete calculation of single particles
- Lagrange or Eulerian
- Only applicable, if particles do not effect the fluid flow
- indirecte coupling scheme for both phases
29
Coupling with CFD
Basics
Scheme of coupled DEM-CFD-simulation
Geometry
in CAD (Import to Ansys) or Design
Modeller
Meshing
Ansys Meshing
Configuration EDEM EDEM EDEM
Fluent simulation adjustments for fluid Adjustments for Creator Simulator
flow coupling
Calculation
Post-Processing
CFD-Post or Ensight
Analysis and graphical output of both
results 30
Coupling with CFD
Basics
Scheme of a coupled simulation
Transfer of fluid EDEM:
field data to EDEM Time step of EDEM
begins after the
converged solution is
delivered by Fluent
Fluent:
Calculation until a
converged solution
EDEM:
Drag and lift forces
determined by Fluent
are applied on particles
at the according position
Fluent:
Particle forces calculated in
EDEM are applied in Fluent
grid cells causing a Particle EDEM:
momentum sink or an
energy source coordinates Particle position is
are transfered updated
to Fluent 31
Coupling with CFD
Basics
Solid volume fraction estimation
Estimation of solid fraction in CFD grid
- Sample points determine if a particle is contained in a grid cell
and how much volume it occupies
- the more sample points are used the more precise and accurate
the solution 2/9
1/9
m
SiVparticle 6/9
i 3/9
DEM
N sampleV fluidcell
5/9 4/9
32
Coupling with CFD
Basics
Drag force on particles
Fluid forces appealing on particles
- various relations are published e.g.:
- Kürten FG FA
- Clift
- Di Felice
- Ergun und Wen & Yu
Ffreestream 0.5cD f Ap v f vp v f vp
FD FL
FD Ffreestream
2
4.8
cD 0.63
Re0.5 33
Example 2
Fluidized bed
Particlde dynamic in fluidized beds
Simulation of mixing, heat and mass transfer processes
34
Example 2
Fluidized bed
Results of simulation
Bi-disperse particle system
d1 = 2mm
d2 = 5mm
ugas = 6 m/s
Facts of interest:
- Mixing effects of both particle
systems
- Discharge behavior
- Fluidization behavior
35
End
Script-Download
Script-Download:
www.ovgu.de/ivt/tvt
Teaching – Download
Login: vorlesung-tvt
Passwort: login-08-x1
Contact: Matthias.Boerner@ovgu.de
36