Investigating strain localisation in clay using mica markers and X-ray computed tomography

Mass movements in clay deposits result in damage to infrastructures and buildings with significant social, economic, and environmental consequences. These processes are characterised by strain localisation, a complex process to investigate experimentally and model. Strain localisation in clays is particularly worrisome and possess huge destructive capabilities because clay is characterised by low shear strength. Conventional laboratory tests are essentially a post-mortem destructive analysis of localised deformations, and do not account for the fundamental physics of soil behaviours. Hence the development of 4-dimensional (4-D) non-destructive imaging approaches to study soil mechanical behaviour. However, due to the small size of clay particles compared to achievable X-ray computed tomography (X-ray CT) resolution, it has not been possible to directly evaluate particle scale clay micromechanics non-destructively using 4-D imaging techniques. This thesis presents a novel technique involving the use of plate-shaped (“platy”) muscovite mica marker for the evaluation of the initiation and propagation of strain localisation in kaolin. First, an investigation was carried out to understand the suitability of the use of mica particle markers for the study of clay by carrying out both chemical and mechanical characterization of mica. Subsequently, sample preparation techniques were experimented to understand the appropriate sampling approach with least microstructure disturbance. Furthermore, a novel miniature triaxial cell instrumented with a high capacity tensiometer and a novel platy particle matching algorithm were developed for the study of mica marker particle kinematics within kaolin. Kinematic analysis (displacement and rotation) of mica particle markers within the kaolin sample was then carried out. The results presented in this thesis demonstrated that (i) The particle configuration of silt sized muscovite samples consistently varied (dispersive and non-dispersive) with pore-water chemistry, regardless of whether the samples being tested were suspension sediments or compacted samples. (ii) By adding both silt sized muscovite or sand sized muscovite to kaolin for up to 30% sand sized muscovite or silt sized muscovite, the compressive behaviour is still clay-dominated. Similarly, the addition of mica (up to 30%) to kaolin does not significantly affect its hydraulic conductivity of kaolin. However, the shear strength characteristics of kaolin may significantly change by the addition of about 2.5-30% of silt sized muscovite or sand in the low normal stress (<100 kPa) but not at higher stress regime. (iii) PLATYMATCH (algorithm developed in this thesis) can effectively match platy particles in consecutive sample scans when adequately registered and the particles adequately segmented. (iv) A conceptual model of the initiation and propagation of strain localisation in kaolin was developed. The findings of this thesis implies that there is the potential to use platy mica particle marker images for early (pre-peak shear strength) detection of the initiation and propagation of strain localisation in kaolin and this may possibly be useful in enhancing available constitutive models such as the double scale constitutive model for improved clay behaviour prediction.Mass movements in clay deposits result in damage to infrastructures and buildings with significant social, economic, and environmental consequences. These processes are characterised by strain localisation, a complex process to investigate experimentally and model. Strain localisation in clays is particularly worrisome and possess huge destructive capabilities because clay is characterised by low shear strength. Conventional laboratory tests are essentially a post-mortem destructive analysis of localised deformations, and do not account for the fundamental physics of soil behaviours. Hence the development of 4-dimensional (4-D) non-destructive imaging approaches to study soil mechanical behaviour. However, due to the small size of clay particles compared to achievable X-ray computed tomography (X-ray CT) resolution, it has not been possible to directly evaluate particle scale clay micromechanics non-destructively using 4-D imaging techniques. This thesis presents a novel technique involving the use of plate-shaped (“platy”) muscovite mica marker for the evaluation of the initiation and propagation of strain localisation in kaolin. First, an investigation was carried out to understand the suitability of the use of mica particle markers for the study of clay by carrying out both chemical and mechanical characterization of mica. Subsequently, sample preparation techniques were experimented to understand the appropriate sampling approach with least microstructure disturbance. Furthermore, a novel miniature triaxial cell instrumented with a high capacity tensiometer and a novel platy particle matching algorithm were developed for the study of mica marker particle kinematics within kaolin. Kinematic analysis (displacement and rotation) of mica particle markers within the kaolin sample was then carried out. The results presented in this thesis demonstrated that (i) The particle configuration of silt sized muscovite samples consistently varied (dispersive and non-dispersive) with pore-water chemistry, regardless of whether the samples being tested were suspension sediments or compacted samples. (ii) By adding both silt sized muscovite or sand sized muscovite to kaolin for up to 30% sand sized muscovite or silt sized muscovite, the compressive behaviour is still clay-dominated. Similarly, the addition of mica (up to 30%) to kaolin does not significantly affect its hydraulic conductivity of kaolin. However, the shear strength characteristics of kaolin may significantly change by the addition of about 2.5-30% of silt sized muscovite or sand in the low normal stress (