BLOG

Redefining Geotechnical Failure Analysis: Introducing MPoint2D and MPoint3D

September 2, 2026

By David DeGagne

Bridging the Large-Deformation Gap in Geotechnics

In geotechnical engineering, understanding when a slope, dam, or foundation fails is only half the battle. What happens after initiation—how far the debris travels, the impact force on downstream infrastructure, and the final runout geometry—is often where the greatest risk and liability lie.

Traditional mesh-based numerical methods like standard FVM (finite volume method) or FEM (finite element method) excel at small-deformation analyses of complex soil and rock behavior, including evaluating stress distributions, material yielding, seepage, consolidation, and the onset of instability. However, once localized shear bands form, and an analysis must extend beyond initial yielding into post-failure runout and large deformation begins, small strain assumptions no longer apply. While large-strain FVM simulations (like FLAC3D) can run large-strain models considerably further, mesh degradation due to severe distortions can eventually stall the model.

Automatic remeshing, available in FLAC2D, improves large-strain solutions by allowing models to progress much further. However, in extreme deformation scenarios, each remeshing step must transfer stresses, pore pressures, and plastic history variables from the distorted grid to a new, clean grid. This process can increase computational costs, struggle with material separation, and is often limited to 2D applications (remeshing is not available in FLAC3D).

The same Mohr-Columb model run in small-strain (left) showing displacement vectors and in large-strain (right) showing zones colored by quality (blue being the best and red being the poorest).

By contrast, Discrete Element Method (DEM) software such as PFC2D and PFC3D can readily capture extreme material flow and separation, but modeling large-scale, continuous geotechnical bodies with high fidelity often require complex property calibration and significant computational effort.

To bridge this critical gap, ITASCA is introducing MPoint3D and MPoint2D, the first commercial software specifically developed to bring the Material Point Method (MPM) directly into mainstream geotechnical practice.

A Brief Introduction to MPM

At its core, the Material Point Method is a hybrid Eulerian-Lagrangian continuum approach designed to handle extreme deformation without mesh distortion by combining moving material points (Lagrangian) to track stress and momentum with a fixed computational grid (Eulerian) to solve the governing equations.

MPoint’s solution cycle consists of four primary operations between Lagrangian material points (P) and a stationary Eulerian background grid (I) (after De Vaucorbell et al., 2020).

Unlike discrete element methods that model granular assemblies individually, MPM remains fundamentally a continuum formulation. Standard constitutive laws (such as Mohr-Coulomb, NorSand, strain-softening, or advanced soil models) apply directly to the material.

Meet MPoint

MPoint brings the Material Point Method directly to geotechnical engineers, offering a powerful solution for simulating both the onset and consequences of geotechnical failures in 2D or 3D. MPoint is part of ITASCA’s common software framework and benefits from multithread CPU optimization.

Within MPoint, material points are easily generated for simple regions using specified spatial ranges. More complex regions can be defined using geometry (i.e., imported DXF or STL files), and if a non-regular grid is desired, FISH scripting can be used to generate complicated arrangements or inlets of material points.

Generate material points within (a) ranges, (b) using geometry-space logic to assign group, delete materials (e.g., topography), or use FISH to generate (c) arrayed material points or (d) sets of material points during cycling that stream into the model.

The background node spacing and the number of material points can be adjusted to balance computational speed and model resolution. Constitutive models, material properties, and initial and boundary conditions can all be applied directly to material points (or background nodes) like in a FLAC model. MPoint supports ITASCA’s built-in FLAC constitutive models seamlessly and, with some minor modifications, existing User Defined Models (UDMs) as well. MPoint analyses include large-strain, quasi-static stability (including pore pressures for effective stress), factor of safety (using shear strength reduction), and dynamics for earthquake simulations (optionally).

If you know FLAC3D or FLAC2D, you know MPoint.

Prevent Numerical Fracture

A major technical challenge in MPM is numerical fracture: when material points stretch too far apart, they lose connection with shared background grid nodes. The simulation can incorrectly treat this gap as a physical tear or crack. To prevent these non-physical gaps, MPoint uses material point splitting.

When a material point exceeds a deformation or strain threshold, it splits into two child points placed side-by-side along the primary axis of stretch. To preserve physical accuracy, these child points inherit the parent’s material properties while taking half of its mass and volume.

MPoint3D models of a material undergoing slumping: (a) without splitting and (b) with splitting active.

Avoid Volumetric Locking

Volumetric locking is a numerical issue that typically happens when modeling nearly incompressible materials (such as saturated soil under rapid loading, undrained clay, or liquids) and makes materials artificially stiff and resistant to deformation during computer simulations. Because the software tracks numerous material points inside a single grid element, it creates too many constraints for the calculation to solve and the simulation can mathematically “lock up,” leading to artificially high stresses and unrealistically small displacements.

MPoint handles volumetric locking by separating the deformation into a shear portion (i.e., distortion changing material shape) and a volumetric portion (i.e., compression changing material volume) to which a special treatment is applied. Instead of letting every individual material point calculate its own volumetric deformation (which is over-constrained and leads to locking), MPoint applies a specialized smoothed volumetric deformation across the material points. By relieving the artificial volumetric constraints, MPoint allows soils to flow and deform naturally without unphysical stiffness, giving engineers accurate predictions for complex, large-scale geohazard simulations.

PFC Capabilities in MPoint

MPoint supports two advanced PFC capabilities: domain conditions and walls, which help simplify controlling moving material points at domain limits and provide explicit structures to simplify model generation and provide realistic object/material-point interactions.

Domain Conditions

Domains define the extent of the simulation space. Domain conditions can be assigned to define what happens to material points encountering the domain limits. For MPoint, domain conditions include destroy, stop, and reflect:

Destroy: Any material point that crosses the outer boundary of the domain is immediately removed (deleted) from the model.
Stop: When a material point reaches the domain boundary, its velocity is set to zero, halting its motion.
Reflect: Acts as a frictionless, rigid boundary. Material points that hit the boundary bounce back into the domain, reversing their velocity component normal to the boundary face while preserving parallel motion.

Walls

In MPoint, rigid walls are explicit geometrical surfaces composed of line segments in 2D or triangular facets in 3D that interact with material points. MPoint evaluates the contact interaction between the material points (represented as spheres of equal volume) and the wall facets. Walls may be created by specifying facets individually, as simple configurations (e.g., boxes, planes, spheres, cylinders, cones), and by importing an STL or DXF CAD file via ITASCA’s geometry logic. Wall surface stiffness and strength properties for all built-in PFC contact models and attributes such as translation, rotation, and conveyor mode can be specified. (Note: Available PFC contact models include null, linear, linearcbond, linearpbond, softbond, flatjoint, smoothjoint, subspringnetwork, springnetwork, hertz, hysteretic, rrlinear, arrlinear, burger, eepa, jkr, mohr, lineardipole, and fish.)

Walls are extremely useful for:

Simplifying models using a wall to represent surface topographies, eliminating the need for large numbers of far field material points (e.g., surface run out of tailings storage facilities or earth dams).
Representing rigid boundary objects (e.g., retaining walls, tunnels, dams, or impoundments).
Simulating contact and interaction mechanics with objects (machine tools, pipe or probe penetration, storage bin advancing TBM cutterhead, conveyor belts).

Ground penetration simulation in 2D using frictionless walls (black lines) as a probe surface. The probe was initially above the ground surface. Total strain contours are plotted on the material points.

Why Geotechnical Engineers Need MPoint

Traditional finite element or finite volume tools force engineers to stop their models right at the onset of failure when analyzing slope stability and runout, dam failures, dynamic liquefaction analyses, massive creep, and granular flows. MPoint, on the other hand, allows you to simulate the entire lifecycle of a failure event.

Standard FLAC (FVM) / FEM	MPoint
Small strain	Large strain / Runout
Pre-failure stress/failure	Post-failure flow
Slope stability / FoS	Impact forces
Small deformations	Runout distance

MPoint is specifically engineered to analyze extreme deformations of soil and rock, liquefaction, dynamic flow, and post-failure slope runout—scenarios where traditional mesh-based methods encounter severe mesh distortion that ultimately stall. MPoint also uses familiar continuum constitutive models, greatly simplifying material property assignment for very large deformation analyses.

Key Benefits of MPoint for Engineers

Continuous Modeling of Consequence of Failure: Simulate tailings dam breaches, landslide runouts, and deep excavation collapses from initiation to final deposition.
Continuum Mechanics Accuracy: Retain familiar soil/rock mechanics principles and well-calibrated constitutive behaviors in a straightforward manner.
Risk & Impact Assessment: Extract realistic kinetic energy, runout distance, and structural impact loads needed for risk mitigation and emergency response planning.

A First in Geotechnical Analysis

MPoint marks a major milestone: it provides a dedicated, commercially supported MPM environment tailored specifically for geotechnical engineering practice. Integrated within ITASCA’s familiar and trusted framework, MPoint brings powerful, yet practical, large-deformation modeling into day-to-day workflow.

Get Even More from MPoint

While MPoint provides standalone capabilities for large-strain post-failure modeling, its real power multiplies when paired with existing continuum tools. Stay tuned for our next post, where we’ll explore how MPoint couples with FLAC3D and FLAC2D to seamlessly transition from zones to material point runout within a single, unified analysis workflow to simplify model building, maximize modeling performance, and add advanced boundary conditions (e.g., free field and quiet) for dynamic analyses.

Author

Dave DeGagne

Senior Engineer, Technical Marketing