How PLAXIS 2D Helps Explain Rainfall-Triggered Debris Flows Sponsored

Debris flows are among the fastest and most destructive forms of slope failure, and predicting when and how they initiate remains one of geotechnical engineering's toughest hazard-assessment problems. A new study presented at the 11th International Conference on Physical Modelling in Geotechnics (ICPMG 2026) in Zurich tackles the problem from two directions at once, large-scale physical testing and finite element simulation, and uses PLAXIS 2D to connect the two.

The research team from the University of Enna "Kore" in Italy, led by V. Lentini together with K.A. Pinargote, E. Basile and F. Castelli, built a Large-Scale Landslide Simulator (LSS): a 4.8 m rigid flume with an adjustable triggering channel and evolution channel, capable of reproducing rainfall-induced debris flows on slopes inclined between 16° and 30°. Reconstituted fine-grained soil, oven-dried and sieved to controlled particle sizes, was mixed to a target water content and left to equilibrate before each test, allowing the researchers to reproduce consistent initial suction conditions run after run.

While the physical flume captured what actually happens as a slope liquefies and moves, the team turned to PLAXIS 2D to explain why it happens, and to see whether a numerical model could reproduce the same failure mechanism under matched boundary conditions. The numerical model was built as a two-dimensional plane-strain representation of the flume's longitudinal section, meshed with 15-node triangular elements refined along the surface and the expected shear zone. Critically, the analysis used PLAXIS's fully coupled hydro-mechanical formulation, letting transient rainfall infiltration and mechanical deformation interact within the same solution rather than being modelled as separate steps, essential for capturing the loss of matric suction and the pore-pressure build-up that actually triggers these failures.

The soil was represented with a Mohr–Coulomb elastoplastic model, while unsaturated flow behaviour was described using a van Genuchten soil-water retention curve, with hydraulic parameters calibrated against the pore pressures measured in the physical tests. Rainfall infiltration was applied as a prescribed flux at the surface of the triggering channel, and the analysis proceeded through initial in-situ stress and suction phases, a transient infiltration phase, and a final fully coupled deformation phase used to pinpoint the onset of instability.

The payoff is visible in the model's displacement output: PLAXIS's total-displacement contour plot reveals a wedge of concentrated deformation developing on the triggering slope and propagating downslope. This rotational–translational failure mechanism closely matches what the physical flume showed during testing. Pore pressure and groundwater head outputs from the same model let the researchers trace how suction collapsed as the simulated rainfall infiltrated, mirroring the accumulation the team recorded with a rain gauge in the lab, where failure occurred once roughly 7.5–8.0 dm³ of rainfall had infiltrated the flume.

By running the physical and numerical models side by side under equivalent conditions, the study shows how PLAXIS can extend the insight gained from a physical flume test, helping validate assumptions about pre-failure deformation and suction loss, while also acknowledging where the two-dimensional idealization cannot fully capture effects such as the flume's lateral confinement. The result is a workflow other researchers can adapt: use physical modelling to observe what happens, and a coupled PLAXIS hydro-mechanical analysis to explain the mechanics behind it, a combination the authors argue improves the reliability of debris flow hazard predictions in real slope environments.