inclinedWall

Overview

Template Info

Target Parameters: c_frict_xx c_rollingFrict_xx roll_damp_xx cohesion_energy_xx cohStrength_xx maxCohStress_exp_xx
Contact Models: tangential no_history tangential history rolling_friction cdt rolling_friction epsd rolling_friction epsd2 cohesion sjkr cohesion sjkr2 cohesion adaptive
Particle Shapes: sphere multisphere

The inclined wall experiment tests if particles on an inclined plate start to roll or slide. When they start to move, the angle of the plate is measured.

Experimental setup

The experiment consists of a plate made of the wall material and particles on it. The plate slowly increases its steepness, and when the particles start to move, the angle of the plate is measured. This angle is called “target angle”. With this experiment, both the the rolling friction and the sliding friction between the particles and the wall material can be characterized. The amount of particles is rather low, no particle-particle interactions should appear, since only particle-wall properties are of interest. Depending on particle shape and surface properties, the particles can either move or slide. This needs to be determined by the experimentator, and the results should be used to calibrate the correct parameter.

Simulation setup

Target parameter

This simulation setup is designed to support calibration of the coefficient of sliding or rolling friction, but only one at a time.

Note

The simulation setup is designed to work with both sliding and rolling friction. If one of the two parameters is to be calibrated, the user should make sure that the other one is set significantly higher than the expected value for the target parameter to avoid sliding in case rolling friction needs to be calibrated or vice versa.

Configuration

This case is so small that it is not execute using MPI, hence the settings for NPROC and nChunks/coresPerChunk are ignored. The only tunable parameter is the number of particles used for calibration via the setting nParticles.

Initialization

During the initialization, 100 particles are placed on a horizontal wall, and the state of the simulation is written to a restart file.

Optimization

The restart file written during initialization is loaded. During the optimization run, the gravity vector is set to emulate an inclination of the plane equal to the measured angle. In the lateral direction, periodic boundary conditions are used. The particles are then allowed to move freely, and the kinetic energy is measured to evaluate the quality function.

Measurement data

The relevant quantity for calibration is the target angle, or the angle at which the particles in the experiment startet to move. It is supplied via a file with the following structure:

targetAngle xx

where xx is the value of your target angle in degrees.

Quality Function

The quality function is computed as

$q = \frac{E_{kin}}{m} \cdot C+c_f$

where $E_kin$ is the total kinetic energy of all particles in the simulation, $m$ their total mass, $c_f$ is the coefficient of (sliding/rolling) friction used in the simulation, and $C=50000$ is a constant determined by experiments.

_images/inclinedWall_qualFnc.png — Quality function for a target angle of 1° and 5°, respectively. When the particles move, the contribution of kinetic energy to the quality function drastically increases, creating a minimum exactly at the onset of motion.

Template parameters

General settings

Parameter	Default
scaleweight	1
pairGranModel	hertz
pairTangential	history
pairCohesion	off
pairRollingFriction	epsd
pairSurface	default
NPROC	1
dumpT	0.01 s
dump	0
G_D_RATIO	25
nChunks	1
coresPerChunk	1

Specific settings

Variable	Default	Meaning
nParticles	1000	Number of particles simulated