The practice of widening levees to mitigate frequent river flooding is globally prevalent. This paper addresses the pressing issue of sand-filled widened levee failures under the combined effect of heavy rainfall and high riverine water levels, as commonly observed in practice. The primary objective is to illuminate the triggering mechanism and characteristics of such levee failures using the well-designed physical model experiment and Material Point Method (MPM), thus guiding practical implementations. Experimentally, the macro-instability of the levee, manifested as slope failure within the sand-filled widened section, is primarily triggered by changes in the stress regime near the levee toe and continuous creep deformation. Upon failure initiation, the levee slope experiences a progressive failure mode, starting with local sliding, followed by global sliding, and ultimately transitioning into a flow-like behaviour, which characterises the slide-to-flow failure pattern. The slope failure along the interface between the original and new levees is the result of shear deformation rather than the cause. Parametric studies conducted using the calibrated MPM model reveal a critical threshold for the widening width, beyond which the volume of sliding mass and travel angle exhibit no further variation. Increasing the cohesion of the river sand used for levee widening demonstrates the most pronounced improvement in levee stability in the face of the combined effect of intense rainfall and elevated river levels. The MPM-based evaluation of common slope protection measures demonstrates the superior protective benefits of grouting reinforcement and impervious armour layer protection, providing valuable insights for reinforcement strategies in levee engineering applications.

In silty (fine) sand aquifer, water-soil gushing (WSG) of shield tunnel may occur, causing structural damage and even collapse. A comprehensive understanding of the mechanism behind ground displacement and tunnel deformation during WSG in stratified soil was required for guiding disaster-relief in practice. In this article, the responses of ground and shield tunnel to WSG in stratified soil were investigated using a material-point method (MPM). First, a typical case of WSG in stratified soil was studied. By comparing the results with those of WSG in a homogeneous sand, both the ground and tunnel responses to WSG in a stratified soil were clarified. It was found that in stratified soil, the tunnel lining may deform first and then became stable, while in homogeneous sand, the tunnel deformation was shown to continuously develop with time due to unremitting soil loss. Then, the effects of the WSG locations, the sand layer position to tunnel, the layers number and permeability of clay and the discharge rate on WSG were further analyzed.

The frequent occurrence of extreme rainfall events often triggers levee slope failure (LSF), which, due to the levee effect, significantly damages the roads behind the levee. This paper presents a novel framework for the quantitative risk assessment of roads posed by LSF. Within the framework, the innovative integration of Monte Carlo simulation (MCS) and Material point method (MPM) provides a unique solution for simulating the complicated dynamic relationship between LSF and road destruction. MCS generates precise failure scenarios for MPM simulations, overcoming the limitations of traditional approaches in addressing uncertainty in complex scenario systems. With its technical superiority in capturing post-failure deformations, MPM offers critical insights for assessing road exposure and vulnerability. The framework also accounts for indirect losses from road disruptions, which have long been overlooked. The application of the framework to the risk assessment of the road behind the Shijiao Levee in the Pearl River Basin fully demonstrates its practicality and robustness. Compared to traditional risk assessment methods, the proposed framework provides a more refined dynamic evaluation, facilitating the formulation of more effective disaster mitigation strategies.

Accurate continuum modelling of granular flows is essential for predicting geohazards such as flow-like landslides and debris flows. Achieving such precision necessitates both a robust constitutive model for granular media and a numerical solver capable of handling large deformations. In this work, a novel unified phase transition constitutive model for granular media is proposed that follows a generalized Maxwell framework. The stress is divided into an elastoplastic part and a viscous part. The former utilizes a critical-state-based elastoplasticity model, while the latter employs a strain acceleration-based mu(I) rheology model. Key characteristics such as nonlinear elasticity, nonlinear plastic hardening, stress dilatancy, and critical state concept are incorporated into the elastoplasticity model, and the non-Newtonian mu(I) rheology model considers strain rate and strain acceleration (i.e., a higher-order derivative of strain) to capture changes in accelerated and decelerated flow conditions. A series of element tests is simulated using the proposed unified phase transition model, demonstrating that the novel theory effectively describes the transition of granular media from solid-like to fluid-like states in a unified manner. The proposed unified model is then implemented within the material point method (MPM) framework to simulate 2D and 3D granular flows. The results show remarkable consistency with results from experiments and other numerical methods, demonstrating the model's accuracy in capturing solid-like behaviour during inception and deposition, as well as liquid-like behaviour during propagation.

Accurate characterization of soil dynamic response is paramount for geotechnical and protective engineering. However, the impact properties of unsaturated cohesive soil have not been well characterized due to lack of sufficient research. For this purpose, impact tests using the Split Hopkinson Pressure Bar (SHPB) were elaborately designed to investigate the dynamic stress-strain response of unsaturated clay with strain rates of 204 similar to 590 s(-1). As the strain rate increased up to 500 s(-1), a lock-up behavior was observed in the plastic flow stage, which can be explained as the breakage and rearrangement of soil gains under a high level of stress. Further, the strain rate dependency of the dynamic strength was quantitatively characterized by the Cowper Symonds (CS) model and the CS coefficients were determined to be the intercept of 55 and slope of 0.8 in the double logarithmic scale of Dynamic Increase Factor (DIF) and strain rate space. Furthermore, the SHPB test was reproduced using a modified Material Particle Method (MPM), which involves an improved dynamic constitutive model for unsaturated soil considering the strain rate effect. The simulated stress-strain curves basically agree with the experimental results, indicating the feasibility of MPM for investigating the dynamic properties of unsaturated soil under SHPB impact loading.

Bio-inspired probes have emerged as a promising solution for in-situ site characterisation, particularly in challenging terrains and extraterrestrial exploration. This study presents a viable and computationally efficient Material Point Method (MPM) framework for studying Bio-Inspired Cone Pressuremeter (BICP) probe mechanism. With its inherent advantage of particle and continuum frameworks, MPM allows seamless simulation of multi-staged BICP probe propulsion that involves large deformation. A novel implementation strategy was developed for this study to simulate the complex movement of the BICP probe in three sequential stages, including penetration, pressuremeter module expansion, and tip advancement. Sensitivity analysis was conducted to achieve an objective solution and determine the optimum mesh size and mass scaling factor for the BICP probe within the realms of current state-of-the-art MPM formulation. Furthermore, investigations were performed on the established MPM framework to study the influence of probe geometry, material state, and layered soil strata. The findings reveal that in probes with longer pressuremeter modules, larger zone of stress relaxation was observed around the cone tip during module expansion stage than their shorter or double-module counterparts. Meanwhile, the BICP probe's response during all stages in different material states corroborates its sensitivity to the soil's mechanical properties. Although the layered strata significantly influenced the BICP probe's response during the penetration and module expansion stages, it had minimal impact during the tip advancement stage.

Background and aimsUrban trees in coastal cities like Hong Kong may suffer from an uprooting failure when subjected to extreme winds. A proper numerical model for tree uprooting simulation can help to select tree species or soil types that better resist uprooting failure. However, modeling tree uprooting is challenging as it is a cross-disciplinary problem involving complex root system architectures (RSAs) and large deformation of both roots and soils. This study aims to develop a hybrid numerical model that combines truss elements and material point method (MPM) to simulate the entire large-deformation uprooting process of trees with complex RSAs.MethodsThe tree uprooting model is developed by coupling truss elements in finite element method (FEM) with MPM. Laboratory pull-out tests using artificial roots and real root cuttings are adopted to validate the developed model. A comparative study is performed to investigate the difference between using complex and simplified RSAs in tree uprooting simulations.ResultsThe developed model provides consistent predictions of peak load, critical displacement and failure mode when compared with results from laboratory tests. Moreover, the comparative study shows that the uprooting resistance obtained with a complex RSA is higher than that with a simplified RSA. The difference varies with the soil and root mechanical properties.ConclusionThe developed hybrid model offers a novel way for simulating an entire tree uprooting process involving large deformations and complex RSAs. The study shows that using a simplified RSA to approximate the complex RSA might result in misleading failure modes.