Frictional Flow of Dense

Granular Materials
Jin Sun and Sankaran Sundaresan
Department of Chemical Engineering
Princeton University
NETL Workshop on Multiphase Flow Science
Morgantown, WV
April 23rd, 2009
/19
Roadmap for dense granular ow
2
Workshop on Multiphase Flow Research, June 6-7, 2006
20
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
!"#$%& '('( Strategic timeline for addressing key topical areas related to
dense phase flows. Block A is primarily concerned with the issues in
Table 1.1; B refers to issues in Table 1.2; C refers to Table 1.3 and D
refers to Table 1.4..
! Paul Mort and Joseph McCarthy
Report on Workshop on Multiphase Flow Research, Morgantown, WV, June 6-7, 2006
/19
Connection to roadmap
3
Workshop on Multiphase Flow Research, June 6-7, 2006
20
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
!"#$%& '('( Strategic timeline for addressing key topical areas related to
dense phase flows. Block A is primarily concerned with the issues in
Table 1.1; B refers to issues in Table 1.2; C refers to Table 1.3 and D
refers to Table 1.4..
! Paul Mort and Joseph McCarthy
Key questions
addressed:
Action taken in our
project:
/19
Connection to roadmap
What denes the stress in
quasi-static regime?
3
Workshop on Multiphase Flow Research, June 6-7, 2006
20
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
!"#$%& '('( Strategic timeline for addressing key topical areas related to
dense phase flows. Block A is primarily concerned with the issues in
Table 1.1; B refers to issues in Table 1.2; C refers to Table 1.3 and D
refers to Table 1.4..
! Paul Mort and Joseph McCarthy
Key questions
addressed:
Identied internal variables
dening stress states.
Action taken in our
project:
/19
Connection to roadmap
What denes the stress in
quasi-static regime?
What parameters control the
transitions between granular
states?
3
Workshop on Multiphase Flow Research, June 6-7, 2006
20
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
!"#$%& '('( Strategic timeline for addressing key topical areas related to
dense phase flows. Block A is primarily concerned with the issues in
Table 1.1; B refers to issues in Table 1.2; C refers to Table 1.3 and D
refers to Table 1.4..
! Paul Mort and Joseph McCarthy
Key questions
addressed:
Identied internal variables
dening stress states.
Demonstrated the connection
between quasi-static transition
with the jamming point.
Action taken in our
project:
/19
Connection to roadmap
What denes the stress in
quasi-static regime?
What parameters control the
transitions between granular
states?
Continuum rheological models
from quasi-static to rapid ow
regimes?(Goal II in our project)
3
Workshop on Multiphase Flow Research, June 6-7, 2006
20
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
!"#$%& '('( Strategic timeline for addressing key topical areas related to
dense phase flows. Block A is primarily concerned with the issues in
Table 1.1; B refers to issues in Table 1.2; C refers to Table 1.3 and D
refers to Table 1.4..
! Paul Mort and Joseph McCarthy
Key questions
addressed:
Identied internal variables
dening stress states.
Demonstrated the connection
between quasi-static transition
with the jamming point.
Developed a plasticity model for
the quasi-static regime and linked
to particle scale properties.
Action taken in our
project:
/19
Connection to roadmap
What denes the stress in
quasi-static regime?
What parameters control the
transitions between granular
states?
Continuum rheological models
from quasi-static to rapid ow
regimes?(Goal II in our project)
How is stress transmitted?
3
Workshop on Multiphase Flow Research, June 6-7, 2006
20
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
!"#$%& '('( Strategic timeline for addressing key topical areas related to
dense phase flows. Block A is primarily concerned with the issues in
Table 1.1; B refers to issues in Table 1.2; C refers to Table 1.3 and D
refers to Table 1.4..
! Paul Mort and Joseph McCarthy
Key questions
addressed:
Identied internal variables
dening stress states.
Demonstrated the connection
between quasi-static transition
with the jamming point.
Developed a plasticity model for
the quasi-static regime and linked
to particle scale properties.
Probed stress inhomogeneity and
transmission in Jenike shear cell.
Action taken in our
project:
/19
Connection to roadmap
What denes the stress in
quasi-static regime?
What parameters control the
transitions between granular
states?
Continuum rheological models
from quasi-static to rapid ow
regimes?(Goal II in our project)
How is stress transmitted?
3
Workshop on Multiphase Flow Research, June 6-7, 2006
20
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
2
0
1
5
>
2
0
1
5
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
1. Fundamental aspects of stress and flow fields in dense
particulate systems.
2. Definition of material properties on relevant scales, along with
efficient ways to represent properties in models and establish
standards for material property measurements.
3. Given the practical need for continuum modeling capability,
identify the inherent limitations and how to proceed forward, e.g.,
hybrid models that connect with finer scale models (DNS, DEM,
finite element, stochastic, etc.) for finer resolution.
4. Size-scaling and process control (particle / unit-op / processing
system) is critical to industrial applications.
2
0
0
9
2
0
1
2
A
B
C
D
!"#$%& '('( Strategic timeline for addressing key topical areas related to
dense phase flows. Block A is primarily concerned with the issues in
Table 1.1; B refers to issues in Table 1.2; C refers to Table 1.3 and D
refers to Table 1.4..
! Paul Mort and Joseph McCarthy
Key questions
addressed:
Identied internal variables
dening stress states.
Demonstrated the connection
between quasi-static transition
with the jamming point.
Developed a plasticity model for
the quasi-static regime and linked
to particle scale properties.
Probed stress inhomogeneity and
transmission in Jenike shear cell.
Action taken in our
project:
/19
Outline
The continuum model development
Its predictive capabilities demonstrated by
applications to unsteady shear ows
Future work
4
/19
Dissipative plasticity model
*
5 * D.G. Schaeffer. J. Differ. Equ. 66, 19, 1987 ; J.D. Goddard. JFM 568, 1, 2006

ij
= p
ij
p
S
ij

D : D
/19
Dissipative plasticity model
*
5
Stress is related to plastic deformation with rate independence
* D.G. Schaeffer. J. Differ. Equ. 66, 19, 1987 ; J.D. Goddard. JFM 568, 1, 2006

ij
= p
ij
p
S
ij

D : D
Strain rate
Deviatoric strain rate
/19
Dissipative plasticity model
*
5
Stress is related to plastic deformation with rate independence
* D.G. Schaeffer. J. Differ. Equ. 66, 19, 1987 ; J.D. Goddard. JFM 568, 1, 2006

ij
= p
ij
p
S
ij

D : D
/19
Dissipative plasticity model
*
5
Stress is related to plastic deformation with rate independence
* D.G. Schaeffer. J. Differ. Equ. 66, 19, 1987 ; J.D. Goddard. JFM 568, 1, 2006
Pressure

ij
= p
ij
p
S
ij

D : D
/19
Dissipative plasticity model
*
5
Stress is related to plastic deformation with rate independence
* D.G. Schaeffer. J. Differ. Equ. 66, 19, 1987 ; J.D. Goddard. JFM 568, 1, 2006
Stress ratio

ij
= p
ij
p
S
ij

D : D
/19
Dissipative plasticity model
*
5
Stress is related to plastic deformation with rate independence
Pressure and stress ratio are modeled as functions of
microstructural variables: coordination number and fabric
tensor.
* D.G. Schaeffer. J. Differ. Equ. 66, 19, 1987 ; J.D. Goddard. JFM 568, 1, 2006

ij
= p
ij
p
S
ij

D : D
/19
Dissipative plasticity model
*
5
Stress is related to plastic deformation with rate independence
Pressure and stress ratio are modeled as functions of
microstructural variables: coordination number and fabric
tensor.
Stress is evolved through microstructural evolution.
* D.G. Schaeffer. J. Differ. Equ. 66, 19, 1987 ; J.D. Goddard. JFM 568, 1, 2006

ij
= p
ij
p
S
ij

D : D
/19
Model construction
Simulate particle dynamics of homogeneous
assemblies under isotropic compression or simple
shear using discrete element method (DEM)
Extract stress and structural information by
averaging; seek constitutive relations.
6
/19
Computational system
3D periodic domain without gravity
2000 mono-dispersed spherical particles
Restitution coefcient: 0.7
Inter-particle friction coefcient: 0.1-1
Simulate using the LAMMPS code
*
7
Stress
The additional contribution to the stress is due to collisions and contacts, which can be derived
from the principle of virtual displacement for soft interaction potentials and can be modied for
hard sphere systems. The static stress is given as
s
= 1/V

N
i

j,j=i
1
2
r
ij
F
ij
, where r
ij
is the
vector pointing from the center of particle j to the center of particle i, and F
ij
is the contact force
acting on particle i by particle j. The summation can also be performed over contacts in the volume
as
s
= 1/V

c
r
ij
F
ij
.
3
Combining the dynamic and static contributions, the stress tensor for the soft sphere system is
=
1
V
N

m
i
C
i
C
i
+

j,j=i
1
2
r
ij
F
ij

, (11)
which can be used to determine solid stresses from soft sphere MD simulations.
For hard spheres, replacing the force vector by momentum change per unit time, obtain
=
1
V
N

m
i
C
i
C
i
+
1
t

j
r
ij
p
j

, (12)
where p
j
is the momentum change at collision n.
3
For one contact c,
1
2
r
ij
F
ij
+
1
2
r
ji
F
ji
=
1
2
r
ij
F
ij
+
1
2
(r
ij
)(F
ij
) = r
ij
F
ij
.
3
S. J. Plimpton. J Comp Phys, 117, 1-19 (1995) http://lammps.sandia.gov
/19
Computational system
3D periodic domain without gravity
2000 mono-dispersed spherical particles
Restitution coefcient: 0.7
Inter-particle friction coefcient: 0.1-1
Simulate using the LAMMPS code
*
7
Stress
The additional contribution to the stress is due to collisions and contacts, which can be derived
from the principle of virtual displacement for soft interaction potentials and can be modied for
hard sphere systems. The static stress is given as
s
= 1/V

N
i

j,j=i
1
2
r
ij
F
ij
, where r
ij
is the
vector pointing from the center of particle j to the center of particle i, and F
ij
is the contact force
acting on particle i by particle j. The summation can also be performed over contacts in the volume
as
s
= 1/V

c
r
ij
F
ij
.
3
Combining the dynamic and static contributions, the stress tensor for the soft sphere system is
=
1
V
N

m
i
C
i
C
i
+

j,j=i
1
2
r
ij
F
ij

, (11)
which can be used to determine solid stresses from soft sphere MD simulations.
For hard spheres, replacing the force vector by momentum change per unit time, obtain
=
1
V
N

m
i
C
i
C
i
+
1
t

j
r
ij
p
j

, (12)
where p
j
is the momentum change at collision n.
3
For one contact c,
1
2
r
ij
F
ij
+
1
2
r
ji
F
ji
=
1
2
r
ij
F
ij
+
1
2
(r
ij
)(F
ij
) = r
ij
F
ij
.
3
S. J. Plimpton. J Comp Phys, 117, 1-19 (1995) http://lammps.sandia.gov
Small
/19
Characterize microstructure
Coordination number: average number of contacting neighbors
Fabric tensor: average of tensor product of unit contact normals
magnitude indicates the microstructure anisotropy strength; sign
indicates the anisotropy direction
8
Exclude particles with zero or one contact
n
p,c
A
xz
Z
2
=

N
p=1

c
p
2
c=1
1
N
2
A =
1
N
c2
N

n=1
c
p
2

c=1
n
p,c
n
p,c

1
3
I
/19
4 4.5 5 5.5 6 6.5 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
Z
2
p
d
/
k


Isotropic compression
Simple shear
Pressure equation
9
= 0.5
/19
4 4.5 5 5.5 6 6.5 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
Z
2
p
d
/
k


Isotropic compression
Simple shear
Pressure equation
9
p = a(Z Z
c
)
b
+
5
(A : A)(Z Z
c
)

6
= 0.5
/19
4 4.5 5 5.5 6 6.5 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
Z
2
p
d
/
k


Isotropic compression
Simple shear
Pressure equation
9
p = a(Z Z
c
)
b
+
5
(A : A)(Z Z
c
)

6
= 0.5
Isotropic pressure
a = 0.0052, b = 2.48
/19
4 4.5 5 5.5 6 6.5 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
Z
2
p
d
/
k


Isotropic compression
Simple shear
Pressure equation
9
p = a(Z Z
c
)
b
+
5
(A : A)(Z Z
c
)

6
= 0.5
Excess pressure caused by
structural anisotropy

5
= 1.1,
6
= 1.2
/19
!0.06 !0.04 !0.02 0 0.02 0.04 0.06
!0.5
!0.4
!0.3
!0.2
!0.1
0
0.1
0.2
0.3
0.4
0.5
A
xz
!
z
x
/
p


DEM
DEM
linear "
quadratic "
long amplitude cyclic shear
Shear stress ratio
Both steady and unsteady shear ratios following similar
variation against anisotropy
Modeled as function of fabric tensor
10
/19
!0.06 !0.04 !0.02 0 0.02 0.04 0.06
!0.5
!0.4
!0.3
!0.2
!0.1
0
0.1
0.2
0.3
0.4
0.5
A
xz
!
z
x
/
p


DEM
DEM
linear "
quadratic "
long amplitude cyclic shear
Shear stress ratio
Both steady and unsteady shear ratios following similar
variation against anisotropy
Modeled as function of fabric tensor
10
=
1
+
2
A : D

D : D
/19
!0.06 !0.04 !0.02 0 0.02 0.04 0.06
!0.5
!0.4
!0.3
!0.2
!0.1
0
0.1
0.2
0.3
0.4
0.5
A
xz
!
z
x
/
p


DEM
DEM
linear "
quadratic "
long amplitude cyclic shear
Shear stress ratio
Both steady and unsteady shear ratios following similar
variation against anisotropy
Modeled as function of fabric tensor
10
=
1
+
2
A : D

D : D
+
3
(A : D)
2
D : D
/19
Evolution equations
Functions of A and D; satisfying frame indifference.
Satisfy stability requirement.
material constants, c and , can be calibrated using DEM data.
11
Jaumann derivative
Quarterly Report (December 2008) DE-FG26-07NT43070 14
4.2.2 Evolution equations
The microstructure is characterized by the coordination number, Z and the fabric tensor, B.
The evolution equations describing their rate of change against strain rate have be developed,
which take the following form:

Z =

1

B : D + [

D : D +
3

D : DZ](1 +

4
tr(D)
B : D
) + Ztr(D), (6)
and

B = c
1
S + c
2
(

D : D)B + c
3
(B : D)B, (7)
where

Z and

B are the time derivative and Jaumann derivative, respectively, i.e.,

Z =
dZ
dt
and

B =
dB
dt
+ B W W B. The
1

4
and c
1
c
3
in the evolution equations are material
constants, which can be calibrated against DEM data of the coordination number and fabric
evolution. The S and tr(D) denotes the deviatoric and trace of the strain rate tensor.
4.2.3 Yield function and ow rule
With D interpreted as the plastic strain rate, equation 3 represents a plastic ow rule. In
particular, we can rewrite the equation as

D =

/p. (8)
By denition, |

D| = 1, it follows from equation 8 that

= 1, (9)
which represent a scalar yield condition on

. The yield function for this plastic model can

be written as
F() =

p
2

2
= 0. (10)
The deviatoric stress tensor in principle stress space can be written as

1
p 0 0
0
2
p 0
0 0
3
p

, (11)
where p =
1
3
(
1
+
2
+
3
). The yield function becomes
(
1
p)
2
+ (
2
p)
2
+ (
3
p)
2
= p
2

2
. (12)
This prescribes a cone-shaped yield surface in the principle stress space with its apex at the
origin and with the space-diagonal as its symmetric axis. The angle of the cone depends on
. As changes with the fabric evolution, the material reaches a series of co-axial cones. The
increasing (decreasing) dictates the isotropic hardening (softening) behavior of the model
material, which is a result of the fabric evolution as revealed by the yield function.

Z =
1
A : D +
2

D : D +
3

D : DZ +
4
tr(D)

A =
dA
dt
+A WW A
Stickel, J. et al. A constitutive model for microstructure and total stress in particulate suspensions.
Journal of Rheology, 50(4):379413, 2006.
Coordination number
Fabric tensor

A = c
1
S + c
2
(

D : D)A + c
3
(A : D)A
/19
0 1 2 3 4 5
!0.06
!0.04
!0.02
0
0.02
0.04
0.06

0
t
A
x
z


DEM
model
Constant volume shear reversal
12
0 1 2 3 4 5
4.85
4.9
4.95
5
5.05
5.1
5.15
5.2
5.25
5.3
5.35

0
t
Z
2


DEM
model
= 0.6
= 0.5
S
h
e
a
r

c
o
m
p
o
n
e
n
t

o
f

f
a
b
r
i
c
C
o
o
r
d
i
n
a
t
i
o
n

n
u
m
b
e
r
/19 0 1 2 3 4 5
!0.4
!0.3
!0.2
!0.1
0
0.1
0.2
0.3
0.4
0.5

0
t
!
z
x
/
p


DEM
model
s
t
r
e
s
s

r
a
t
i
o
0 1 2 3 4 5
1
2
3
4
5
6
7
8
9
10
11
x 10
!3

0
t
p
d
/
k


DEM
model
S
c
a
l
e
d

p
r
e
s
s
u
r
e
0 1 2 3 4 5
!0.06
!0.04
!0.02
0
0.02
0.04
0.06

0
t
A
x
z


DEM
model
Constant volume shear reversal
12
0 1 2 3 4 5
4.85
4.9
4.95
5
5.05
5.1
5.15
5.2
5.25
5.3
5.35

0
t
Z
2


DEM
model
= 0.6
= 0.5
S
h
e
a
r

c
o
m
p
o
n
e
n
t

o
f

f
a
b
r
i
c
C
o
o
r
d
i
n
a
t
i
o
n

n
u
m
b
e
r
/19
Constant pressure shear reversal
13
0 0.5 1 1.5 2 2.5 3 3.5 4
!0.05
!0.04
!0.03
!0.02
!0.01
0
0.01
0.02
0.03
0.04
0.05

0
t
A
x
z


DEM
model
0 0.5 1 1.5 2 2.5 3 3.5 4
5.35
5.4
5.45
5.5
5.55
5.6
5.65

0
t
Z
2


DEM
model
C
o
o
r
d
i
n
a
t
i
o
n

n
u
m
b
e
r
= 0.5
= 0.60
S
h
e
a
r

c
o
m
p
o
n
e
n
t

o
f

f
a
b
r
i
c
/19
Constant pressure shear reversal
13
0 0.5 1 1.5 2 2.5 3 3.5 4
!0.05
!0.04
!0.03
!0.02
!0.01
0
0.01
0.02
0.03
0.04
0.05

0
t
A
x
z


DEM
model
0 0.5 1 1.5 2 2.5 3 3.5 4
5.35
5.4
5.45
5.5
5.55
5.6
5.65

0
t
Z
2


DEM
model
0 0.5 1 1.5 2 2.5 3 3.5 4
0.598
0.6
0.602
0.604
0.606
0.608
0.61
0.612
0.614
0.616
0.618

0
t
!


DEM
model
V
o
l
u
m
e

f
r
a
c
t
i
o
n
0 0.5 1 1.5 2 2.5 3 3.5 4
!0.4
!0.3
!0.2
!0.1
0
0.1
0.2
0.3
0.4

0
t
!
x
z
/
p


DEM
model
s
t
r
e
s
s

r
a
t
i
o
C
o
o
r
d
i
n
a
t
i
o
n

n
u
m
b
e
r
= 0.5
= 0.60
S
h
e
a
r

c
o
m
p
o
n
e
n
t

o
f

f
a
b
r
i
c
/19
0 1 2 3 4
0.58
0.585
0.59
0.595
0.6
0.605
0.61
0.615
0.62
0.625

0
t
!
Constant pressure cyclic shear
Small strain amplitude ( ) cyclic shear under constant
pressure condition
Lead to compaction as observed in experiments, e.g., Okada, 1992
14
!2 !1.5 !1 !0.5 0 0.5 1 1.5 2
0.59
0.595
0.6
0.605
0.61
0.615
0.62
0.625
0.63
0.635

0
t
!
= 0.60 = 0.5
DEM Model
Okada, N., Energy dissipation in inelastic ow of cohesionless granular media. PhD thesis, University of California, San Diego

A
= 0.5
V
o
l
u
m
e

f
r
a
c
t
i
o
n
/19
0 0.5 1 1.5 2 2.5 3
!0.05
!0.04
!0.03
!0.02
!0.01
0
0.01

0
t
A
x
z


DEM
model
0 0.5 1 1.5 2 2.5 3
0.6
0.605
0.61
0.615

0
t
!


DEM
model
Reynolds dilatancy
Shear an initially isotropic assembly under constant
pressure
Model predicts correct dilation dynamics and steady
state without tting the dynamic data.
15
V
o
l
u
m
e

f
r
a
c
t
i
o
n
s
h
e
a
r

c
o
m
p
o
n
e
n
t

o
f

f
a
b
r
i
c

/19
1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
2
p
d
/
k
(CV and CN)
0 1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
p
d
/
k
H
K

=0
=0.1
=0.2
=0.3
=0.4
=0.5
=0.7
=1
=0.15
=0.25
Unsteady
0 1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
p
d
/
k
H
K

=0
=0.1
=0.2
=0.3
=0.4
=0.5
=0.7
=1
=0.15
=0.25
Unsteady
0 1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
p
d
/
k
H
K

=0
=0.1
=0.2
=0.3
=0.4
=0.5
=0.7
=1
=0.15
=0.25
Unsteady
= 0.5
= 0.1
= 1
Friction dependence: pressure
16
C.Song, P.Wang, and H.A. Makse. A phase diagram for jammed matter. Nature, 453(7195):629632, 2008.
FIG. 10: Mechanical coordination number versus friction obtained in our numerical simulations
explained in Section VI for dierent preparation protocols characterized by the initial volume
fractions
i
indicated in the gure. The symbols and parameters used in these simulations are the
same as in the plot of Fig. 4.
geometry of the packing. Thus, the coordination number z appearing in Eq. (1) is the
geometrical coordination number related to volume, which is dierent from the mechanical
coordination number Z that can be measured through the force contact network relating to
the isostatic condition. In general we expect Z z, since some geometrical contacts may
carry no force. To show this, imagine a packing of innitely rough ( ) spheres with
volume fraction close to 0.64. There must be z = 6 nearest neighbors around each particle
on the average. However, the mechanical balance law requires only Z = 4 contacts per
particle on average, implying that 2 contacts have zero force and do not contribute to the
contact force network.
Such a situation is possible: starting with the contact network of an isostatic packing
of frictionless spheres having z = 6 and all contacts carrying forces (then Z = 6 also), we
simply allow the existence of tangential forces between the particles and solve the force and
torque balance equations again for this modied system of innitely rough spheres. Such
a solution is guaranteed to exist due to the isostatic condition. The resulting packing is
mechanically stable and is obtained by setting to zero the forces of two contacts per ball, on
30
[f`1('%('*/&eXkli\'6981 JLGGC<D<EK8IP @E=FID8K@FE
nnn%eXkli\%Zfd&eXkli\ (+
Our simple shear data Song et al. Jamming transition
Pressure depends on friction.
Transition point to quasi-static
regime, Z
c,
is related to jamming
transition.
/19
1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
2
p
d
/
k
(CV and CN)
0 1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
p
d
/
k
H
K

=0
=0.1
=0.2
=0.3
=0.4
=0.5
=0.7
=1
=0.15
=0.25
Unsteady
0 1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
p
d
/
k
H
K

=0
=0.1
=0.2
=0.3
=0.4
=0.5
=0.7
=1
=0.15
=0.25
Unsteady
0 1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
p
d
/
k
H
K

=0
=0.1
=0.2
=0.3
=0.4
=0.5
=0.7
=1
=0.15
=0.25
Unsteady
= 0.5
= 0.1
= 1
Friction dependence: pressure
16
C.Song, P.Wang, and H.A. Makse. A phase diagram for jammed matter. Nature, 453(7195):629632, 2008.
FIG. 10: Mechanical coordination number versus friction obtained in our numerical simulations
explained in Section VI for dierent preparation protocols characterized by the initial volume
fractions
i
indicated in the gure. The symbols and parameters used in these simulations are the
same as in the plot of Fig. 4.
geometry of the packing. Thus, the coordination number z appearing in Eq. (1) is the
geometrical coordination number related to volume, which is dierent from the mechanical
coordination number Z that can be measured through the force contact network relating to
the isostatic condition. In general we expect Z z, since some geometrical contacts may
carry no force. To show this, imagine a packing of innitely rough ( ) spheres with
volume fraction close to 0.64. There must be z = 6 nearest neighbors around each particle
on the average. However, the mechanical balance law requires only Z = 4 contacts per
particle on average, implying that 2 contacts have zero force and do not contribute to the
contact force network.
Such a situation is possible: starting with the contact network of an isostatic packing
of frictionless spheres having z = 6 and all contacts carrying forces (then Z = 6 also), we
simply allow the existence of tangential forces between the particles and solve the force and
torque balance equations again for this modied system of innitely rough spheres. Such
a solution is guaranteed to exist due to the isostatic condition. The resulting packing is
mechanically stable and is obtained by setting to zero the forces of two contacts per ball, on
30
[f`1('%('*/&eXkli\'6981 JLGGC<D<EK8IP @E=FID8K@FE
nnn%eXkli\%Zfd&eXkli\ (+
Our simple shear data Song et al. Jamming transition
Pressure depends on friction.
Transition point to quasi-static
regime, Z
c,
is related to jamming
transition.
/19
1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
2
p
d
/
k
(CV and CN)
0 1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
p
d
/
k
H
K

=0
=0.1
=0.2
=0.3
=0.4
=0.5
=0.7
=1
=0.15
=0.25
Unsteady
0 1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
p
d
/
k
H
K

=0
=0.1
=0.2
=0.3
=0.4
=0.5
=0.7
=1
=0.15
=0.25
Unsteady
0 1 2 3 4 5 6 7
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Z
p
d
/
k
H
K

=0
=0.1
=0.2
=0.3
=0.4
=0.5
=0.7
=1
=0.15
=0.25
Unsteady
= 0.5
= 0.1
= 1
Friction dependence: pressure
16
C.Song, P.Wang, and H.A. Makse. A phase diagram for jammed matter. Nature, 453(7195):629632, 2008.
FIG. 10: Mechanical coordination number versus friction obtained in our numerical simulations
explained in Section VI for dierent preparation protocols characterized by the initial volume
fractions
i
indicated in the gure. The symbols and parameters used in these simulations are the
same as in the plot of Fig. 4.
geometry of the packing. Thus, the coordination number z appearing in Eq. (1) is the
geometrical coordination number related to volume, which is dierent from the mechanical
coordination number Z that can be measured through the force contact network relating to
the isostatic condition. In general we expect Z z, since some geometrical contacts may
carry no force. To show this, imagine a packing of innitely rough ( ) spheres with
volume fraction close to 0.64. There must be z = 6 nearest neighbors around each particle
on the average. However, the mechanical balance law requires only Z = 4 contacts per
particle on average, implying that 2 contacts have zero force and do not contribute to the
contact force network.
Such a situation is possible: starting with the contact network of an isostatic packing
of frictionless spheres having z = 6 and all contacts carrying forces (then Z = 6 also), we
simply allow the existence of tangential forces between the particles and solve the force and
torque balance equations again for this modied system of innitely rough spheres. Such
a solution is guaranteed to exist due to the isostatic condition. The resulting packing is
mechanically stable and is obtained by setting to zero the forces of two contacts per ball, on
30
[f`1('%('*/&eXkli\'6981 JLGGC<D<EK8IP @E=FID8K@FE
nnn%eXkli\%Zfd&eXkli\ (+
Our simple shear data Song et al. Jamming transition
Pressure depends on friction.
Transition point to quasi-static
regime, Z
c,
is related to jamming
transition.
p = a(Z Z
c
)
b
+
5
(A : A)(Z Z
c
)

6
a, b,
5
,
6
depend on friction
/19
Friction dependence: stress ratio
Simple shear data
averaged over volume
fractions
Shear stress ratios
increase magnitude as
particle friction increases.
17
0 0.2 0.4 0.6 0.8 1 1.2 1.4
0.26
0.28
0.3
0.32
0.34
0.36
0.38
0.4
0.42
0.44

!
!
z
x
/
p
/19
Friction dependence: stress ratio
Simple shear data
averaged over volume
fractions
Shear stress ratios
increase magnitude as
particle friction increases.
17
0 0.2 0.4 0.6 0.8 1 1.2 1.4
0.26
0.28
0.3
0.32
0.34
0.36
0.38
0.4
0.42
0.44

!
!
z
x
/
p
=
1
+
2
A : D

D : D
+
3
(A : D)
2
D : D

1
,
2
,
3 depend on friction
/19
Summary: model recapitulation
18
Stress constitutive equation

ij
= p
ij
p
S
ij

D : D
/19
Summary: model recapitulation
18
Closure relations linked to microstructure.
Material constants depend on particle friction and elasticity.
Stress constitutive equation
p = a(Z Z
c
)
b
+
5
(A : A)(Z Z
c
)

6
=
1
+
2
A : D

D : D
+
3
(A : D)
2
D : D

ij
= p
ij
p
S
ij

D : D
/19
Summary: model recapitulation
18
Closure relations linked to microstructure.
Material constants depend on particle friction and elasticity.
Stress constitutive equation
p = a(Z Z
c
)
b
+
5
(A : A)(Z Z
c
)

6
=
1
+
2
A : D

D : D
+
3
(A : D)
2
D : D
Microstructure evolution equations.
Material constants depend on volume fraction and friction.

ij
= p
ij
p
S
ij

D : D

Z =
1
A : D +
2

D : D +
3

D : DZ +
4
tr(D)

A = c
1
S + c
2
(

D : D)A + c
3
(A : D)A
/19
Work in progress
Simulate quasistatic triaxial compression/extension;
Further test the continuum model against these
DEM data.
Study incipient yield behaviors and incorporate to
the model.
Extend the model to include strain rate-dependence.
19
This work is supported by a DOE-UCR grant DE-FG26-07NT43070.
Acknowledgment
NETL Publications Standards Manual
Logo
The NETL logotype illustrated on this page is the institutional signature for the U.S. Department of Energy,
National Energy Technology LaboratoryNETL. Its function is to be the graphic identication for that
organization. The relationship among the elements of this logo is essential to preserve this identity.
The specications included on these pages will assist in the proper display of this logo and should be fol-
lowed exactly. Questions concerning this logo and its application may be addressed to the NETL Oce of
Public Aairs Coordination, Contact.PublicAairs@NETL.DOE.GOV.
Color
NETL Logo Color
In color the NETL logo is Pantone Matching System ink, PMS 286, 40%PMS 286 for theburst. t
As RGB color, PMS 286 translates to R 0, G 56, B 168; 40%PMS 285 to R 153, G 175, B 220. t
As CMYK color, PMS 286 translates to C 100, M60, Y 0, K 6; 40%PMS 286 to C 40, M24, Y 0, K 2. t