Stress relaxation of quartz sands is simulated using a recently proposed physically based time-to-fracture discrete element method framework. The framework incorporates time-dependency through stress-corrosion-induced grain fracture. This feature is embedded into a pre-existing particle-splitting-based rough-contact crushable model. The model is calibrated to represent Fontainebleau sand, a quartz sand. A controlled on–off computational strategy is adopted to advance the simulation efficiently. Model predictions are shown to compare favourably with laboratory results in oedometric and triaxial conditions in terms of stress relaxation and relaxation rate. Grain size distribution evolution is also tracked and shown to compare well with available laboratory results. The influence of initial mobilized strength q/qmax on stress relaxation is recovered by the model, and explained through increased grain breakage. The simulated relaxation results are examined at the microscale and compared with those from creep experiments. The model displays the nonisochronous behaviour characteristic of sands. The relaxation tests display a state shift towards higher dilatancy conditions that may offer a possible explanation for some observations of pile set-up.

Creep of granular soils is frequently accompanied by grain breakage. Stress corrosion-driven grain breakage offers a micromechanically based explanation for granular creep. This study incorporates that concept into a new model based on the discrete-element method (DEM) to simulate creep in sands. The model aims for conceptual simplicity, computational efficiency and ease of calibration. To this end a new form of normalised Charles power law is incorporated into a DEM model for rough-crushable sands based on the particle splitting technique. The model is implemented using a controlled on–off computational strategy. The model is validated by simulating creep in quartz sands in oedometric and triaxial conditions. Model predictions are shown to compare favourably with experimental results in terms of creep strain, creep strain rates and particle breakage. The model proposed would facilitate the calibration of phenomenological continuum models, but may also be useful to directly investigate structural scale phenomena, such as pile ageing.

Pile set-up, defined as the pile shaft capacity increase with time, is a well-established behavioural feature of displacement piles installed in sand. To date, the mechanisms underlying set-up are poorly understood and, as a consequence, is difficult to reliably incorporate this phenomena into design. Micromechanical investigations have recently shown that pile set-up may be directly related to the time dependent breakage of sand around the pile. Following the idea that fracture propagation has an inherent timescale, a fracture based quartz sand discrete element method (DEM) model, previously validated on creep/stress relaxation experimental results, was implemented into a DEM centrifuge chamber to simulate pile set-up. The particle refinement method was implemented to ensure a manageable number of elements. The model pile was either jacked or driven to the target depth. One month creep period under load was applied to age the system. Pull-out pile tests were simulated for the fresh pile and aged pile conditions. The results show a mobilized shaft capacity increase of over 20% for jacked piles and over 50% for driven piles, within the range reported from field pile tests.

Pile setup is now generally recognized as a characteristic feature of displacement pile axial response in sand. However, a clear understanding of the phenomena underlying pile setup is still missing because reliable and practical methodologies to study the problem are lacking. It is very difficult to measure and control all possible contributing factors in field testing experiments, whereas physical laboratory tests often fail to replicate field observations. The resulting ambiguities undermine proposals to incorporate setup in design procedures. A new methodology to study pile setup based on the discrete-element method (DEM) is proposed here. We build a virtual centrifuge chamber using a granular material DEM model calibrated to represent Fontainebleau sand, a quartz sand. The DEM material model, which incorporates delayed grain fracture, was previously verified by reproducing creep and relaxation in element tests, but relies on much simpler measurements for material calibration. Model piles are installed in the chamber by jacking, performing long (up to 1 month) creep stages and measuring shaft resistance at regular intervals by means of pullout tests. The simulated pile creep and pullout responses were well-aligned with previous experimental observations. The simulations presented strongly support the hypothesis that delayed grain fracture is a significant contributor to pile setup. The methodology presented opens the way to quantify this and other physically based explanations of pile setup and creep.

Stress relaxation of quartz sand is simulated using a recently proposed physically based time-to-fracture discrete element method (DEM) framework. The relaxation of stress on sands under high confining pressure is modelled using stress-corrosion induced grain fracture. This feature is embedded into a pre-existing particle-splitting based rough crushable model for Fontainebleau sand. A controlled on-off computational strategy is adopted to advance the simulation. The computed oedometer stress relaxation curve compares favourably with previous experimental results. Triaxial stress relaxation results indicate that stress relaxation increases linearly with time. Contact forces become more homogenous during stress relaxation; as a result, the sample becomes more stable. These microscopic behaviours agree with some observations in pile ageing, suggesting that the model may be adopted directly to investigate such engineering scale phenomena due to its conceptual simplicity, computational efficiency and ease of calibration.

The time-dependent pile settlement under sustained service load has been reported in the literature and defined as pile creep. Particle breakage has been found as the main mechanism for pile creep in sand by Leung et al. [1]. Based on the time-to-fracture Fontainebleau sand DEM model proposed by Lei et al. [2], particle breakage induced pile creep is first investigated in a centrifuge virtual chamber using the particle refined method. The creep settlement is highly similar to the results reported by Leung et al. [1]. The model shows great potential for investigating the pile creep in a microscopic view and the relevant pile ageing (set-up) phenomenon.

In situ soil characterization often involves penetration tests. The response to both static and dynamic penetration in volcanic sands is known to be controlled by particle crushing, obscuring the effect of density and confinement on the response. This severely limits the use of penetration tests for soil characterization. In this work a virtual calibration chamber (VCC) built using the discrete element method (DEM) is employed to explore the relationship between cone penetration tests CPTs (static) and standard penetration tests SPTs (dynamic) results on a volcanic sand. The simulated CPT results are shown to capture well experiments from the literature over a range of density and confining stress. The comparison between the simulated CPT and SPT results shows good agreement with empirical expressions based on particle size. It was found that energy-based equivalent dynamic tip resistance obtained from the SPT is a good match for static cone tip resistance if shaft resistance effects are discounted. The results presented confirm the potential of DEM based VCC to examine the performance of in-situ tests in complex soil conditions.

The European offshore wind sector has grown rapidly, with significant advances in turbine technologies, in creased sizes, and construction locations further from the shore in deeper waters than ever before. A critical challenge has emerged relating to a growing lack of knowledge surrounding how to design foundations to support these new turbines, with the safety, life-span, cost, and environmental implications coming increasingly into question. To maintain Europe’s stance as a World-leader in offshore wind, Foundations for Offshore Wind Turbines (FRONTIErS) Doctoral Network has been designed to bring together research-intensive universities and major industry stakeholders to train the next generation of graduates with the appropriate skills to tackle the emerging issues presenting as a barrier to continued development of the sector. Eleven Doctoral Candidates have been recruited to tackle significant challenges related to foundation design and performance. Projects focus on topics such as: understanding soil variability, effect of cyclic loading on axial capacity, pile aging, dynamic modelling informed from in-situ testing, time and spatial variation in soil properties, driveability modelling, gravity-base scour effects, multi-directional loading effects, centrifuge testing of cyclic loading response, and dynamic features of wind turbine foundations. This paper presents a preliminary overview of the various PhD projects ongoing as part of this network, which are in the early stages, as well as a summary of training conducted to date.

A contact model able to capture creep of crushable sands within a discrete element method (DEM) framework is presented. Time dependency is established through stress-corrosion induced grain fracture. This is grafted onto a pre-existing particle-splitting model developed to simulate rough-crushable sands. The model is calibrated for Fontainebleau sand and applied to simulate soil creep at high-confining pressures. Creep simulation is advanced using the off-DEM ageing technique. The crackvelocity evolution in particle scale and laboratory oedometer creep curve were successfully captured. Moreover, the onset of creep failure during triaxial creep curve under high deviatoric stress can also be simulated using this model. The results obtained indicate significant potential of this model for micromechanically based investigation of time-dependent soil behaviour.

A recently proposed DEM model for materials with rough crushable grains (Zhang et al. 2021; Ciantia et al. 2015; Otsubo et al. 2017) is here employed to examine the effect of contact roughness on the critical state line, a property of granular materials which is a) fundamental for the evaluation of liquefaction risk and liquefied responses and b) easily accessible through DEM simulation (Ciantia et al. 2019).

Chalk is present in a large area of the East and South coast of the UK where renewables are now being installed. In most cases they are founded on displacement piles, whose design uses empirical methods or partial (wished-in-place) simulations. Installation effects such as grain crushing and pore pressure generation are conservatively estimated (if considered at all), as the change of chalk properties around the pile due to pile installation cannot be easily quantified. In this work it is shown how advanced numerical modelling can be used to guide the design of the foundations in these complex soils.