On Creep during Primary Consolidation of Clays

Abstract

Soft clay deposits are characterized by their strong tendency to undergo significant creep deformation even during the primary consolidation phase. This creep behaviour is usually studied based on information obtained from short-term laboratory tests and therefore involves extrapolation of laboratory measurements far beyond the time of observation.

Two distinct schools of thoughts, referred to as creep hypotheses A and B, have been used as a basis of discussion to assess the effect of creep during the primary consolidation phase. In hypothesis A, the strain at end of primary (EOP) is assumed to be independent of the consolidation period, whereas hypothesis B predicts an increasing EOP strain with increasing consolidation period. Advocates of both hypotheses support the view that creep exists during primary consolidation. However, the two hypotheses describe creep during primary consolidation in totally different ways. In hypothesis A, the role of creep is controlled by total strain rate as a function of stress state and effective stress rate resulting in a unique EOP void ratio–effective stress relationship independent of consolidation duration. In hypothesis B, the creep rate is given by the current void ratio and effective stress state. In other words, any combination of void ratio (strain), effective stress and rate of change of void ratio (strain rate) is considered to be unique throughout the primary and the secondary consolidation phases. An interesting feature of these opposing hypotheses was that both of them were supported by ample laboratory and field data. Advocates of the two hypotheses, independently, presented voluminous laboratory and field data as well as numerical simulations to support their own opinions. As a result, they have formed a controversial topic during the last three decades.

There have been arguments posed against either of the two hypotheses. However, these arguments were mainly focused on tackling theoretical assumptions rather than explaining experimental observations; due to this, these arguments have lacked coherency. Therefore, the main objective of this thesis was to thoroughly study both hypotheses and consistently explain the evidence presented for each of them within a common and consistent  framework. Both laboratory and field evidence were investigated and explained.

Thorough investigation of relevant laboratory tests was carried out and it was concluded that the tests implied hypothesis B. In fact, laboratory measurements which were previously claimed to support hypothesis A, do not actually support hypothesis A. In this study, the reasons for laboratory tests seemingly supporting hypothesis A were pointed out. The laboratory measurements were also numerically studied by an isotache model that represents hypothesis B. It was demonstrated that the isotache model can explain and convincingly capture important features of the various types of laboratory tests considered in this work.

Numerical analyses of several field cases using the hypothesis A model as well as comparison of in-situ preconsolidation stress (σ′p) with the laboratory σ′p were used to substantiate hypothesis A. In this work, an attempt was made to systematically combine studies of reported field cases so that the two hypotheses are conveniently evaluated based on common and well-documented field measurements. Hence, predictions from three constitutive models were compared based on three test fills. The three constitutive models represented hypothesis A, hypothesis B (isotache) and a rate-independent elastoplastic model. Comparisons showed that even though the hypothesis A model was proposed to model creep, the formulations and assumptions adopted in the model effectively hinder it from doing so. In fact, comparison of model formulations during the primary consolidation phase indicated that the hypothesis A model gives settlement predictions equivalent to the rate-independent elasto-plastic model. It was also shown that the acceptable agreements between the hypothesis A simulation results and measurements were coincidental and mainly attributed to sample disturbance counterbalancing creep effect along with some incorrect simplifying assumptions. When the field cases were modelled under appropriate and realistic conditions, the isotache model yielded acceptable prediction of settlement as well as excess pore pressure histories. The main conclusion from this study is, therefore, that the isotache models are well suited to predict settlements of water saturated soft clay deposits when the input data is deduced from laboratory tests of good quality soil samples