Physical validation of DEM: comparison with GDR-MiDi inclined-plane flow experiments - Main Page

Benchmark Lead(s)

@Thomas Weinhart

Confirmed collaborator(s)

DEM Code name

Responsible persons

MercuryDPM

LAMMPS

Blaze

@Thomas Weinhart

@dalila vescovi

@Nicolin Govender

Benchmark Type

Benchmark

Verification

Validation

Challenge

Target Result

Compare DEM predictions with the experimental GDR MiDi chute-flow dataset.

Benchmark Geometry and Boundary Conditions

This benchmark concerns dense granular flow down an inclined plane, using the experimental data published in the GDR MiDi collective paper on granular rheology (Figure 5). In these experiments, monodisperse glass beads flow steadily on a rough incline of inclination θ. The measured relationships between flow thickness, mean velocity, shear rate, and the stopping height hstop(θ), the flow height below which the flow arrests, are one of the most widely-used physical datasets for validating granular rheology.

Participants must simulate granular flow on a rough plane using their DEM code, reproducing the experimental geometry, particle size, and material properties. The benchmark focuses on comparing:

density, velocity, and velocity fluctuation profiles
stopping height h_stop(θ), i.e. the flow height below which the flow arrests
flow rule: Froude number Fr=U/sqrt(g_zh) vs h/h_stop(θ), where U denotes the mean flow velocity, g_z the gravitational acceleration in normal direction to the chute surface, and h the steady-state flow height.
rheology: effective friction vs inertial number

This benchmark is explicitly designed as physical validation, comparing not only different codes but also against experimental data. It tests whether DEM implementations reproduce real granular rheology in the dense, steady-flow regime.

The detailed simulation properties (below) will be decided by all participants once there is sufficient expressions of interest in this benchmark.

We will use the setup defined in this paper: https://link.springer.com/article/10.1007/s10035-012-0355-y

Image(s) of the geometry

Additional notes

For now this page describes only a spherical particle case. Yannick expressed interest in doing a non-spherical case as well, a description of that will follow.

Particle Type

Select all applicable

Mixture Type

Select all applicable

Particle properties

Parameters have been nondimensionalised such that the flow particle diameter d=1, mass m=1 and the magnitude of gravity g=1.

Contact model(s) and properties

We use a linear spring-damper contact law; normal spring and damping constants are k = 2e5 and γ = 25; thus the contact duration is tc = 0.005 and the coefficient of restitution is e = 0.88. We use sliding friction with a linear spring-damper in the elastic regime. The tangential spring and damping constants are k_t = (2/7) k and γ_t = γ, such that the frequency of normal and tangential contact oscillation and the normal and tangential dissipation are equal. The microscopic sliding friction coefficient is μ = 1/2.

Simulation setup

The chute is periodic and of size 20 × 10 in the x- and y-directions and has a layer of fixed particles as a base. To make a chute flow of approximate flow height H, N = 200 H particles are inserted into the chute (this assumes a packing fraction of pi/6, so the measured steady-state flow height h will be slightly lower than H). Note, the particles inserted at a low packing fraction, so the particle bed will compact initially.

Time-step

A time step of dt = tc/50 = 1e-4 is used.

Data output interval

Every 10 seconds between t=2000 and 2500 (51 time steps)

Initial configuration & Output data format

Initial conditions and outout format can be found here: GitHub - weinhart/GranularChuteFlows: ON-DEM validation benchmark for dense granular flow down an inclined plane

I only ask that you return particle data, the analysis of continuum quantities can be done using the attached script [TO BE DONE]. The analysis is described in here: https://link.springer.com/article/10.1007/s10035-012-0355-y

The flow height is determined by evaluating the downwards stress profile σ_zz and linearizing it; top and bottom surface is then defined as σ_zz=0 and σ_zz=N m g_z.
The density, velocity, and velocity fluctuations are depth-averaged over the flow height and time-averaged over 2000<t<2500

Tentative schedule

Expression of interest by:

Start date (data received from):

Submission deadline for data:

Data analysis to be completed by:

Tentative publication date for draft submission

Benchmark Lead(s) @Thomas Weinhart
Confirmed collaborator(s)
DEM Code name		Responsible persons
MercuryDPM LAMMPS Blaze		@Thomas Weinhart @dalila vescovi @Nicolin Govender
Benchmark Type	Benchmark Verification Validation Challenge
Target Result	Compare DEM predictions with the experimental GDR MiDi chute-flow dataset.

Benchmark Geometry and Boundary Conditions This benchmark concerns dense granular flow down an inclined plane, using the experimental data published in the GDR MiDi collective paper on granular rheology (Figure 5). In these experiments, monodisperse glass beads flow steadily on a rough incline of inclination θ. The measured relationships between flow thickness, mean velocity, shear rate, and the stopping height hstop(θ), the flow height below which the flow arrests, are one of the most widely-used physical datasets for validating granular rheology. Participants must simulate granular flow on a rough plane using their DEM code, reproducing the experimental geometry, particle size, and material properties. The benchmark focuses on comparing: density, velocity, and velocity fluctuation profiles stopping height h_stop(θ), i.e. the flow height below which the flow arrests flow rule: Froude number Fr=U/sqrt(g_zh) vs h/h_stop(θ), where U denotes the mean flow velocity, g_z the gravitational acceleration in normal direction to the chute surface, and h the steady-state flow height. rheology: effective friction vs inertial number This benchmark is explicitly designed as physical validation, comparing not only different codes but also against experimental data. It tests whether DEM implementations reproduce real granular rheology in the dense, steady-flow regime. The detailed simulation properties (below) will be decided by all participants once there is sufficient expressions of interest in this benchmark. We will use the setup defined in this paper: https://link.springer.com/article/10.1007/s10035-012-0355-y
Image(s) of the geometry
Additional notes For now this page describes only a spherical particle case. Yannick expressed interest in doing a non-spherical case as well, a description of that will follow.
Particle Type Select all applicable	Spheres Sphere clumps Ellipsoidal Super-quadrics Level-set Potential surfaces Polygonal Polyhedra Sphero-polyhedra Other (specify):
Mixture Type Select all applicable	Monodisperse Bidisperse Polydisperse Uniformly graded Other (specify):
Particle properties Parameters have been nondimensionalised such that the flow particle diameter d=1, mass m=1 and the magnitude of gravity g=1.
Contact model(s) and properties We use a linear spring-damper contact law; normal spring and damping constants are k = 2e5 and γ = 25; thus the contact duration is tc = 0.005 and the coefficient of restitution is e = 0.88. We use sliding friction with a linear spring-damper in the elastic regime. The tangential spring and damping constants are k_t = (2/7) k and γ_t = γ, such that the frequency of normal and tangential contact oscillation and the normal and tangential dissipation are equal. The microscopic sliding friction coefficient is μ = 1/2.
Simulation setup The chute is periodic and of size 20 × 10 in the x- and y-directions and has a layer of fixed particles as a base. To make a chute flow of approximate flow height H, N = 200 H particles are inserted into the chute (this assumes a packing fraction of pi/6, so the measured steady-state flow height h will be slightly lower than H). Note, the particles inserted at a low packing fraction, so the particle bed will compact initially.
Time-step	A time step of dt = tc/50 = 1e-4 is used.
Data output interval	Every 10 seconds between t=2000 and 2500 (51 time steps)
Initial configuration & Output data format Initial conditions and outout format can be found here: GitHub - weinhart/GranularChuteFlows: ON-DEM validation benchmark for dense granular flow down an inclined plane I only ask that you return particle data, the analysis of continuum quantities can be done using the attached script [TO BE DONE]. The analysis is described in here: https://link.springer.com/article/10.1007/s10035-012-0355-y The flow height is determined by evaluating the downwards stress profile σ_zz and linearizing it; top and bottom surface is then defined as σ_zz=0 and σ_zz=N m g_z. The density, velocity, and velocity fluctuations are depth-averaged over the flow height and time-averaged over 2000<t<2500
Tentative schedule
Expression of interest by:
Start date (data received from):
Submission deadline for data:
Data analysis to be completed by:
Tentative publication date for draft submission