Home · Blog · How-to

How to calibrate materials in DEM: angle of repose, restitution and friction

🎯 How-to · Aug 5, 2026 · 11 min read

Angle of repose test simulated in BrixDEM. The slope of the pile is the target of particle-particle friction calibration.
In DEM, the result is only as good as the material parameters. And those parameters don't come from a catalogue: you can't directly measure a powder's contact stiffness or rolling friction. You calibrate them, tuning each one until the simulation reproduces the real material's behavior. This guide shows calibration from scratch, test by test, and how BrixDEM's automatic calibration does the heavy lifting.
Golden rule: calibrate one parameter at a time. Each bench test isolates a single parameter, so order matters. Calibrating everything at once is the most common recipe for a result that never converges.

Why calibrate (and why the catalogue won't do)

DEM contact parameters (stiffness, restitution, frictions) don't map directly to the material's catalogue properties. On top of that, the model simplifies reality: particles are usually spheres, and almost always larger than the real grains to keep the cost feasible. So you don't copy a table value. You tune the parameter until the simulation reproduces the macroscopic behavior you measure on the bench, such as the angle of repose. Wrong parameter in, wrong result out.

The parameters you will calibrate

  • Particle density, to match the bulk density of the material.
  • Contact stiffness (Young's modulus or kn), which controls the overlap between particles.
  • Coefficient of restitution (en), how much a collision bounces.
  • Sliding friction, between particles and between particle and wall.
  • Rolling friction, essential because spheres roll far too easily and, without it, the pile collapses.
  • Cohesion, only when the material is fine or wet.

Step by step, from scratch

The recipe has an order to it: each step calibrates one parameter using a test that isolates it, and you always move forward with the previous ones already fixed.

Measure real material Simulate in DEM Compare angle / rebound Validate reference case adjust parameter
The calibration loop: measure the real material, reproduce the test in DEM, compare, adjust the parameter and repeat until it matches, then validate on a reference case.

In practice, there are three bench tests, and each one is reproduced identically in DEM:

h₀ h₁ Rebound restitution θ Inclined plane wall friction α Angle of repose friction + rolling
The three core tests. Each one is done on the bench with the real material and reproduced identically in DEM (the colored particles): the rebound test calibrates restitution, the inclined plane calibrates wall friction, and the angle of repose calibrates the friction and rolling between particles.

1. Density

Measure the bulk density of the material (mass per occupied volume). In the simulation, tune the particle density until the assembly's bulk density matches, remembering there are always voids between particles.

2. Restitution: rebound test

Drop a single particle from a known height onto a plate of the wall material and measure the rebound height. The ratio between heights gives the coefficient of restitution. Reproduce the same drop in DEM and tune en until the rebound height matches.

3. Wall friction: inclined plane

Place the material on a plate made of the equipment's real material and tilt it slowly. The angle at which the material starts to slide measures the friction between particle and wall. Repeat it in the simulator and tune that friction until the sliding angle matches.

4. Particle friction and rolling: angle of repose

Form a pile, by lifting a full cylinder or pouring through a funnel, and measure the angle of repose. This is the central test: tune the particle-particle friction and the rolling friction until the simulated pile forms the same angle. It's the test in the video at the top of this page.

5. Cohesion (if the material is fine or wet)

Fine powders and wet materials stick together. If that's your case, add cohesion to the model and recalibrate the angle of repose, which is now steeper. Without cohesion, the simulation of a cohesive powder flows far too freely.

6. Validate on a reference case

With the parameter set ready, run a case with known behavior, different from the ones used in calibration (for example a real equipment result). If it reproduces what's expected, your calibration generalizes. If not, go back and revisit the test most sensitive to the deviation.

Coarse-graining: particles larger than the real grains

Coarse-graining is a technique in which one larger computational particle represents a group of smaller real particles, thereby reducing computational cost. Calibration should first be performed to represent the real material, after which scaling laws are applied to preserve its behavior with enlarged particles. Therefore, the material should not be fully recalibrated for each coarse-graining factor; instead, properties such as angle of repose, bulk density, and discharge rate should be validated. The higher the enlargement factor, the greater the risk of losing effects related to particle size, cohesion, segregation, and interactions with the geometry.

BrixDEM's automatic calibration

Running this loop by hand, test by test, is laborious. BrixDEM automates it: you choose the protocol, enter the measurement from the real experiment, and the program estimates the parameter and builds the calibration case on its own, ready to run.

  • Rebound test: estimates the coefficient of restitution from the rebound height.
  • Inclined plane: estimates the particle-wall friction from the sliding angle.
  • Angle of repose: estimates the particle-particle friction and the rolling friction from the pile formed, with robust and repeatable angle detection.

The test variables (geometry, material quantity, duration) are part of the case and can be changed and re-run without rebuilding it. In practice, the loop in the diagram above becomes a few clicks.

BrixDEM's automatic calibration on video: choose the protocol, enter the measurement and run the case.

See the full walkthrough, alongside other video tutorials, and the list of protocols in Features.

Common mistakes

  • Copying catalogue values straight into DEM. Those are not calibrated contact parameters.
  • Calibrating everything at once instead of one parameter per test.
  • Forgetting rolling friction. Without it, spheres roll too much and the pile never reaches the real angle.
  • Running a single test and trusting it. The minimum is restitution, inclined plane and angle of repose.
  • Not validating on an independent case after calibrating.

Frequently asked questions

Do I need an expensive lab to calibrate?

No. The core tests (rebound, inclined plane and angle of repose) are simple bench tests. The work is in reproducing them faithfully in the simulator, which is exactly what automatic calibration makes easier.

Can I use particles larger than the real ones?

Yes, that's coarse-graining. You just calibrate the scaled-up particles to reproduce the macroscopic behavior of the material, such as the angle of repose.

How many tests do I need at a minimum?

For a dry material: restitution (rebound), wall friction (inclined plane) and particle friction with rolling (angle of repose). Add a cohesion test if the material is fine or wet.

Does one calibration work for any equipment?

A well-calibrated set tends to generalize across equipment handling the same material, but it's always worth validating on a reference case before trusting it in a new design.

With the material calibrated, you can simulate silos, mixers, drums, screw conveyors and more. BrixDEM is free to download.

Calibrate your material

BrixDEM is free to download and includes automatic calibration for angle of repose, inclined plane and rebound.

Back to blog