The steady increase in demand on underground utilities experienced over the last decade, combined with a greater appreciation of the indirect costs of open trench techniques and a heightened environmental awareness, has resulted in the utilisation of trenchless techniques becoming increasingly financially and politically attractive. Inadequancies in existing percussive moling techniques employed in the installation of small diameter (<300 mm) pipes and cables over short, straight distances (<35 m), have shown that a need exists for a device with improved penetrative abilities through difficult ground conditions whilst reducing the risk of damage to existing services. The vibro-impact principle, originally developed in the USSR c.1940, employs a combination of vibration and impact to provide the driving energy to a driven element. A resilient connection between the vibrator unit and the driven element permits a vibro-impact device to efectively self-adust the relative level of transmitted vibration and impact. This self-adjustment capability has been shown to increase pentration depths through difficult ground conditions over those possible with pure vibratory or pure impact techniques, and reduce disturbance to the surrounding soil. This thesis considers the application of the vibro-impact principle in the design of a ground moling system. Both the theoretical and practical development of the vibro-impact principle to date has been concerned with vertical application of such a device. The existing theoretical models and design practice have been reviewed and developed to consider the horizontal application of a vibro-impact machine. The development of a laboratory based model vibro-impact moling system is presented. Through a variable parameter investigation on this model system a range of optimum parameters were identified for horizontal penetration of a driven element into a sand test bed. Furthermore the model investigation identified three distinct zones of horizontal vibro-impact penetration, different in form from those previously identified for vertical penetration. Based on the results of both the theoretical development of the vibro-impact principle and the model investigation, a design process was established for a full scale vibro-impact moling system. A theoretical, computer based model was developed to optimise the parameters of the proposed design. The design of the full scale prototype was finalised utilising computer aided design software. Following manufacture the prototype system was subject to a series of laboratory based commissioning trials. Further developments to the pneumatic actuation system based on the results of these laboratory tests led to the development of a fully operational vibro-impact ground moling system. The prototype moling system was then subject to a series of full scale field trials. The results of these trials confirmed the operational response indentified by the model investigation. Furthermore the optimum operating parameters derived from the results of the field trials correlated well with the results of the computer based theoretical model. The field trials thus validated the theoretical analysis and validated the design procedure established by the author. The penetration rates achieved with the full scale prototype vibro-impact moling system indicated an improvement in performance over that which could be expected with conventional impact moles. Successful trials were also conducted in conjunction with a coring tube end fitting, identifying the possibility of developing the system into a portable ground investigation system. The work described in this thesis thus represents a development of existing vibro-impact theory and practice, and highlights the commercial potential of this line of research.

The work presented in this thesis concerns the behaviour of isolated piles subjected to repeated loading and placed at various depths in a medium dense sand upon which either static or cyclic surcharge acted. The piles, which were of laboratory scale, were instrumented by strain gauged load cells located along the inner surfaces of the pile shafts. The behaviour of tension as well as compression piles was examined. It was found that the behaviour of the pile was governed to a large extent by the repeated load level, the number of load cycles and the initial boundary stress conditions existing along the pile shaft. In compression, the pile life-span decreased when the embedment depth increased while the reverse trend was observed,for tension piles. The movement of both tension and compression piles decreased when the surcharge pressure was increased or was cycled, and it was of a minimum value when the upper repeated load acted in-phase with the higher surcharge pressure. For tests performed with static surcharge pressure, repeated loading was found to decrease the bearing capacity and the pulling resistance of the pile. The higher percentage of reduction was recorded for the tension pile. In contrast, after cyclic surcharge tests the pile capacity always increased. At any depth of embedment or surcharge pressure,' as the nun-her of load cycles was increased the shaft load of a compression pile increased up to a peak value then decreased gradually until it reached a limiting value. This limiting value increased when the load level, the pile depth or the surcharge pressure was increased and it was independent of the pile loading history. iii For a tension pile the shaft load decreased progressively as the number of cycles increased until failure occurred.

For the assessment of Masonry Arch Bridges (MAB), many structural and material models have been applied, ranging from sophisticated non-linear finite element analysis models to much simpler rigid-block limit analysis models. i.e. elastic and plastic methods respectively. The application of elastic analysis to MAB suffers many drawbacks since it requires full mechanical characterization of ancient masonry structures. The mechanical characterization of ancient masonry is difficult since these structures have typically undergone a century or more of environmental deterioration and in many cases have been already subjected to extensive modification. Also, sophisticated material models generally require specialized parameters that are hard to assess, particularly if non-destructive tests are used. In these cases practicing engineers typically favour simpler material models, involving fewer parameters. Thus non-linear finite element methods or other sophisticated models may not be a good choice for the assessment of MAB, while simplified approaches for example based on limit analysis principles are likely to be more appropriate. In this research. a holistic computational limit analysis procedure is presented which involves modelling both soil and masonry components explicitly. Masonry bridge parts are discretized using rigid blocks whilst the soil fill is discretized using deformable triangular elements and modelled a.'i a Mohr-Coulomb material with a tension cut-off. Lower and upper bound estimates of the collapse load are obtained. Results are compared with those from recently performed bridge tests carried out in collaboration with the University of Salford. A key project finding is that the use of peak soil strength parameters in limit analysis models is inappropriate when the soil is modelled explicitly. However, use of mobilized strengths appears to be a promising way forward, yielding much closer correlation with experimental data.

No description available.

Rock Engineering Systems are a collection of ideas, mathematical tools and computer technology all of which are designed to solve problems in rock engineering with interacting components. The interactions between components can be complex and the rock engineering problems themselves contain a high degree of uncertainty. The research described in this thesis investigates the incorporation of computational techniques known as parallel distributed processing methods into the disciplines of rock mechanics and rock engineering. Two main applications of parallel distributed processing methods in rock engineering are investigated in this thesis. 1) Multilayered perceptron artificial neural networks are used successfully to encapsulate the laboratory behaviour of rocks under triaxial compression. Trained artificial neural networks are then used to replace conventional constitutive models within finite difference geomechanical numerical modelling codes. 2) Two multilayered perceptron artificial neural networks are developed to assist in the task of discrimination of rock fracture presence within digital imagery of rock exposures. The first is trained using samples of the image that contain fracture image content and samples that do not, and provides a probability-like measure of fracture presence. It was sufficiently successful to permit estimation of fracture intensity parameter , . The second was developed specifically to identify fracture termination condition by matching samples to a set of fracture termination condition templates. Seven original contributions to the rock mechanics and rock engineering disciplines have resulted across the three application areas. These contributions are itemised, with details, at the beginning of the final Chapter of the thesis.

The study reported in this thesis forms part of a continuing programme of research on the performance of anchors and anchored supported structures being carried out at the University of Sheffield. In addition to a review of some of the recent laboratory studies, field observations and analytical investigations, the present study incorporates two main parts. In the first part, the failure mechanism and load carrying capacity of multi-plate anchors were investigated in a series of small scale studies. The study showed the difference in behaviour between single and multi-plate horizontal anchors and led to a better understanding of their failure mechanism. It also provided information for the design of the multi-plate anchors to be used in the second part of the research programme. In the second part, the behaviour of a 0.6 m high rigid retaining wall in a normally consolidated sand and supported by up to four rows of' anchors, was studied. Different design methods were employed to examine the overall stability of the wall-anchors-soil system. These were investigated experimentally in a series of tests in which field. construction procedure was simulated. After construction was completed, the - iii - retained backfill was subjected to surcharge loading in an attempt to approach failure. Normal earth pressure distribution on both sides of the wall, normal and shear components of the wall base reaction, anchor load changes, anchor movements, wall movements, sand surface subsidence and sand movements within the retained sand mass were monitored. The study showed the reliability of the different design methods and the effect of parameters such as anchor lengths and prestress loads on the overall behaviour of the systems.

The bulk of the rock material in the neighbourhood of an excavation, beneath a foundation or inside mine pillars is in fact triaxially loaded over long periods. Strata exposed in mines and excavations is subjected to bending over long periods also. In many mines and underground workings the rock is nearly saturated, also information on the three dimensional creep behaviour of any rock is very limited. It was felt that studying the three dimensional creep behaviour under confinement in both dry and saturated conditions could add valuable information to the knowledge available on rock behaviour in such conditions. Gypsum was chosen as a suitable evaporite rock for carrying out this work. To assess the significance and the nature of the time-dependent behaviour of gypsum in engineering applications, triaxial compression creep tests were carried out at 10,20 and 30 N/mm2 confining pressure. Bending and uniaxial compression creep tests were also performed on the rock. Axial and lateral creep strains were measured from which volumetric creep strain was calculated. All the creep tests were carried out in dry and water saturated conditions at room temperature. Instantaneous strengths of gypsum under bending, uniaxial tensile, uniaxial and triaxial compression were found. The effects of specimen size and water saturation on these strengths were determined at room temperature. Empirical equations describing the individual creep curves in bending, uniaxial compression and triaxial compression were determined for both dry and saturated conditions, in the latter two cases equations for both axial and lateral strains were-obtained.

Mudrocks (or stiff clays) cover much of south-east England. These materials were deposited from the Triassic to the Eocene in various marine environments. They have been submitted to complex sedimentation and post-sedimentation processes, resulting in the development of a wide range of micro and macro structures. Despite their engineering signi cance, advanced experimental investigations of their properties are limited. From 2007 to 2012, an Imperial College research project investigated the structure and mechanical properties of British mudrocks with the help of advanced laboratory testing. As part of this project, the research presented here focused on the stiffness and shear strength anisotropy of mudrocks in relation to their structure. Three UK mudrocks, deposited from the Jurassic to the Cretaceous, were selected for this project: Oxford clay, Kimmeridge clay and Gault clay. High quality block and rotary cored samples were tested in two Hollow Cylinder Apparatus (HCA), which control independently four stress parameters, among which the intermediate principal stress factor (b) and the orientation of the major principal stress (α ). The present study concentrated on the effects of α on the stiffness and strength properties. A set of six tests was performed on each soil: specimens were re- consolidated to their estimated in-situ stresses before being sheared undrained at various α values while controlling b. The results were interpreted together, taking account of an earlier comprehensive study on London Clay (Nishimura, 2006; Anh-Minh, 2007). The value was found to affect both the strength and stiffness properties, at various degrees for each mudrock. While the considered constitutive model influences the recorded degree of shear strength anisotropy, a good correlation was observed between the latter and the micro-structure anisotropy, as quantified by Wilkinson (2011)'s associated work. Qualitative assessment also revealed the strong influence of the macro- structure on the strength anisotropy.

The glacier regime is described and the classifications of glacier types are outlined. The modes of acquisition and transportation of the debris in glaciers is considered and the deposition process briefly considered. The various systems of classification of unstratified and stratified drift are given and the several morphological expressions of the unstratified drift are described. The crushing mechanism producing basal tills, of which the materials studied are examples, is reviewed in terms of their transportation environment in the glacier. The mechanical crushing of rocks is also considered and the particle size distribution laws developed to describe the products of this industrial process considered for application to glacial comminution products. The nature of glacial comminution is then further reassessed in terms of the shape of particles developed by the process. A full description of the locations of study sites and samples included in the investigations is then given including a detailed description of the local geological and glaciological environments where these are known. Previous investigations of the textural variations in tills are reviewed and the sampling and testing techniques for particle size distribution are outlined. The textural variations measured, or previously reported for Scottish, Norwegian, American and Canadian Pleistocene tills are analysed and compared to the measured variations in tills taken from three active modern glaciers in Norway, Iceland and Antarctica. These variations are then interpreted in terms of the mode of formation of the tills. Close correlations between these glacial products and industrial crushing products are thus shown to exist. Based on this, it is suggested that till variability can be much more easily understood if the material is treated as two size fractions mixed in variable proportions. The size fractions in any one till are then found to be fairly consistent with the mix proportions varying often very locally. Lithological variations are, however, found to have much less effect than might be expected. The organisation of the particulate matter in the tills is then considered in terms of the mode of deposition of the basal melt out tills. The orientation of clasts in the tills were measured in the field. This study was then extended in the laboratory using the contact goniometer to include the orientation of particles down to the sand size grade. Further qualitative studies using the Scanning Electron Miorosoope then extended the appreciation of the particle organisation down to the clay size particles. This attempt at determining the nature of particle organisation at all size levels was then proven to be most useful in determining the actual mechanisms operating during deposition. Essentially it was shown that the clasts react to the englacial stress field whereas the fines are organised during the melt out process and record in a more sensitive manner than the clasts, the stress history of the till during and after deposition. Some of the microfabrics in the tills were found to be very open and field data on in-situ density confirmed both the variability and the very low densities that can be found in the tills. Investigations of the factors influencing achieved density when recompacting tills showed that the variations in grading in these soils could be as important as differences in compactive effort. It is suggested that for engineering situations where quality control of compaction is required it is necessary to include gradational variations in the assessment procedures otherwise variations in compactive effort will be masked. A review of the factors affecting permeability and shear strength of soils was also undertaken with particular reference to the tills. Detailed investigation of some tills were undertaken and comparisons made to recorded data for tills used in Norwegian and Canadian dams. The interrelated influences of grading, density and particle arrangement were shown to critically affect permeability and shear strength and a system of classification based on percentage fines and the identity of the fines was suggested which attempted to include many of the variables. Limited success was met with respect to permeability but none with shear strength, the many other variables influencing the tills strength properties overriding the basic parameters used in the system. The use of percentage fines and the identification of the coarse and fine fractions in the tills was still however shown to be the best means of understanding the variations within anyone till. The inherent variations in the basal melt out tills were therefore shown to be attributable to their modes of formation and deposition and by recognising the fundamental nature of the till composition and particle organisation it is considered that a better understanding of the variations has been achieved.

Techniques for testing unsaturated soils have been investigated by the author where the measurement and control of parameters were undertaken directly. Suction was measured and controlled with a new high suction tensiometer and water content through mass measurements with a balance. These techniques have been used for the determination of soil water retention curves and for the development of a suction . control system using air circulation and water injection. The techniques allow the soil to be subject to the same drying and wetting conditions that occur in nature and avoid the need for elevated air pressures, as are traditionally involved in testing using the axis translation technique. The performance of the new high suction tensiometer was evaluated, followed by its applications to soil testing. The tensiometer performance focused on the factors controlling cavitation, calibration in the negative pressure range and measurement. It was found that isotropic unloading is the most accurate technique for calibration in the negative range and that axis translation techniques can lead to errors. The research confirms high suction tensiometers are easy to use and versatile devices. Techniques were developed to measure and control suction and water content in unconfined and confined samples. Research on the unconfined samples focused on the procedures to obtain the soil water retention curve: discrete (soil dried or wetted in stages) and continuous (soil drying continuously). While both procedures were found not to influence the curves significantly, it is demonstrated that the continuous procedure is sensitive to factors such as the exposed surface area to drying or wetting and so should be used carefully. For confined conditions, wetting, drying, and water content measurement systems were developed. Wetting was based on the injection of water; drying was based on air circulation through a desiccant within a closed loop system. Water content was determined from the. difference between water injected and that adsorbed by the desiccant. This has been applied as part of a· double cell triaxial testing system that allows continuous determination of suction, water content and volume change. A challenge of such a system was imposing an air tight environment. The suitability of environmental scanning electron microscopy to observe unsaturated soils at the particle level was explored. The imaging of micron-sized materials at different relative humidities allowed a series of observations previously undocumented, among them: water menisci were visible, including their shape and interaction with surfaces; the contact angle between the air-water and water-solid interfaces was measurable.