Discrete Element Modelling of Transitional Soils

点击翻译按钮获取中文摘要

The mathematical constitutive models for engineering soils that are principally used today in design analyses and in predicting the responses of soils are formulated based on laboratory data of either clean sand or clean clay. These soils follow consistent patterns of deformation called the Critical State framework. It is clear that to optimise the design of geotechnical structures it is necessary to understand how and why natural soils might behave differently to idealised clean unbreakable soils, and to characterise these differences. Recent work on residual soils from Brasil, silts from the Po River or a glacial till from Iceland has highlighted the existence of other types of soils that do not follow the Critical State framework, but in a more dramatic way. Being neither sands nor clays these soils have been called transitional soils. The range of soils that can be classified as transitional soils is increasing, but at present their patterns of behaviour cannot be unified in a consistent manner. Whether or not a high proportion of small-sized particles (fines), the particle size distribution or the crushability of coarse-grain soils are important effects to the formation of transitional behaviour is still unclear, and it is proposed to investigate this using the computational Discrete Element Method (DEM).DEM can be used to understand the hidden physical factors governing these different patterns by visualising what happens at the level of individual particle interactions. It is a powerful tool as on one hand the DEM can simulate macroscopic deformation data that are comparable to laboratory testing data, on the other hand it can capture information at the level of discrete particle contacts and provide microscopic information to assist the search process of physically plausible parameters to be used in constitutive soil models. With the Itasca PFC3D program, soil particles can be each represented by a group of bonded balls. By adding different proportions of very fine soil particles into an assembly of crushable particles, the effects of particle size distribution and particle breakage can be investigated numerically. Microscopic information that happens at particle contacts (for example breakage) will be calculated and compared with experimental data. It will then become possible to identify when and why the unconventional behaviour happens, and help form bases for a new paradigm for transitional soils.