Modeling of Ground Improvement by Prefabricated Vertical Drains in Highly Variable Soils

Abstract

The geotechnical community is well aware of the fact that soils are highly variable in the ground due to the uneven soil micro fabric, geological deposition and stress history. In recent years, there has been an increasing interest worldwide to consider the inherent spatial variability of soil properties in design of many geotechnical engineering applications. However, owing to the complexity of the problem,soil spatial variability has never been taken into consideration in any available design method of ground improvement by prefabricated vertical drains (PVDs). Current design methods of ground improvement by PVDs are typically carried out by assuming a single best estimate of the degree of consolidation based on “average” soil properties that are used to define an “equivalent” homogeneous soil. In reality, however, soil is rarely homogeneous and the assumption of soil homogeneity usually results in unsatisfactory agreement between the desired (predicted) and actual (observed) consolidation values, leading to unreliable and uneconomical design. This paper describes a rigorous stochastic modelling approach that allows the nature of soil spatial variability to be considered in a quantifiable and economically viable manner in design using statistical methods. The approach involves development of advanced 3D models that merge the random field theory and finite element method into a Monte-Carlo frame work. The results highlight the significant role of soil spatial variability in ground improvement by PVDs, and elucidate the importance of proper modelling of this role in design.