An understanding of the behaviour of newly deposited soil is important because of its direct applications in fields such as harbour siltation and storage of dredge slurries. In coastal regions the dominant mode for deposition of fine-grained cohesive material is through flocculation- the electrostatic aggregation of mud particles. The present study focuses on the measurement of both flocculation conditions and bed properties, towards the understanding of the influence of the former on the latter. A computerised laboratory setup, complete from sedimentation and flocculation to the long term consolidation allows each of these processes to be observed in instrumented perspex columns without the many variables present in an estuary. The bed deposition rates resemble those found in natural estuaries, and therefore this technique is a much improved method for examining natural processes than the method of slurry deposition, which is normally employed in soil mechanics. An image acquisition and analysis system has been designed specifically for this work. Image sequences are analysed in real time to give information about the dimensions and concentration of the particles, and the particle velocities. Floc sizes increase as the particle concentration increases in the column. Increasing the concentration further leads to an overall hindering of the sedimentation rate and breakup of the flocs. Images of beds formed through slow sedimentation clearly show aggregate features that are not present in slurry experiments. From low to medium sedimentation rates the bed height, normalised by the mass of sediment in the bed, increases. From medium to high rates of sedimentation, however, the normalised bed heights decrease. Overall slurry experiments have lower bed heights than slowly deposited experiments. Image analysis of the bed surface (top 0.5mm) has revealed that aggregates which are present immediately after deposition are broken down over tens of hours. The ISIS instrument has been modified to gather information about the resistance to erosion of the beds. A bed is more easily eroded after it has had a long period of consolidation. X-ray density measurements and imaging techniques are used to link the strength of the bed to biological factors. X-ray bulk density and pore pressure measurements allow calculations of void ratio, porosity, and effective stress. Bender element apparatus has been constructed to measure shear stiffness of the soil, and bed strength measurements are made using a shear vane. It is found that variations in the sedimentation conditions have significant effects on bed density and on void ratio. These effects continue to be visible in the consolidated bed, even after an order of magnitude increase in the total vertical bed stress (equivalent to the range of metres of overburden pressure). The variations in the bed structure can have profound effects on bed strength. Methods to compare floc measurements to bed measurements are complicated by the inherent variation in floc data. Possible techniques include the use of solids volume fraction, fractal dimension and intrinsic sediment properties. The usefulness of each of these is assessed. It is suggested that it is the frequency at which the bed is bombarded by flocs, and not the properties of the flocs themselves, that largely determines the bed properties.

Introduction: When fine grained natural sediments or artificially produced waste materials are transported and deposited through water, several different phases of behaviour are observed. These have been described variously as suspension, free settling, compression settling, intermediate and consolidating soil phases. Transitions between each are not always distinct in terms of material properties or behaviour and time spent in the early phases before a conventional soil state is attained can be a significant proportion of the total period of engineering interest. The eventual state attained following procession through these stages may be very different to that if the soil had been remoulded at the same final density. Standard engineering models exist which can describe soil behaviour well in separate phases under particular conditions, but these are of limited validity when extended to more general conditions and wider volumetric ranges than those for which they were formulated. The number of factors required to describe the entire range of behaviour is consequently larger than that for any one phase, and many of these factors are more familiar in fields of chemistry, geology or sedimentology than in classical soil mechanics. This thesis discusses, in engineering terms, the engineering behaviour observed in a particular soil during the general sedimentation and self weight consolidation process. In the second chapter existing knowledge about behaviour at zero or low stresses is reviewed and evaluated with respect to common assumptions made, often implicitly, in formulating predictive models. It is shown that while these models have been extensively developed to a stage where they can approximate many aspects of soil behaviour, the lack of fundamental investigations carried out in parallel with their development has often led to inadequate appreciation of the causes of discrepancies between modelled and real behaviour. This has occurred particularly where standard geotechnical testing equipment and methods devised for stiff soils have been used to obtain global average relationships between engineering parameters. Even where modified tests have been developed, instrumentation has sometimes been inadequate and measurements too infrequent, so that data available have necessarily been analysed only in terms of constitutive forms assumed already. In chapter three experimental techniques are proposed which, where possible will allow soil behaviour to be examined under the least restrictive conditions of one dimensional compression so that basic engineering concepts may be analysed. Chapter four describes the testing programme and presents direct results of experiments. Chapter five analyses compression behaviour and establishes some trends which can be observed for particular parameters and relationships, and which exist between experiments under different initial and boundary conditions. Similar analysis of strength behaviour is undertaken in chapter six, where results obtained using different testing methods are compared. In the final chapter the general relevance of these results and their implications for engineering problems are discussed. Some suggestions are made for future work. Areas of application Improved knowledge about cohesive waterborne sediments can result in considerable savings for related industries. In the United Kingdom the annual cost of maintenance dredging is £25m (I.C.E. Costal Engineering Research Panel, 1985). In East Coast ports alone reduction of the distances travelled by each dredger would lead to a saving of £270,000 per annum, per kilometre reduction. Studies at Rotterdam Europort (Kirby, Parker, van Oostrum, 1979) show that although a channel dredged recently may quickly refill with sediment to a depth which echo-sounding techniques might indicate to be unnavigable, the strength may be so low as to allow passage of vessels virtually unimpeded. A density of 1.2 Mg/m3 is now used by the Rijkswaterstaat to define the "Nautical Depth" of a channel, stated to be "a density within the suspension above whose altitude vessels can safely sail." Dredging control using information from gamma ray densimeters has enabled production increases of up to 50% to be obtained in the Europort area. In the United States $30m was spent in a 5 year period on a dredging research programme aimed at improving disposal methods (Haliburton, 1977). Considerable volumes of waste material are also produced by the mining industry. The phosphate industry in Florida produces 40 million tons by dry mass per annum at an initial 3% solids by mass which even after two years retains void ratios around 10, due to the high content of attapulgite, a clay mineral consisting of long fibrous particles with large specific surface. Disposal areas for these clays occupy over 50,000 acres and are surrounded by 300 miles of dams, posing significant environmental and safety problems (Bromwell, Oxford, 1977). Failures of underwater slopes have been well documented. In muds deposited recently in the Mississippi Delta area very low shear strengths combine with apparently high excess pore pressures and presence of gas bubbles to cause instability for slope angles less than 1°. Recent research carried out a Oxford suggests that presence of gas may cause high excess pressures to be deduced where none exist. Duncan and Buchignani (1973) analysed a slope failure in San Francisco Bay which occurred during cutting of a slope from a normally consolidated clayey silt. The importance of accurate determination of an in situ parameters for analysis was shown by the estimated saving of $200,000 through using a slope of 7:8 rather than 1:1, decreasing the supposed safety factor from 1.26 to 1.17. Analysis of error sources showed that an error of only 4% in the soil density could reduce this safety factor by 10%. Similar problems due to changes in loading or boundary conditions occur where natural changes, such as increase in water current, cause erosion of a sediment layer which might, for example, be supporting an underwater cable or pipeline. In all these areas in situ property determination in solid of low density provides major problems. Density is often the only quantity that can be measured both accurately and continuously and then only when a stable platform can be maintained. Recovery of high quality samples from these layers is virtually impossible, so that there is a strong need for correlations between density and other properties such as strength and compressibility.

The subject of this thesis is the verification of a three-dimensional numerical modelling approach for the prediction of settlement damage to masonry buildings due to tunnelling in soft ground. The modelling approach was developed by previous researchers at Oxford, and was applied to three sites, representative of a range of practical configurations. The first involved the excavation of a shaft close to the corner of an eighteenth century church in London. The second involved tunnelling with very low cover beneath the foundations of a terrace of cottages at Ramsgate, Kent. The third was the relatively well-known case of tunnelling beneath the Mansion House, London, for construction of the extension to the Docklands Light Railway in the late 1980’s. The overall conclusion of the project is that the modelling procedures are suitable for application to the detailed assessment of the response of buildings to tunnelling. Particular features of the procedures are that the building is modelled together with the ground and a representation of the tunnel excavation, and in three dimensions. It has been confirmed that all these features are necessary to model the building response, which may include a combination of shear deformation, arching and bending behaviour. Further lessons have been learned concerning the importance of the self-weight of the building in determining overall settlements, how to model openings such as doors and windows in façades, and whether it is necessary to model the building foundation. It has not proved possible, through lack of time, to model the advance of tunnels beneath buildings within this thesis. This, however, is observed to be an important effect in the field, particularly in causing damage to internal walls. It is recommended that further research be carried out in this area. This project made use of large-scale non-linear finite element analysis. The demand on computing resources was high, stimulating many enhancements to the software, the most important of which was parallelisation of the analysis program for use on the Oxford Supercomputer. To obtain optimum results, larger model sizes are required. The computing resources to enable this should become more commonly available within the next few years, enabling the modelling techniques to be used routinely.

The research described in this thesis focussed on the bond resistance of grouted soil nails in clay. Physical modelling took the form of large scale element tests in the laboratory of drilled and grouted nails in a stiff clay. Nails were installed under different boundary stresses; testing was conducted at different constant rates of pull-out and also under sustained load conditions. Observed behaviour was compared with that predicted by measurements of interface shear resistance obtained in a conventional shear box. Numerical modelling was carried out in an attempt to improve understanding of the effects of installation processes on nail performance, and to investigate the trends in behaviour observed during laboratory tests. For this purpose, a one-dimensional finite element computer program was developed to include the effects of consolidation and the out-of-plane soil displacements associated with nail axial loading. The physical modelling showed that the loading rate has a significant effect on bond resistance. This has consequences for the interpretation of data from constant rate of displacement pull-out testing, the conventional method of verifying bond resistance in the field. It is suggested that this type of testing is inappropriate in low permeability soils, because capillary suctions develop which lead to increased bond resistance. Results from laboratory sustained load tests show that lower values of bond resistance are mobilised under the static load conditions more likely to exist in a real soil nailed structure. The numerical modelling confirms the behaviour observed during the laboratory tests, and shows that the mechanisms by which bond is mobilised are complex, depending critically on the dilation and consolidation characteristics of the soil. Nail installation procedures are modelled, and grout pressures are shown to strongly influence bond resistance. Interface tests show trends in behaviour similar to those observed during pull-out testing. However, difficulties exist in the qualitative use of interface test data to predict nail bond resistance.

The quasi-static penetration of a cone penetrometer into clay can be formulated as a steady state problem by considering a steady flow of soil past a stationary cone. The soil velocities are estimated from the flow field of an inviscid fluid, and the incompressibility condition is achieved by adopting a stream function formulation. Emphasis is placed on obtaining an accurate velocity estimate and this is accomplished by a solution of the Navier-Stokes equations. The strain rates are evaluated from the flow field using a finite difference scheme. The clay is modelled as a homogeneous incompressible elastic-perfectly plastic material and the soil stresses are computed by integrating along streamlines from some initial stress state in the upstream region. These stresses do not in general obey the equilibrium equations, although one of the two equations can be satisfied by an appropriate choice of the mean stress. Several attempts have been made to use the remaining equilibrium equation to obtain an improved velocity estimate and three plausible iterative methods are detailed in this thesis. In a second study, a series of finite element calculations on the cone penetration problem is performed. In modelling the penetration process, the cone is introduced in a pre-formed hole and some initial stresses assumed in the soil, incremental displacements are then applied to the cone until a failure condition is reached. Although the equilibrium condition is satisfied very closely in the finite element calculations, it is extremely difficult to achieve a steady state solution. In a third series of computations, the stresses evaluated by the strain path method are used as the starting condition for the finite element analysis. This is believed to give the most realistic solution of the cone penetration problem because both the steady state and equilibrium conditions are approximately satisfied. Numerically derived cone factors are presented and these are found to depend on the rigidity index of the soil and the in situ stresses. The pore pressure distribution in the soil around the penetrometer is estimated using Henkel's empirical equation. The dissipation analysis is based on Terzaghi's uncoupled consolidation theory. The governing equation is formulated in the Alternating-Direction-Implicit finite difference scheme. This formulation is unconditionally stable and variable time steps are used to optimise the solution procedure. The dissipation curves are found to be significantly affected by the rigidity index of the soil and a dimensionless time factor is proposed to account for this effect.

This dissertation describes a study of the interaction between closely spaced tunnels during shield tunnel construction, concentrating on the study of the short-term incremental behaviour of the liner. Carefully controlled physical model tests were carried out and the test results were complemented by a limited amount of numerical analysis. In the physical model tests described in this dissertation, two groups of tests were carried out at a laboratory scale; one set of tests studied closely spaced parallel tunnels and the other set investigated perpendicular tunnels. An important feature of the study was that a novel model tunnelling machine was designed and developed as part of the research. Thin steel tubes were used to model the tunnel linings. The experimental technique adopted in the preparation of clay samples (which is a well-established procedure) was found to produce high quality samples. Good repeatability was achieved in preparing the kaolin samples. The tunnelling machine allowed tunnel liners to be installed using similar procedures to those adopted in the construction of full scale shield tunnels using an earth pressure balance approach. The instrumentation system used in this experimental programme are shown to produce reliable data. During the model tests measurements were made of liner strains, pore water pressures and total stresses acting on the liner. Errors in the data logging system were shown to be very small (of the order of less than 2% of peak values). The mechanisms governing the structural interaction between closely spaced tunnels are highly complex. The tunnel installation was shown to modify the stresses acting on the liner of the adjacent tunnel. These stress increments led, in turn, to line deformations and induced bending moments. The nature of the interaction mechanisms depends on the geometric configuration of the tunnels, the liner properties, and overconsolidation ratio. For the parallel tunnels, the pillar width ratio (W/D) is an important parameter governing the magnitude of the interaction effects. the interaction effects increase as the pillar width ratio is reduced. Increasing the liner flexibility was found to reduce the induced bending moments but to increase the induced displacements. The interaction effects were larger in overconsolidated clay than normally consolidated clay. The worst case for interaction effects occurs when the pillar width is small, the liner is flexible and the value of OCR is large. Three-dimensional considerations suggest that interaction between parallel tunnels may be more severe than those measured in the corresponding perpendicular tunnel tests. However, the different nature of the mechanisms in the two cases appear to be more significant than this geometric effect.

Wave loading on offshore structures founded on sand can result in partially drained response of the foundation soil. The characteristics of the rate of loading, the permeability of the soil, and the size of the foundation affect the degree of partial drainage. Partial drainage refers to situations where pore pressures develop in the soil, and the response of the soil is neither fully drained nor undrained. This thesis is concerned with the effects of loading rate, and consequent drainage, on the behaviour of a flat footing that is founded on the surface of a saturated sand base. The results of physical tests performed in the laboratory on a model-sized footing are reported. The footing was founded on oil-saturated fine sand and was subjected to combined loading. The effect of the vertical, horizontal, and rotational displacements are reported. The response of the footing is analysed in the context of existing drained foundation models that are based on work hardening plasticity theory. The rate dependency of the vertical load:deformation behaviour and the combined yield surfaces are described.

The research was intended to enhance the understanding of penetrometer behaviour in normally consolidated clay. The effects of varying penetration rate and clay shear strength were studies and the distribution of pore pressures determined. The meet these objectives laboratory testing was undertaken. Penetration tests were to be performed in tanks of clay consolidated from slurry and maintained under known stresses. An initial series of tests examined the effect of stiff tank walls. These showed unexpectedly high radial boundary stresses were generated with tank to probe diameter ratios of 50:1. This finding dictated laboratory tests with small scale penetrometers in the largest practical size of tank. Field tests with small scale and full size penetrometers demonstrated an absence of scale effect in the penetrometer pore pressures and total stress data. Two consolidation tanks of 580mm diameter and one of 1000mm diameter were built with the facility for maintaining constant stresses on top of the sample during penetration. Piezocone penetrometers of 5cm2 and 1cm2 cross-section were built. Penetration tests were performed from 2cm/s to 3m/s with hydraulic insertion equipment. At lower rates the total cone resistance factor Nkt was shown to be 10.3±0.9 in normally consolidated kaolin. The corresponding pore pressure factor NΔu was 8.25±1.0 at the cone shoulder. A hundredfold increase in the penetration rate increased the Nkt factor by 40% but the NΔu factor was unchanged. Various subsidiary points emerged. The type of strength test used for comparison with penetrometer data is significant. Major strength reductions occur on sample depressurization. These were demonstrated with vane and triaxial tests. Baligh and Levadoux’s method for determining ch from pore pressure dissipation around penetrometers matched experimental data. Ratios of excess pore pressures on the cone face and shoulder show some promise in the evaluation of OCR.

The application of an electric field to contaminated soil specimens and the resulting electrokinetic phenomena have been combined with the degradative actions of bacteria to create a novel method for the remediation of contaminated land. Currently, the vast majority of remediation projects involve disposing of contaminated soil in landfill. With the introduction of, and increases in, landfill tax in the UK, this option is becoming less desirable, and so robust, more sustainable techniques are required, such as bioremediation. A major problem with the implementation of bioremediation is the bioavailability of contaminating chemicals. Reduced accessibility of bacteria to the chemical(s) they are attempting to degrade can lead to significant increases in required remediation times, as well as the possibility of significant residual contamination, the levels of which cannot be easily further reduced. This thesis addresses the problem of bioavailability of contaminants in soils, and has investigated the use of electrokinetic phenomena as tools to bring about an increase in this factor, leading to improved biodegradation. Soil contaminated with pentachlorophenol (sodium salt) was subjected to an electric field in a number of experiments, with significant transport of the initially ionic chemical noted. The transport and fate of this chemical were tracked throughout each experiment, along with properties of the soil pore fluid. Significant changes in soil chemistry were noted (particularly pH or moisture content, depending on the experiment). The effect of pH change was found to be particularly important in this respect, with acidic conditions hindering both movement and bioavailability. A method of applying an electric field to contaminated soil containing degrading bacteria was developed which minimised the changes to these parameters within the soil. A significant increase in the effectiveness of the remediation was noted with this technique, with substantially faster degradation found to occur.

The in-situ mechanical properties of Huntly coal are investigated. Laboratory tests on core specimens and small scale in-situ pressuremeter testing are used to determine elastic properties and failure criteria of the coal and coal measure rocks. The in-situ stress field that exists in the Waikato basin and in particular in the Test Panel of the No 1 West Mine is determined. An instrument is developed especially to measure in-situ stress in coal. To check in-situ behaviour of the coal seam, a full scale underground roadway is excavated in the Test Panel. Stress changes and deformations are monitored from nearby drives allowing all the changes to be observed, even those that are normally missed because they occur ahead of excavation. Using elastic properties and in-situ stresses determined from small scale tests, the roadway excavation is mathematically modelled using the finite element technique. Stress changes and deformations thus calculated are compared to those measured in the full scale test. With evidence gained in the Test Panel investigation, general observations of tunnel behaviour elsewhere in the mine are investigated. Some problems that have beset the mine are identified and solutions are found.