UC Berkeley

Abstract

Constitutive Modeling of Weakly Cemented Sands

By

Chukwuebuka Nweke

Doctor of Philosophy in Civil and Environmental Engineering

University of California, Berkeley

Professor Nicholas Sitar, Chair

Weakly cemented sands are prominent in nature and can be found in many geologic deposits all over the world. The same is true for loose sand deposits, which conversely, create undesirable conditions for engineering design and execution. The primary difference between weakly cemented sand and loose sand is the presence of cementation, which enhances the mechanical properties and behavioral response of the former. As a result, the ability to replicate this cementation feature and use it to improve loose sand deposits has been (and is currently) an area of intense investigation and research. Traditional ground improvement methods employ the use of Portland cement via jet grouting, deep soil mixing, compaction grouting, and many others. These methods are considered “environmentally unfriendly” due to their use of “high-embodied” energy materials. A potential solution may lie in the realm of biocementation where sustainable ground improvement technologies use microbial metabolic activity to activate chemical reactions that inevitably induce precipitation of calcium carbonate, which accumulates at the grain contacts and binds the soil skeleton. These artificial cemented granular materials (biocement or Portland cement), as well as naturally cemented materials often serve as the foundation material or support foundation structures of varying overlying infrastructures. For this reason, there is a need for tools that are capable of assessing the improvements (or enhanced characteristics) of these types of sands, and predicting the performance under varying loading conditions.

The focus of this dissertation is to gain a better understanding of the mechanics of cemented and uncemented sands, considering the vast similarities between both states, but highlighting the distinctions that may give insights into how the effects of cementation alter the mechanical properties of sands.

Laboratory triaxial tests were used as a means to investigate the mechanical behavior. It was observed that light cementation preserved the characteristics of stress-strain response typical of uncemented sands, while it also significantly enhanced the strength and stiffness for even low levels of cement content. This was attributed to strengthening of the soil fabric that resulted from the formation of “cement bridges” at the interparticle contacts, which induced increased strength in shear and in compression. Furthermore, it was found that the stress-dilatancy theories used in modeling uncemented sands also apply to weakly cemented sands. Specifically, it was shown that the critical state conditions were relatively unaffected by cementation, leaving the dilatancy to harbor most, if not all, of the cementation enhancement effects. Nor-Sand “bounding plasticity” model was employed as the foundation model due to its ability to represent dilation of the material at low confining pressures. As such, attention was placed on enhancing the dilation component of the model via the inclusion of a cementation parameter that is a function of the amount of cement. In addition, the cementation parameter is capable of evolving with accumulated deformation, allowing for the transition from the cemented to uncemented state. The new model, N-We-Ce (Nor-Sand for Weakly Cemented sands) maintains usage of the majority of parameters from the Nor-Sand base model, while adding 4 – 6 new parameters (depending on the type of test data) describing the contribution of cementation to the strength and stiffness of the sand.