Geotechnical systems are temporally variable due to both endogenous and exogenous processes. The prediction of their future states is often very relevant due to the potentially significant modifications in their behavior with respect to design limit states. The quantitative forecasting of the future evolution of geotechnical systems such as slopes and embankments subjected to rainfall or earthquake loading is strongly dependent on the spatial variability of hydro-mechanical constitutive model parameters. This temporal variability further adds to the well-known challenges stemming from aleatory and epistemic uncertainties. Suitable modeling of this complexity requires non-deterministic approaches involving, for instance, statistical and probabilistic methods. The acquisition of monitoring data—such as pore water pressure, or settlements—is required to dynamically update engineering models and refine the reliability of forward predictive system response. This work proposes a Bayesian approach for the dynamic updating of spatially variable parameters in numerical models of temporally evolving geotechnical systems. The methodology is implemented within a multivariate random finite element model that recalls Ensemble Kalman Filter–based scheme to update constitutive model parameters as new empirical evidence becomes available. Application to a simplified embankment model under water level fluctuation and transient seepage analysis demonstrates the capability of the proposed approach to capture the temporal evolution of the spatial attributes of the system accounting for uncertainty, thereby enhancing the confidence of predicted temporal evolutions of geotechnical systems.
The increasing emphasis on seismic design following the adoption of Eurocode 8 has necessitated the evaluation of both new and existing structures under earthquake loading conditions. This study addresses the seismic stability assessment of a channel embankment that flows into the Danube River and also serves as part of the river’s primary flood protection infrastructure. Seismic effects were modeled in PLAXIS 2D using the simplified pseudostatic method within a finite element framework. Soil parameters were derived from seismic cone penetration tests, and soil behavior was described by the Hardening Soil Small-Strain (HS-small) constitutive model. The investigation focused on a critical cross-section of the embankment, considering a range of water levels and horizontal ground accelerations. The results indicated that the magnitude of the seismic load fundamentally influences the overall stability, while the effect of water levels is less significant. The critical horizontal soil acceleration coefficient kx,crit, representing the onset of instability, is greater than 0.2. The total displacement is around 50 cm at the limit-equilibrium state. The results provide insight into the embankment’s performance under seismic loading and contribute to a broader understanding of the resilience of flood protection infrastructure.
This study presents field monitoring data of soil temperature, humidity, and piezometric levels collected at the levee of Grizzaga stream in Modena, Italy, over the period from March 2022 to February 2025. The field data reveals how the transient thermo-hydraulic processes occurring within the embankment during the flooding are affected by the initial humidity conditions presenting seasonal variations and affecting the levee behaviour during the successive flood waves. The monitoring data is used to calibrate a coupled thermo-hydraulic numerical model to analyse the levee behaviour in transient saturation conditions, developed with the commercial software FEFLOW. The validated model provides a predictive tool for analyzing the performance of levees during periods of intense rainfall and flooding. It can support risk assessment and decision-making for similar embankment structures subjected to climate change impacts.
In seismically active regions, understanding how existing earth dams respond to earthquakes is essential for developing effective rehabilitation strategies and ensuring their safety. Since well-documented case studies are limited, small-scale physical modelling in a geotechnical centrifuge is a key tool for studying their complex seismic behaviour. Such experiments, however, require compromises: the limited number and duration of tests make it necessary to maximise the information obtained from each one. In particular, within a performance-based framework, it is often of interest to examine dam responses to progressively stronger seismic inputs. In centrifuge modelling, this is typically done by applying ground motions of increasing intensity at defined time intervals. This study presents numerical analyses investigating the effects of sequential seismic loading. A PLAXIS 2D finite element model of a small zoned earth dam was calibrated against centrifuge test data. After simulating the impoundment phase, four seismic inputs of increasing intensity were applied sequentially to reproduce the experimental procedure, and then individually to allow comparison. The results show that applying seismic motions sequentially influences the dam’s response through soil nonlinearity and the accumulation of excess pore water pressure. These findings offer useful insights for the design and interpretation of centrifuge experiments on the seismic behaviour of earth dams and similar geotechnical structures.
It is known that slope movements induced by earthquakes may be responsible for damages on structures and infrastructures interacting with the slope. Depending on its characteristics (e.g. mechanical and dynamic soil properties, morphology of the slope and landslide body, pore water pressure distribution), different seismic slope responses are expected during earthquakes, which may induce differential displacements beneath a structure located on its surface. In this study, a finite element (FE) numerical investigation is carried out to evaluate the permanent displacement fields induced by seismic actions and their effect on the seismic response of a multi-storey reinforced concrete (RC) building, located on a slope characterized by a pre-existing landslide mechanism. The study is conducted adopting an uncoupled approach, consisting in performing 2D FE geotechnical analyses of the slope including a simplified structure for evaluating the seismically induced permanent displacements and 3D FE detailed structural analyses of the RC building. The geotechnical simulations showed that accounting for the slope-structure interaction provides more realistic displacement fields at the base of the structure, while the structural analyses predicted larger inter-storey drift, mainly related to the horizontal displacements correlated to the activation of the landslide mechanism.
