published_or_final_version / Applied Geosciences / Master / Master of Science

Modeling of the mechanical behavior of geomaterials is a fundamental yet very difficult problem in geotechnical engineering. The difficulty lies in that the engineering behavior of geomaterials is strongly nonlinear and anisotropic, depending on confining pressure, void ratio, stress history, and drainage conditions. A traditional approach to the modeling of geomaterials is to formulate empirical equations to fit experimental data. Generally, this approach is not able to provide physical insights into the diverse responses observed in the soil mechanics laboratories. Another conventional approach is to make use of the classical plasticity theory, established mainly for metals, to develop constitutive models for geomaterials. While this approach is capable of shedding light on the mechanisms involved, it has been recognized that such models may violate the basic laws of physics. The objective of this thesis is to apply a new approach to constructing constitutive models for geomaterials, by making use of thermomechanical principles. The essence of the new approach is that the constitutive behavior of geomaterials can be completely determined once two thermomechanical potentials, i.e. the free energy and dissipation rate functions, are specified. The yield function and flow rule in the classical plasticity theory can be established from the two potentials, and the models so derived satisfy the basic laws of physics automatically. In this thesis, the theoretical framework for constructing thermomechanical models is introduced. Several concepts in relation to plastic work, dissipated and stored energy are discussed. Both the isotropic and anisotropic models are formulated and realized in this framework and the generated predictions are compared with the test data of a series of triaxial compression tests on sand. To address the important density- and pressure-dependent behaviors of sand in the framework, a state-dependent thermomechanical model is developed, by introducing the state parameter into the dissipation rate function such that a unique set of model parameters is able to predict the behaviors of sand for a wide variation of densities and pressures. Finally, a thermomechanical model for predicting the complex unloading and reloading behaviors of sand is developed by modifying the hardening laws, and the performance of this model is investigated. / published_or_final_version / Civil Engineering / Master / Master of Philosophy

Not available / text

Natural hazards related to ground movement that directly affect the safety of motorists and highway infrastructure include, but are not limited to, rockfalls, rockslides, debris flows, and landslides. This thesis specifically deals with the evaluation of rockfall hazards through the evaluation of LiDAR data. Light Detection And Ranging (LiDAR) is an imaging technology that can be used to delineate and evaluate geomechanically-controlled hazards. LiDAR has been adopted to conduct hazard evaluations pertaining to rockfall, rock-avalanches, debris flows, and landslides. Characteristics of LiDAR surveying, such as rapid data acquisition rates, mobile data collection, and high data densities, pose problems to traditional CAD or GIS-based mapping methods. New analyses methods, including tools specifically oriented to geomechanical analyses, are needed. The research completed in this thesis supports development of new methods, including improved survey techniques, innovative software workflows, and processing algorithms to aid in the detection and evaluation of geomechanically controlled rockfall hazards. The scientific research conducted between the years of 2006-2010, as presented in this thesis, are divided into five chapters, each of which has been published by or is under review by an international journal. The five research foci are: i) geomechanical feature extraction and analysis using LiDAR data in active mining environments; ii) engineered monitoring of rockfall hazards along transportation corridors: using mobile terrestrial LiDAR; iii) optimization of LiDAR scanning and processing for automated structural evaluation of discontinuities in rockmasses; iv) location orientation bias when using static LiDAR data for geomechanical analysis; and v) evaluating roadside rockmasses for rockfall hazards from LiDAR data: optimizing data collection and processing protocols. ii The research conducted pertaining to this thesis has direct and significant implications with respect to numerous engineering projects that are affected by geomechanical stability issues. The ability to efficiently and accurately map discontinuities, detect changes, and standardize roadside geomechanical stability analyses from remote locations will fundamentally change the state-of-practice of geotechnical investigation workflows and repeatable monitoring. This, in turn, will lead to earlier detection and definition of potential zones of instability, will allow for progressive monitoring and risk analysis, and will indicate the need for pro-active slope improvement and stabilization. / Thesis (Ph.D, Geological Sciences & Geological Engineering) -- Queen's University, 2010-03-26 11:25:15.741

No description available.

No description available.

The Amethyst hydro project was proposed on the West Coast of New Zealand as an answer to the increasing demand for power in the area. A previous hydro project in the area was deemed unviable to reopen so the current project was proposed. The scheme involves diverting water from the Amethyst Ravine down through penstocks in a 1040m tunnel and out to a powerhouse on the floodplain of the Wanganui River. The tunnel section of the scheme is the focus of this thesis. It has been excavated using drill and blast methods and is horseshoe shaped, with 3.5x3.5m dimensions. The tunnel was excavated into Haast Schist through its whole alignment, although the portal section was driven into debris flow material. The tunnel alignment and outflow portal is approximately 2km Southeast of the Alpine Fault, the right lateral thrusting surface expression of a tectonically complex and major plate boundary. The Amethyst Ravine at the intake portal is fault controlled, and this continuing regional tectonic regime has had an impact on the engineering strength of the rockmass through the orientation of defects. The rock is highly metamorphosed (gneissic in places) and is cut through with a number of large shears. Scanline mapping of the tunnel was completed along with re-logging of some core and data collection of all records kept during tunneling. Structural analysis was undertaken, along with looking at groundwater flow data over the length of the tunnel, in order to break the tunnel up into domains of similar rock characteristics and investigate the rockmass strength of the tunnel from first principles. A structural model, hydrological model and rockmass model were assembled, each showing the change in characteristics over the length of the tunnel. The data was then modeled using the 3DEC numerical modelling software. It was found that the shear zones form major structural controls on the rockmass, and schistosity changes drastically to either side of these zones. Schistosity in general steepens in dip up the tunnel and dip direction becomes increasingly parallel to the tunnel alignment. Water is linked to shear position, and a few major incursions of water (up to 205 l/s) can be linked to large (1.6m thick) shear zones. Modeling illustrated that the tunnel is most likely to deform through the invert, with movement also capable of occurring in the right rib above the springline and to a lesser extent in the left rib below the springline. This is due to the angle of schistosity and the interaction of joints, which act as cut off planes. The original support classes for tunnel construction were based on Barton’s Q-system, but due to complicated interactions between shears, foliations and joint sets, the designed support classes have been inadequate in places, leading to increased cost due to the use of supplementary support. Modeling has shown that the halos of bolts are insufficient due to the >1m spacing, which fails to support blocks which can be smaller than this in places due to the close spacing of the schistosity. It is recommended that a more broad support type be used in place of discreet solutions such as rock bolts, in order to most efficiently optimize the support classes and most effectively support the rock mass.

The Mw 7.1 Darfield earthquake generated a ~30 km long surface rupture on the Greendale Fault and significant surface deformation related to related blind faults on a previously unrecognized fault system beneath the Canterbury Plains. This earthquake provided the opportunity for research into the patterns and mechanisms of co-seismic and post-seismic crustal deformation. In this thesis I use multiple across-fault EDM surveys, logic trees, surface investigations and deformation feature mapping, seismic reflection surveying, and survey mark (cadastral) re-occupation using GPS to quantify surface displacements at a variety of temporal and spatial scales. My field mapping investigations identified shaking and crustal displacement-induced surface deformation features south and southwest of Christchurch and in the vicinity of the projected surface traces of the Hororata Blind and Charing Cross Faults. The data are consistent with the high peak ground accelerations and broad surface warping due to underlying reverse faulting on the Hororata Blind Fault and Charing Cross Fault. I measured varying amounts of post-seismic displacement at four of five locations that crossed the Greendale Fault. None of the data showed evidence for localized dextral creep on the Greendale Fault surface trace, consistent with other studies showing only minimal regional post-seismic deformation. Instead, the post-seismic deformation field suggests an apparent westward translation of northern parts of the across-fault surveys relative to the southern parts of the surveys that I attribute to post-mainshock creep on blind thrusts and/or other unidentified structures. The seismic surveys identified a deformation zone in the gravels that we attribute to the Hororata Blind Fault but the Charing Cross fault was not able to be identified on the survey. Cadastral re-surveys indicate a deformation field consistent with previously published geodetic data. We use this deformation with regional strain rates to estimate earthquake recurrence intervals of ~7000 to > 14,000 yrs on the Hororata Blind and Charing Cross Faults.

Geotechnical investigations undertaken by GHD Pty Ltd uncovered a previously undescribed rock type in the suburbs of Footscray and Alphington approximately 5 km west and 6.5 km east of Melbourne CBD respectively. The rock encountered appeared to be a breccia type rock with angular high strength fine gravel to boulder sized fragments of relatively unweathered grey to dark grey basalt surrounded by a matrix of orangish brown fine grained brittle material resembling hard clay. Pillow basalts were also encountered in the deposits in the form of 0.6 m or larger globular but highly fractured basalt bodies within the rock mass. The rock was eventually identified as a hyaloclastite, a rock type formed when basalt lava flows into water bodies and is quench fragmented. The debris forms piles of basalt and volcanic glass fragments. The volcanic glass fragments are thermodynamically unstable and are altered to palagonite within as little as 20 years from initial deposition. No prior reference to the occurrence of hyaloclastite in the Melbourne region could be found. As such, the location, extent and geotechnical properties of this rock type are unknown, posing a potential risk to infrastructure and construction projects. This study aimed to investigate the possible origins of the hyaloclastite; develop a theory of emplacement/origin; identify other locations where this rock type may exist; determine the geotechnical properties and engineering geological behaviour of the rock; and develop a classification system for the rocks encountered. A variety of methods were used to gather sufficient information to allow the occurrences and geological and geotechnical nature of hyaloclastites and pillow basalts in the Melbourne area to be better understood. Samples of the rock were obtained during the geotechnical investigations undertaken in Footscray and Alphington and outcrop mapping was completed on exposures identified during the course of this study. Historical borehole logs and as built drawings were obtained to assist in the understanding of the previous description terminology associated with the rock now identified as hyaloclastite. Standard and “non-standard” laboratory testing was undertaken as well as classification testing. The field of block-in-matrix rocks “bimrocks” was assessed as a possible method to assist in the understanding of the behaviour and geotechnical properties of the hyaloclastite rock with or without pillow basalts. The RMR, Q-System and GSI rock mass classification systems were used to help understand the rockmass characteristics. A weak rock classification system, a weathered rock characterisation system and a ground behaviour characterisation system were also used to provide information on the possible behaviour of the hyaloclastite type rocks. Development of both 2D and 3D geological models of the two sites indicate that the hyaloclastites encountered in Melbourne were deposited in “lava-deltas”. The hyaloclastites were deposited on advancing subaqueous delta fronts with an overlying layer of subaerial basalt above what has been termed the “passage zone” which represents the historical level of water into which the lava flowed. Strength testing undertaken on the various samples suggested that the hyaloclastite should be classified as a weak rock, with UCS values of ranging from approximately 1 MPa to 10 MPa, and a median UCS value of 1.37 MPa. Using the compiled UCS data and PLT data an estimate of the PLT Is50 to UCS conversion factor “k” was calculated as 10.4. The results of jar slake testing and weatherability index testing were variable: whilst the majority of samples showed no sign of slaking, one sample showed a strong reaction. The samples of disaggregated rock were classified as sandy gravel as per AS1726:1995. Whilst the fine to medium gravel was of subangular grains of basalt the sand was found to be made up of angular fragments of palagonite. Plasticity index and XRD testing of fines obtained from the disaggregation process indicated that the fines are comprised of illite and smectite clay minerals and behave as a high plasticity silt. Several categorisation methods utilised indicated that the hyaloclastite type rockmass strength parameters are controlled partly by the strength of the matrix and partly by the discontinuities and that the rock mass strength is dominated by the pillow basalt behaviour (typical hard rock type behaviours) only once the content of these structures in these rocks exceeds a volume content of 75% pillows to 25% hyaloclastite. Rock mass strength and deformation calculations indicate that the hyaloclastite rock mass is both very weak and also highly deformable (rock mass modulus <100 MPa) when compared with the highly weathered subaerial basalt (~500 MPa) and the fresh/slightly weathered basalt (~15000 MPa). A value of petrographic constant mi used in the Generalised Hoek Brown Criterion was also determined to be 7.01. This is considerably different to the values suggested for “breccia” in the literature of 19±8. A modulus ratio of 150 was also estimated using testing data from Melbourne and also Iceland. The extent of hyaloclastite in the Melbourne region remains unknown. Whilst the location of these deposits is associated with the base of palaeovalleys now infilled by volcanic products, hyaloclastite does not occur in the base of all the palaeovalleys and is expected to be controlled by sea level change and also disruption of drainage lines by damming caused by earlier subaerial flows. Geotechnical practitioners must be aware of the potential occurrence of hyaloclastite as both the hyaloclastite and hyaloclastites with pillow basalt rock masses were found to be significantly weaker and more deformable than the highly weathered subaerial basalt rock. Misidentification of the rock as highly weathered basalt during geotechnical investigation may result in significant under-design. In addition, rock mass behaviour categorisation indicates that block-falls of pillow basalt from excavation walls and roofs may be a risk. Increased excavation effort to remove the pillow basalt structures should also be factored in to projects. To aid identification and understanding of the potential hazards associated with hyaloclastite type rocks, a series of reference sheets has been developed. These reference sheets aim to increase practitioners’ knowledge of hyaloclastites, and the implications for excavation and construction. The reference sheets also provide geomechanical details. Three-dimensional simplified engineering geological block models have also been included to provide graphical information on the relationships and possible geohazards of the various rock types. Future research should aim to further define the extent and engineering properties of hyaloclastites in the Melbourne region and to further define the petrographic constant mi, a better estimate of modulus ratio based on instrumented UCS tests. It is also hoped that now this rock has been recognised in Melbourne that the geotechnical community will reassess previous projects and start to build knowledge on the whereabouts of hyaloclastite and pillow basalt type rocks in the Melbourne area.

Discrete pile rows are widely used for improving the stability of potentially unstable slopes, where columns of reinforced concrete are constructed in the ground to reinforce it and inhibit instability. The method becomes more cost effective with wider pile spacings, but simultaneously there is also increasing risk that the soil will flow through the gap between adjacent piles, rather than arching across it. The impact of pile spacing along the row, which is likely to have a significant effect on stability, is not clearly understood from a current design perspective. In this study the effects of pile spacing on passive interaction with the slope are investigated using a series of geotechnical centrifuge model tests which are interpreted with a proposed theoretical framework. A total of 23 geotechnical centrifuge model tests were successfully carried out (Chapters 3 and 4): • A plane strain model slope was subjected to up to 50 g centrifugal acceleration, with the upper layer of the slope tending to fail on an underlying predefined surface. The model piles were instrumented to measure bending moment, and hence the shear force and pressure on the piles resulting from interaction with the unstable layer were deduced using a curve-fitting technique. Cameras ‘on-board’ the centrifuge model allowed in-flight photogrammetry to be used to determine soil or pile displacement. • Pile spacing (s/d) was varied, which determined limiting pile-soil interaction for the row, and variation of other geometrical parameters (l/h) for the slope controlled the total load on the pile row. • A number of mechanisms of behaviour for the reinforced slope were identified ranging from a successfully stabilised slope to shallow and deeper slips passing through the pile row, as well as slips which occurred upslope of the pile row and thus did not interact with it. A theoretical framework was developed and used to interpret the results (Chapter 5): • The centrifuge model test results have been successfully interpreted using the proposed analytical approach. • The centrifuge test results confirm previous numerical modelling results, and hence a simple theory which can be used for calculation of the maximum stabilising force available from interaction of the pile row with the slope. The work presented here also confirmed that another previous theoretical model, although quite widely used, is somewhat flawed. Comparison with a field study where stabilisation has been successful (to date) indicated consistency with the experimental results and associated interpretation.