Skip to content
GeoEQ · Geotechnical software research lab · University of Utah

Solutions · Granular and research modelling

Granular material simulation: particle-scale modelling of soils for research and teaching

Virtual element tests, micromechanics and scripted analysis, and which GeoEQ tools fit each step.

The short answer

Simulate the grains, and the soil behaviour follows.

Granular material simulation represents a soil as a collection of individual particles and calculates how they move and interact, so strength, stiffness and failure can be studied from the particle scale up; GeoEQ DEM is being built for this in 2D, with geoeq for Python covering the scripted analysis around it.

This page is about the task. For the product itself, see GeoEQ DEM. Not to be confused with a digital elevation model, which is a terrain dataset.

The workflow

From question to result, step by step.

Six steps make up a typical granular simulation study. Each is mapped to the GeoEQ tool that covers it, with its true status. Steps GeoEQ does not cover are marked as done in your own tools.

  1. Step 01

    Frame the question

    Decide what the simulation must answer: how a grading or particle shape changes strength, why a specimen dilates, how force is carried through the assembly. A clear question fixes the test, the measurements and how many particles are needed.

    Outside GeoEQ

    Done in your own tools and notes.

  2. Step 02

    Characterise the real material

    Collect the laboratory data the model will be compared with: grain size distribution, index properties, and triaxial or direct shear results for the same material.

    geoeq for Python Available

    geoeq for Python can reduce laboratory test data in a script. Producing the data is done in the laboratory.

  3. Step 03

    Build and run the particle assembly

    Generate an assembly of particles, bring it to a stable starting state under a confining stress, then load it. This is the particle-scale simulation step.

    GeoEQ DEM 2D releasing soon 3D in development
  4. Step 04

    Calibrate against laboratory tests

    Particle-level parameters are not measured directly. They are adjusted until a simulated element test reproduces the measured stress-strain and volume change, then checked on a test that was not used for tuning.

    Outside GeoEQ

    Done in your own tools. GeoEQ does not claim automated calibration.

  5. Step 05

    Post-process and script the analysis

    Turn the output into the quantities a paper or a lecture needs: stress-strain curves, volumetric strain, strength parameters, comparisons across many runs.

    geoeq for Python Available

    Use Python for the batch analysis; geoeq supplies the geotechnical functions. Micromechanical measures such as fabric are computed in your own scripts.

  6. Step 06

    Report or teach

    State the particle count, the loading path and the calibration, and compare with the laboratory result. For teaching, the same run shows students what the continuum equations hide.

    Outside GeoEQ

    Done in your own tools.

Explainer

What particle-scale modelling of soil involves.

What granular simulation involves
A soil is made of grains that touch, slide and rotate. Granular simulation keeps those grains as separate bodies, applies a contact law wherever two touch, and steps the whole assembly through time. The behaviour of the soil is not assumed; it comes out of the interactions. The method most used for this in geotechnics is the discrete element method, introduced for granular assemblies by Cundall and Strack (1979).
Virtual element tests
An element test is a laboratory test, such as triaxial compression or direct shear, treated as a single point of material behaviour. A virtual element test does the same on a simulated assembly: the specimen is confined, loaded along a chosen path, and the stresses and strains are read from the particle motions. Because every particle is observable, a virtual test shows things a physical test cannot.
Micromechanics
Micromechanics asks how particle-scale quantities produce the macroscopic response. Typical measures are the coordination number (average contacts per particle), the anisotropy of contact directions (fabric), and the chains of larger contact force that carry most of the load. Published DEM studies relate these to stiffness, strength and critical state behaviour.
Key outputs
Stress-strain and volumetric strain curves from the simulated test, derived strength and dilatancy, and, for research, the particle-scale quantities above. How the results are compared with laboratory data is what makes them credible.

Common pitfalls

  • Parameters that are not unique. Different combinations of particle friction and stiffness can give similar curves, so calibrate against more than one test.
  • Too few particles. A specimen with a handful of grains across its width is dominated by boundary effects, not by the material.
  • Idealised particles. Round discs or spheres cannot reproduce the interlocking of angular sand without extra modelling; say so when reporting.
  • Treating 2D as 3D. A 2D assembly is a plane-strain idealisation, so compare it with plane-strain data, and treat quantitative agreement with 3D tests with care.
  • Skipping the comparison. A simulation that has not been compared with a laboratory test shows what the model does, not what the soil does.

Reference: Cundall, P. A. and Strack, O. D. L. (1979). A discrete numerical model for granular assemblies. Géotechnique, 29(1), 47-65.

Tools for this work

GeoEQ tools for granular and research modelling.

  • 2D releasing soon 3D in development

    GeoEQ DEM

    Discrete element modelling

    Discrete element modelling of soils and granular materials: particle-scale simulation, starting in 2D with 3D in development.

  • Available

    geoeq for Python

    Open-source library

    Open-source geotechnical functions in Python for the work around a simulation: laboratory test reduction, classification and scripted calculations. It does not simulate particles.

Direct answers

Questions about granular simulation.

What is granular material simulation?

Granular material simulation represents a soil or other granular material as individual particles and calculates how they move and interact. The behaviour of the assembly, such as strength and volume change, then comes from the particle interactions instead of being assumed.

What is a virtual element test?

A virtual element test is a laboratory test such as triaxial compression or direct shear carried out on a simulated particle assembly. The specimen is confined and loaded along a chosen path, and stresses and strains are read from the simulation.

What is the micromechanics of soils?

It is the study of how particle-scale behaviour, such as contacts, sliding and force transmission, produces the macroscopic strength and stiffness of a soil. Simulation is a common tool because every particle and contact can be observed.

Which GeoEQ tool simulates granular materials?

GeoEQ DEM, discrete element modelling for soils and granular materials. It is not available yet: the 2D version is releasing soon and a 3D version is in development. Choose Get notified to hear when it is released.

Can I use geoeq for Python for granular simulation?

No. geoeq is an open-source geotechnical library for Python, not a particle simulator. It is useful around a simulation, for example to reduce laboratory test data or script geotechnical calculations, and it installs with pip install geoeq.

Does a granular simulation replace laboratory testing?

No. A simulation of an idealised assembly is calibrated against laboratory tests and compared with them. It adds insight into why a soil behaves as it does; it does not remove the need to measure the real material.

Why start with a 2D simulation?

A 2D assembly needs far fewer particles and runs faster, which suits teaching, method development and plane-strain problems. It is an idealisation, so results are compared with plane-strain data and treated with care against 3D tests.

Who uses particle-scale soil modelling?

Geotechnical researchers, graduate students and instructors studying granular behaviour, and engineers who want to understand a mechanism before choosing a design method. Universities teach it in graduate geotechnics courses.