Search

Large soil movements arising from the installation of jack-up spudcan foundations may induce detrimental effects on piles supporting adjacent permanent jacket platforms. Particularly, in a dense sand-over-clay profile, there is a risk of spudcan rapid penetration due to underlying clay bearing capacity failure, known as spudcan punch-through failure, which can cause further damage to piles in the proximity. This study investigates the foundation interaction between the spudcan and adjacent pile in a sand-overlying-clay soil profile by way of centrifuge modelling. Issues investigated include thickness of the upper sand layer with incidence of sand plug beneath the spudcan and spudcan–pile clearance. The experimental findings reveal that spudcan punch-through hazard and the magnitude of induced pile moments amplifies primarily due to the increase in the thickness of the upper sand. In addition, the magnitude of induced pile moment in a sand-over-clay profile is significantly higher than that in a single clay profile because of the post-peak bearing behaviour observed in stratified soil, yielding to a greater risk to pile integrity. Soil pressure profiles on the piles are also back-analysed. Two generalised induced soil pressure profiles are deduced, with the first profile representing the onset of spudcan punch-through and the second profile denoting the situation of post-punch-through failure.

The paper studies strike-slip fault rupture propagation through dense sand and its interaction with surface foundations, combining physical and numerical modelling. A series of centrifuge tests is conducted using a three-section split-box, which allows the modelling of two strike-slip faults per test. A free-field test is initially conducted, followed by four interaction tests. Eight different foundation configurations are studied, varying the foundation location, surcharge load, aspect ratio and rigidity. The experiments are numerically simulated employing three-dimensional finite-element modelling, combining periodic boundaries and a relatively simple yet efficient constitutive model, developed as part of this study. Based on a Mohr–Coulomb yield criterion, the model incorporates post-yield isotropic frictional hardening and softening (MC–HS). Carefully calibrated on the basis of triaxial tests, the model is validated against the centrifuge model tests, and exploited to derive further insights. The MC–HS model covers the entire range from elastic to fully softened response, capturing the deviatoric and volumetric behaviour of dense sand, and especially its pre-softening volumetric response, which is proven to be crucial for the simulation of the complex mechanisms of strike-slip faulting. Both physical and numerical modelling reveal the formation of diagonal shear ruptures at the ground surface (Riedel shears). These are complex helicoidal structures, formed due to the spatial variation of shear stresses. Foundation response is mainly governed by the kinematic constraint offered by its presence. Fault rupture locations close to its sides typically lead to a translational mechanism, whereas locations close to its centreline lead to a rotational one. Foundation rigidity is proven to be a prerequisite for the development of both mechanisms, which rely on the ability of the foundation to resist the developing normal and shear stresses.

This paper details a programme of centrifuge tests (100g) on the monotonic and cyclic lateral behaviour of large-diameter semi-rigid monopiles in kaolin clay. Piles of different diameters were used, with a maximum prototype value of 5·9 m. The pile–soil interaction was investigated in terms of the overall load–displacement behaviour, distribution of bending moment and pile deflection, p–y curves and excess pore pressure response. For monotonic loading, procedures were developed to fit the monotonic p–y curves obtained from different pile and soil properties using a hyperbolic tangent model. In addition, an explicit model was derived to assess the development of excess pore pressure within the surficial soil, which was found to approach the vertical effective soil stress at large displacement. For cyclic loading, the accumulation of lateral pile displacement could be accompanied by either increase or decrease in soil resistance, which strongly depends on the loading rate. Strong correlation between the development of peak soil resistance and a dimensionless velocity group was obtained, and was supported by the measured excess pore pressure. Under undrained conditions, the peak soil resistance decreases almost linearly with the accumulation of excess pore pressure.

Discretely spaced reinforced concrete (RC) energy pile rows have been proposed to be installed at the mid-height of soil slopes. Although the pile row can provide mechanical reinforcement to slopes, the piles can also potentially be used to (a) intercept solar energy from roadways for heat storage in the ground to mitigate extremely high carriageway temperatures and (b) to extract shallow geothermal energy for road surface de-icing. In this study, a series of centrifuge model tests was conducted to evaluate the shearing behaviour of unsaturated silt with and without reinforcement by conventional piles and energy piles. Three-dimensional finite-element coupled water–vapour–heat transport analysis was also performed to understand further the effects of pile heating on the responses of temperature and pore-water pressure of the soil. The measured and computed results revealed that the primary effect of pile heating was an increase in soil hydraulic conductivity due to the heat-induced reduction in water viscosity. The heated soil had enhanced water flow and hence developed higher suction. When subjected to translational slip, the silt reinforced by a row of closely spaced RC energy piles exhibited a more ductile shearing response and a lower peak shear resistance, compared to that reinforced by conventional piles. The peak bending moments mobilised in the energy piles in the stable stratum were also smaller. At larger shear displacements, however, the shear resistance converged to a similar value, regardless of the suction and temperature. These findings suggest that piles modified with additional energy transfer functionality can continue to act as reinforcement, potentially preventing soil from a sudden brittle failure, without attracting additional flexural stresses onto the piles.

This research provides an insight into the previously unexplored aspects of kinematic pile bending, especially for large-intensity earthquakes where the soil behaviour is highly non-linear. In this study, a series of dynamic centrifuge experiments was conducted on pile foundations embedded in a two-layered soil profile to investigate the kinematic effects on pile foundations during model earthquakes. A single pile and a closely spaced 3 × 1 row pile group were used as model pile foundations, and the soil model consisted of a soft clay underlain by dense sand. It was observed that the peak kinematic pile bending moment occurs slightly beneath the interface of the soil layers and this depth is larger for the pile group compared to a single pile. Also, the piles in a group attract lower bending moments but carry larger residual kinematic pile bending moments compared to a single pile. Furthermore, the elastic solutions available in the literature for estimating the kinematic pile bending moments are shown to yield satisfactory results only for small-intensity earthquakes, but vastly underestimate for large-intensity earthquakes, if methods are applied injudiciously. The importance of considering soil non-linearity effects and accurate determination of shear strain at the interface of layered soils during large-intensity earthquakes for a reliable assessment of kinematic pile bending moment from methods in the literature is demonstrated using dynamic centrifuge test data.

Seismic design of soil-nailed walls requires demonstrations of tolerable ranges of wall movements, especially when a surcharge load exists near the wall. In this study, the effect of surcharge location on seismically induced wall movements was investigated using four centrifuge tests. The axial tensile forces, developed along the soil nails during the seismic loadings, were also measured during the tests. At 50g centrifugal acceleration, model tests represented a 12 m high prototype wall reinforced with five rows of soil nails. To apply a surcharge stress of 30 kPa at the specified location relative to the wall for each model test, a rigid footing was placed on the soil surface. The model soil-nailed walls were subjected to three successive earthquake motions. Surprisingly, it was found that the model wall with the footing located behind the soil-nailed region experienced the largest seismic movements, even more than when the footing was directly behind the wall. Furthermore, the tests showed that the lower soil nails played a key role in the wall stability during earthquake shaking, acting as a pivot for the pre-collapse cases tested, whereas the upper soil nails needed to be sufficiently extended to contribute properly to the seismic stability of the wall.

This paper presents the results of centrifuge model tests and three-dimensional large-deformation finite-element analysis on axial pipe–soil sliding behaviour of partially embedded subsea pipelines. A three-dimensional re-meshing and interpolation technique with small strain approach was adopted using the Abaqus mesh-to-mesh solution mapping feature with the Cam Clay model. The parameters studied include soil properties, interface friction coefficient, embedment depth and post-lay reduction in vertical load. Embedment depth and post-lay load reduction were found to exercise the strongest influence on the post-consolidation and sliding behaviour. Empirical relationships correlating centrifuge experiments and numerical modelling are proposed for the embedment enhancement factor in terms of the embedment and vertical load ratios, under conditions of very slow sliding, corresponding to the drained condition. For very fast sliding, corresponding to effectively undrained conditions, similar relationships involving embedment and vertical load ratios and interface friction coefficients are also proposed. Finally, the variation of embedment enhancement factor with sliding velocity is expressed in terms of a relative enhancement index, Ψ, which can be adequately described by hyperbolic relationships for the complete range of consolidation history. These relationships provide a basis for the evaluation of the embedment enhancement factor corresponding to an arbitrary sliding velocity, which can be used in routine analysis and design of as-laid underwater pipelines in soft clayey soils.

Fault deformation can cause severe damage to structures located within the fault zones. A logarithmic spiral model was developed by D. A. Cole and P. V. Lade in 1984 to predict the shapes and locations of fault ruptures in sand caused by dip–slip faulting, based on 1g sandbox tests. However, it shows poor predictions for reverse fault ruptures observed in centrifuge tests. In this study, centrifuge modelling of reverse faulting was performed to explore the reasons for the poor predictions by the logarithmic spiral model. These faulting tests were conducted in free-field conditions with different sand densities and soil thicknesses. Based on results of centrifuge tests from both the present study and previous research studies, a hyperbolic spiral model is proposed to predict approximately the shapes and locations of reverse fault ruptures in sand. The results show that the 45° − ϕmax/2 rather than 45° − ψmax/2 controls the surface direction angle of the outcropping rupture. Validation of the hyperbolic spiral model shows a promising prediction in centrifuge tests both in the present study and previous research studies.

This paper presents results from a systematic investigation of footing response in four different types of sands placed and tested under controlled conditions in a geotechnical centrifuge. A parallel series of cone penetration tests (CPTs) was performed to examine correlations between CPT end resistance (qc) and footing bearing pressure (q). The experimental results show that there is a unique relationship between the normalised bearing stress (q/qc) and the settlement normalised by footing diameter (s/B), irrespective of footing diameter, stress level, sand grading, mineralogy and angularity. These results are shown to compare well with findings from published databases of field-scale footing tests when allowance is made for the significant effects of creep. The study enables simple and robust recommendations to be established for the ultimate capacity (defined at s/B = 0·1) and allowable bearing pressure of a footing on sand.

This paper details analysis of deformation behaviour of silica and carbonate sands under a rectangular foundation subject to uniaxial vertical load based on results from a series of centrifuge model tests. A multiscale particle image velocimetry/digital image correlation (PIV/DIC) technique was used to record and analyse the foundation tests with high resolution and measurement precision. Cone penetrometer and pressuremeter tests provide in situ soil characterisation of the tested sand sample in the centrifuge environment. The soil behaviour is analysed through foundation load–settlement response and the observed soil deformation measurements. Different soil deformation mechanisms and strain behaviours were observed in the different sands tested, and the particle shape effect is considered, with data from scanning electron microscopy, to explain the differences. The results and analyses contribute towards better understanding of different soil behaviours under shallow foundations in different sands.

The present study refers to the effects of reverse faulting in dry sand on the mechanical response of a tunnel lining. Seven centrifuge tests (five tests with tunnel and fault rupture and two tests without tunnel and with fault rupture as free field) were performed in this research and finite-element models are utilised to gain further insight into the interaction between reverse faulting and the tunnel. The effects of tunnel location and tunnel rigidity are investigated based on the experimental results. Results show that the tunnel location with respect to the location of the fault tip and the rigidity of the tunnel are important parameters affecting the mechanical response of the tunnel lining. The centrifuge tests, in prototype scale, were simulated numerically using a two-dimensional explicit finite-element model in a plane-strain condition. Good agreement between the numerical analysis and centrifuge test results was obtained, indicating proper numerical modelling. A numerical parametric study was then conducted to investigate the safe zone around the fault region for tunnel construction. Finally, it was concluded that the internal forces in the tunnel lining are highly affected by soil stress redistribution caused by faulting.

This paper investigates the seismic performance of rooted granular slopes using dynamic finite-element analysis, validated against recently published centrifuge test data. The importance of selecting suitable strength parameters to represent soil response within a strain-hardening constitutive model was demonstrated and the simulations suggested that any boundary effects introduced through the use of the equivalent shear beam container in the centrifuge are negligible and can be represented by a semi-infinite lateral boundary condition. Using the validated model, a parametric study investigated the effects of different rooted soil properties on the performance of slopes of different heights. Vegetation was effective in reducing deformations at the crest of slopes of modest height, although the benefit reduced as slope height or soil apparent cohesion increased. The effectiveness was significantly affected by the extent of the root system, but only moderately sensitive to root cohesion, and insensitive to stiffness or damping of the rooted soil. Plant species possessing deep and extensive root systems are therefore recommended for seismic stabilisation rather than those with the strongest roots. For modelling purposes, it is sufficient to be able to quantify only the strength of the rooted soil and its area of influence. The magnitude of improvement from vegetation in terms of decreasing seismic permanent slip was also found to be insensitive to the construction method used (i.e. compacted/uncompacted embankment or cutting) for drained granular slopes.

This paper reports a forensic investigation on the unexpected collapse of a 13-storey high-rise residential building in Shanghai soft clay. Prior to the building collapse, a shallow excavation was carried out on one side of the building while a slope was filled on the opposite side. Despite the observed breakage of all the piles in the field, a preliminary forensic numerical analysis adopting a typical undrained shear strength (cu) profile in Shanghai suggests that the piles are very unlikely to break. The surprising contrast has led to this detailed forensic study on the underlying mechanisms for the building collapse, through a site-specific characterisation of the cu profile, three-dimensional centrifuge modelling and numerical back-analysis. Although the typical cu profile in Shanghai increases almost linearly with depth, the clay in this site consists of two major layers with distinctive strength – that is, a relatively soft clay layer (average cu = 23 kPa, thickness = 13·5 m) underlain by a relatively stiff clay layer (average cu = 58 kPa, thickness = 18·2 m). By reproducing this cu profile in the centrifuge test and numerical analysis, the resulting direction of the building collapse and the depths of pile breakage were broadly consistent with those observed in the field. It was found that the surcharge of the fill slope caused the bearing capacity to be exceeded in the upper soft clay layer. This led to the development of a sliding wedge under the surcharge, which extended horizontally (at a depth of 11 m) across the pile foundations to reach the excavation side, and therefore the breakage of the piles. If a typical strength profile in Shanghai (with cu linearly increasing with depth) were adopted, the failure mechanism could be significantly altered from the aforementioned deep slip failure to a shallow passive failure, causing much smaller bending moment in each pile.

The rocking response of embedded foundations in sand is studied, combining centrifuge modelling and numerical analysis. In total, 51 centrifuge model tests are conducted at ETH Zurich, varying the embedment ratio D/B and the factor of safety against vertical loading (bearing capacity), Fs. The experimental results are used to validate three-dimensional finite-element models, which are subsequently employed for a parametric study. The initial two phases of non-linear response are studied, namely quasi-elastic and non-linear response, while the third phase (plastic response) will be examined in a forthcoming publication. Regarding small-strain rocking stiffness, Kr (first phase), an extended formula is proposed, accounting for soil inhomogeneity along the embedded sidewalls. Kr is shown to decrease with Fs, due to initial soil yielding from vertical loading. The formula is further extended to account for the role of Fs, offering improved estimation of small-strain rocking stiffness. With respect to the non-linear response (second phase), it is shown that the degradation curves of rocking stiffness with rotation can become approximately dimensionless, if the rotation θ is normalised to a characteristic rotation, θs. The increase of D/B leads to a reduction of the normalised rotation θ/θs at which the non-linear response is initiated, due to the increasing participation of sidewall friction.

This note examines the seismic response of a pile–raft system using the random field finite-element method. The numerical method is validated by recently published centrifuge test results. The randomness in the maximum shear modulus of soils in the system is considered by three-dimensional random fields in order to investigate the statistical characteristics of the maximum pile bending moment response. The computed pile bending moment results indicate that the deterministic results can be linked up with the random analysis results by an amplification factor. Therefore, a reliability-based geotechnical design can be applied.

Punch-through failure of mobile jack-up rigs is still a main concern for spudcan foundations. Using centrifuge model tests and numerical modelling, this paper examines the mitigating effect of top-mounted sleeves and a downward skirt on spudcan punch-through in medium dense sand overlying clay. The findings show that the top-mounted sleeves and downward skirt are both useful separately for reducing punch-through distance. The top-mounted sleeve functions by interfering with soil backfilling above the spudcan and maintaining confinement to the sides. This results in an earlier recovery in penetration resistance and reduces the peak-to-trough drop in resistance. The downward skirt appears to mobilise the soil further ahead of the spudcan, which reduces the peak penetration resistance and increases the rate of post-trough recovery in penetration resistance. A combination of a top-mounted sleeve with a downward skirt appears to confer a smaller peak-to-trough drop in resistance and higher rate of post-peak recovery, and thereby a much smaller punch-through distance. The results would suggest that, compared with the downward skirt, the top-mounted square sleeve with a similar area ratio to the existing lattice leg is supposed to be a more effective option in mitigating spudcan punch-through failure.

A series of lateral tests conducted in a centrifuge on displacement piles in kaolin is used to examine the effects of pile shape, pile end-condition and clay overconsolidation on the lateral load transfer (p–y) curves in soft clay. These experimental tests are supported by finite-element analyses, which examine the responses of the circular, square and H pile sections used in the centrifuge tests. The experimental and numerical results reveal an important effect of pile shape on p–y curves, but no discernible effect of the pile end condition. These results also indicate comparable dependencies of net ultimate pressures on normalised depth and undrained shear strength. The findings enable development of a new formulation for p–y curves in soft clay, which incorporates the effects of pile shape and uses the triaxial compression consolidated undrained shear strength as the reference undrained strength.

A newly developed sequential limit analysis (SLA) technique is used to perform large-displacement numerical simulations relevant to thermally induced lateral buckling of untrenched subsea pipelines. A rigid plane-strain pipe segment, partially embedded in undrained clay, is subjected to cyclic lateral displacements with amplitudes of up to eight pipe diameters. A constant vertical dead load is applied to the pipe during each lateral sweep, but in some analyses this load is varied from one sweep to the next. The SLA method directly models the evolution of the soil surface profile, including active and dormant berms, and incorporates strain softening behaviour caused by soil remoulding. Comparisons of the numerical results with published centrifuge model test data, for a range of loading cases, are provided in terms of the pipe invert trajectory and lateral soil resistance. There is good overall agreement between the numerical and experimental results, demonstrating the suitability of SLA for solving such problems. Detailed aspects of the cyclic loading behaviour are discussed with reference to soil failure mechanisms and bearing capacity failure envelopes for combined vertical and horizontal loading. Finally, two brief parametric studies are used to explore the effects of the initial pipe embedment and vertical loading history on the subsequent lateral loading behaviour.

This paper presents an investigation into the influence of slope height on the role of vegetation to improve seismic slope stability. Dynamic centrifuge modelling was used to test six slope models with identical soil properties and model slope geometry within different centrifugal acceleration fields (10g and 30g, respectively) representing 1:10 and 1:30 scale slopes, that is, slopes of different height at prototype scale. A three-dimensional (3D) root cluster analogue representing a tap-root system, with root area ratio, root distribution and root length representative of a 1:10 and 1:30 scale tree root cluster (of rooting depth 1·5 m at prototype scale) was modelled using 3D printing techniques. A sequence of earthquake ground motions was applied to each model. The influences of filtering out low-frequency components of the earthquake motion, such as was necessitated at the lowest scaling factor owing to the practical limitations of the earthquake simulator, on dynamic amplification of motions within the slopes and the seismically induced slip, were first revealed. Subsequently, the effects of slope height on acceleration and deformation response of vegetated slopes were illustrated. It was found that the beneficial effects of roots on improving the seismic performance varied with the height of the slope. As an individual engineering technique for slope stabilisation, root reinforcement will not be such an effective solution for taller slopes, and complementary hard engineering methods (e.g. piles, retaining walls) will be necessary. For slopes of smaller heights (e.g. low-height embankments along transport infrastructure), however, vegetation appears to represent a highly effective method of reducing seismic slip.

This paper is concerned with the vertical penetration resistance of conical spudcan and flat footings in layered soils. Centrifuge tests are reported for a clay bed with strength increasing with depth interbedded with dense and medium dense sand. Both non-visualising (full-model) and visualising (half-model) tests were conducted with high-quality digital images captured and analysed using the particle image velocimetry technique for the latter. The load–displacement curves often show a reduction in resistance on passing through the sand layers, which creates a risk of punch-through failure for the foundations when supporting a jack-up drilling unit. For a given foundation, the peak punch-through capacity (qpeak) is dependent on the thickness of both the overlying clay and the sand layer. The failure mechanism associated with the peak resistance in the sand layer involves entrapment of a thin band of top clay above the sand layer that subsequently shears along an inclined failure surface before being pushed into the underlying clay. The top clay height when normalised by the foundation diameter affects the soil failure pattern in this layer and along with the sand layer thickness controls the severity of the punch-through failure (i.e. the additional penetration before the resistance returns to the peak value). Comparisons are made with current industry guidelines for predicting qpeak and the risk of punch-through failure for sand overlying clay. These methods are shown to be conservative in their prediction of qpeak but inconsistent in predicting punch-through.

The paper presents results from a centrifuge-scale study of the lateral response of piles in variably cemented sands. These sands were characterised using in situ cone penetration tests (CPTs) as well as laboratory element tests. Experimental results and numerical back-analysis of the centrifuge tests confirm that the lateral ground stiffness and capacity correlate well with the CPT end resistance (qc). The centrifuge data are combined with measurements made in other field-scale tests obtained in a strongly cemented sand, enabling development of a CPT-based formulation for load transfer curves that is consistent with observations and numerical analyses. The formulation is shown to provide reasonable predictions of lateral load response in a number of cemented sands with qc values in the range 2–50 MPa.

A series of centrifuge tests were performed on a 4 × 3 pile–raft system embedded in soft kaolin clay to investigate the pile bending moment and raft acceleration response under seismic excitation. Two types of far-field ground motions were employed, these being representative of the short- and long-duration ground motions arising from the Great Sumatran Fault and the Sunda Subduction Trench, respectively. The influences of the peak base acceleration, pile type and raft mass on raft acceleration and pile bending moment responses were studied in the centrifuge tests. Furthermore, based on the present and previous relevant centrifuge test results, correlations were derived to respectively estimate the peak raft acceleration and the maximum bending moment generated near the pile head during the shaking events. Despite some scatter in these correlations, they can be useful in engineering design practice for obtaining quick first-order estimates of the seismic response of pile–raft systems embedded in soft soils.

This paper presents a comparison of extraction resistance between spudcan and skirted foundations in sand-over-clay deposits. A series of centrifuge model tests was undertaken on a spudcan and two skirted foundations, varying the length of the skirt relative to the foundation diameter. The thickness of the sand layer was also varied to cover a range of practical interest. It was found that the ultimate breakout forces of the spudcan and ‘sealed’ skirted foundations in the sand-over-clay deposits are lower than that of a spudcan in a single-layer clay deposit. On the sand-over-clay deposits, when the skirted foundations are sealed, the breakout forces are 60∼61% higher than that of the spudcan. However, when the skirted foundations are vented, this difference becomes minimal, regardless of the skirt length and the effective sand layer thickness beneath the foundation, with the height of the sand plug left in the clay layer being similar to the thickness of the sand layer.

For calculating spudcan penetration resistance in an interbedded sand layer between two clay layers, the recently developed mechanism-based design methods for surface sand-over-clay deposits are extended to account for the influence of the overlying clay layer. A series of large-deformation finite-element (LDFE) analyses were carried out, varying layer geometries and strength parameters. The numerical results and existing centrifuge test data are used to calibrate the extended design methods. The performance of the modified design methods and those suggested in the ISO standard 19905-1 is compared, confirming the accuracy of the former methods in estimating the peak penetration resistance in the sand layer. LDFE analyses were also carried out adding a strong fourth layer, with the corresponding effects on the increasing peak resistance in the second (sand) layer and squeezing behaviour (or limiting squeezing depth) in the third (soft clay) layer discussed. Design formulas are also developed to estimate th...

The paper presents a series of centrifuge tests simulating a level site consisting of granular soil deposits subjected to various biaxial and uniaxial base excitations. The tests were conducted at RPI NEES centrifuge facility to assess the dynamic response characteristics of level deposits under multidirectional shaking. Synthetic sinusoidal waves were used as base excitations to test loose models under biaxial and uniaxial shaking. Dense arrays of accelerometers were used to monitor the deposit response along with pore water pressure transducers. Two uniaxial tests and one biaxial shaking test were conducted on three similar soil models to study the impact of multidirectional shaking on the generation of soil liquefaction. The two uniaxial shaking tests consisted of: (a) a test with input energy content using an Arias intensity similar to that of the biaxial input shaking and (b) a test with 10% increase in one of the components of the biaxial shake amplitude, as commonly done in practice for uniaxial simulation of multidirectional field shaking. The observed acceleration and pore pressure are used along with non-parametric identification procedures to estimate the corresponding dynamic shear stress–strain histories. The measured results along with the stress and strain histories obtained are used for two purposes: first, to shed light on the properties of liquefaction occurring through the stratum under uniaxial shaking; second, to show the difference in soil behaviour when it is subjected to biaxial shaking. The latter purpose is evident in the strain energy generated in the biaxial test compared to that of equivalent and traditional uniaxial tests.

In practice, laterally loaded piles are most often analysed using a ‘beam-on-non-linear-Winkler-foundation’ approach, whereby the soil–structure interaction is modelled by means of p–y curves. Although well-calibrated p–y curves exist for non-liquefied soils (e.g. soft clay and sand), the profession still lacks reliable p–y curves for liquefied soils. In fact, the latter should be consistent with the observed strain-stiffening behaviour exhibited by liquefied samples in both element and physical model tests. It is recognised that this behaviour is induced by the tendency of the liquefied soil to dilate upon undrained shearing, which ultimately results in a gradual decrease in excess pore pressure, and consequent increase in stiffness and strength. The aim of this paper is twofold. First, it proposes an easy-to-use empirical model for constructing stress–strain relationships for liquefied soils. This only requires three soil parameters which can conveniently be determined by means of laboratory tests. Second, it introduces a method for the construction of p–y curves for liquefiable soils from the proposed stress–strain model, based on the scaling of stress and strain into compatible soil reaction p and pile deflection y, respectively. The scaling factors for stress and strain are computed following an energy-based approach that is analogous to the upper-bound method used in classical plasticity theory. To validate the proposed p–y curves, results from a series of centrifuge tests are employed to back-calculate p–y curves for liquefied soils. The latter are compared with those obtained from the proposed method and the conventional p-multiplier approach.

Steel catenary risers are pipelines that convey fluids from the seabed to floating structures. The stiffness of the pipe–seabed response, which is the ratio between soil resistance and pipe embedment, in the touchdown zone strongly affects the fatigue accumulation rate, so is an important design parameter. This paper reports a centrifuge modelling study into the long-term pipe–seabed interaction forces on soft clay seabeds, with tests representing many months of behaviour at prototype scale. The results show that the penetration and extraction resistance during large-amplitude cycles degrades during the initial few tens of cycles, in the same way that cyclic penetrometer tests capture the fall in soil strength from the intact to the remoulded state. Calculations using bearing capacity factors for a cylinder provide good predictions of this response, although if the cycles of movement involve the pipe breaking away from the soil then the resistance reduces by more than the ratio of intact to remoulded strength, and this is attributed to entrainment of water in the soil around the pipe. However, with further cycles, as pore pressure dissipation occurs, the seabed stiffness recovers due to the gain in soil strength from consolidation. Eventually, the remoulding and water entrainment effects are wholly erased, and the stiffness exceeds the initial state. These observations suggest that current design practice – which factors down the soil stiffness to represent the influence of the cyclic degradation and remoulding process – may overlook a significant effect that raises the seabed stiffness, and potentially also reduces the fatigue life.

When a jack-up rig is installed in close proximity to an existing piled jacket platform, the massive soil movements induced by spudcan penetration may damage the adjacent piles. In 2012 the authors reported the findings of an extensive centrifuge model study on spudcan–pile interaction in soft clay. As a seabed profile may consist of soft clay over stiff soils and the spudcan–pile interaction phenomenon in such case is expected to be significantly different from that in soft clay, the earlier study is extended to investigate the interaction of a spudcan with adjacent piles in a soft clay overlying sand profile. Besides centrifuge tests on model spudcans, tests on half-spudcans placed behind the transparent window of the model container have been conducted to evaluate the soil movements and failure mechanism for various model configurations and soil conditions. The experimental results and findings are presented in this paper.

Jack-up installation is typically discontinuous with seawater ballast being taken on and shed again periodically to elevate the hull out of the ocean as the spudcan footings penetrate the soil. Any pauses in the footing penetration provide the opportunity for consolidation to occur in sufficiently permeable cohesive soils. The undrained shear strength underneath the spudcan is increased during consolidation, which enhances the penetration resistance within a limited extent immediately after the consolidation. This phenomenon has been explored in a limited number of centrifuge tests, but cannot be reproduced using conventional finite-element methods. An effective stress large-deformation finite-element approach based on periodic mesh regeneration is employed to capture the entire process of ‘penetration–consolidation–penetration’, with the modified Cam-clay model being incorporated to represent the behaviour of normally consolidated kaolin clays. The numerical model established is verified by comparison with two series of centrifuge tests before exploring the influence of key parameters, including consolidation depth, consolidation duration, coefficient of consolidations, anisotropy of permeability and loads held during consolidation. The normalised consolidation duration is identified as the dominant factor that affects the post-consolidation peak in penetration resistance. A simple method is proposed to estimate the post-consolidation penetration resistance.

Soil scour around a shallowly embedded pile can significantly compromise its lateral response, reducing both stiffness and capacity. Estimation of the lateral pile response must take into account both the scour-hole geometry and the overconsolidation effects on the remaining soil. A series of centrifuge model tests with various scour profiles were conducted at a scale of 1:250 to investigate the effects of both local and general scour on the response of a laterally loaded pile. Measured pile moment distributions and force–displacement data at the pile head were used to derive p–y curves quantifying the lateral pile–soil interaction. The p–y curves derived from various scour profiles were compared for equivalent depths below the new scour base, and below the original soil surface. For the general scour cases, the p–y curves for given depths below the post-scour surface are essentially identical to those at the same depths without scour. In contrast, for the local scour cases, the p–y response at a given depth below the scour-hole base is much stiffer than at the same depth below the original soil surface. As a practical approach to evaluate the effects of scour, the concept of an effective soil depth is introduced to determine the corresponding p–y curves for shallowly embedded piles in sand.

The use of plate anchors in permanent moorings for deep-water floating structures requires a thorough understanding of their performance under sustained and cyclic loading. This paper addresses the former aspect through a series of centrifuge tests on vertically loaded plate anchors in normally consolidated kaolin clay, supported by large-deformation finite-element analyses. The centrifuge experiments, which included particle image velocimetry measurements to identify the conditions leading to anchor failure, demonstrate that sustained loading can be maintained indefinitely if the anchor displacement at application of the sustained loading does not exceed 40% of the displacement reached at monotonic failure. This corresponds to a sustained load that is about 88% of the monotonic capacity, as identified from the numerical analyses. An important observation from both the centrifuge tests and the numerical analyses was a gap at the base of the anchor that developed under moderate load levels when the vertical effective stress at the base of the anchor reached zero. Although this so-called breakaway condition is generally associated with a lower capacity factor and a heightened potential for anchor failure, the conditions leading to breakaway also allow for consolidation-induced strength increase at the top face of the anchor and the mobilisation of partially drained or drained soil strength, both of which maintain anchor stability.

Liquefied ground loses its stiffness and strength, and it flows laterally in sloping areas and backfill areas behind collapsed quay walls. A method of installing piles such as cement-treated ones into the ground to prevent the lateral flow has been proposed and studied. The authors are investigating the most effective arrangement of piles to optimise cost-effectiveness. It is proposed to irregularly displace the piles to prevent lateral flow in various directions. In the present study, centrifuge model tests were conducted to verify the efficacy of piles against the lateral flow of liquefied ground and the effect of the pile arrangement in detail. This study first demonstrated the similarity law for the lateral flow of liquefied ground in a centrifuge, where both the Froude and Reynolds numbers can be satisfied. Numerical analyses assuming the flow of liquefied ground to be fluid-like were conducted to consider the improvement mechanism and the effect of the pile arrangement. The results of the model tests and numerical analyses indicated that the improvement by piles dramatically reduced the lateral displacement and that the average total flow velocity was small in the irregular arrangement case.

The strength of soil improved by deep mixing processes often exhibits significant spatial variability. This paper presents a statistical model for strength distribution in deep-mixed columns, based on the assumption that spatial variation in strength arises from spatial variation in binder mass fraction. The statistical inputs of the model are the probability density function and coefficient of variation of the binder distribution, both of which were obtained from centrifuge model data and were assumed to be independent of the amount of the binder introduced into the ground. The theoretical framework of the model is first developed, followed by an assessment of the probability density function for the binder distribution. The variation of binder mass fraction within a deep mixing column is fitted by a truncated normal distribution. The model is benchmarked against core strength data from three phases of deep mixing operation in a construction project. The simplicity of the model precludes consideration of the effects of detailed mixing blade configuration, in situ variation in soil properties as well as workmanship. In spite of this, the computed histograms of strength distribution gave a reasonable reflection of the measured strength distribution. This suggests that despite the variety and complexity of deep mixing processes, one may still be able to develop a simple and workable model of the strength distribution taking into account just a few important factors.

Relative ground movement represents a severe seismic hazard to pipelines crossing faults, and the maximum expected pipeline strains are the primary design concern. Past research has documented how stiff steel pipelines respond to permanent ground deformation and has produced calibrated empirical models of pipeline response to strike-slip and normal faulting. However, what little data exist on flexible pipelines illustrate how ‘stiff pipeline’ soil reaction models significantly overestimate peak strains for flexible pipelines. Results for four centrifuge tests conducted on model pipelines are presented to quantify the flexural response of pipelines to normal faulting over a wide range of pipeline stiffness. Continuous functions of pipeline and soil displacement calculated using digital image correlation are used to assess pipeline curvature, soil reaction and relative displacement. Comparisons of peak measured curvatures to current design analysis methods confirm that empirical stiff pipeline soil reactions progressively overestimate maximum curvature as the pipeline's stiffness decreases. Two strategies to modify current empirical soil reactions to account for flexible pipelines were then investigated, with the strategy of reducing the stiffness of the empirical soil reaction (while maintaining the advantage of a simplified, abrupt, step-like fault) providing better outcomes than modification to account for more realistic ground displacement profiles.

Many natural and man-made slopes are planted with vegetation, and it is known that this can increase the stability of slopes under static conditions. There is anecdotal evidence that vegetated slopes also perform better than fallow slopes during earthquakes. However, the study of the dynamic behaviour of slopes planted with species having dichotomous (‘woody’) roots is relatively rare owing to the extreme expense and difficulty involved in conducting full-scale dynamic testing on shrubs and trees. In this paper, dynamic centrifuge testing and supporting numerical modelling have been conducted to study this problem. In the centrifuge modelling, ABS plastic rods are used to simulate repeatably the mechanical properties of real roots. The numerical modelling work consisted of two parts. First, a computationally-efficient beam-on-non-linear-Winkler-foundation (BNWF) model using existing p–y formulations from piling engineering was employed to produce a macro-element describing the individual root and soil interaction both pre- and post-failure. By adding contributions from the different root analogues of different diameters, smeared continuum properties were derived that could be included in a fully dynamic, plane-strain continuum, finite-element model in a straightforward way. The BNWF approach was validated against large direct shear tests having stress conditions simulating those in the centrifuge at different potential slip plane depths. The conversion to smeared properties for global time-history analysis of the slope was validated by comparing the continuum finite-element results with the centrifuge test data in terms of both the dynamic response and permanent deformations at the crest, and these demonstrated good agreement. Owing to the simplicity of the BNWF approach and its ability to consider variable root geometries and properties, along with variation of soil properties with depth, it is suggested that the validated approach described will be useful in linking individual root–soil interaction characteristics (root strength and stiffness, diameter variation, root spacing and so on) to global slope behaviour.

Centrifuge tests investigating the effect of compressive preloading and consolidation on the subsequent uplift capacity of shallow skirted foundations that rest on lightly over-consolidated kaolin clay are reported. The uplift resistance, total and pore pressures at the foundation invert were monitored during the installation, consolidation and uplift. The results show that uplift performed immediately after the preloading generates a lower capacity due to the remoulding of the clay and reduction in soil strengths that result from the application of preloading. The soil strength can be recovered and eventually enhanced by the subsequent consolidation process. A theoretical framework, where the variation of the soil operative strength with preloading and consolidation is interpreted using the critical state concept, is proposed to explain and justify the experimental observations. The framework is validated through retrospective prediction of the uplift resistance of the centrifuge model tests with good agreement for a wide range of levels of preloading and degrees of consolidation demonstrated.

Penetration problems are important in many areas of geotechnical engineering, such as the prediction of pile capacity and interpretation of in situ test data. The cone penetration test is a proven method for evaluating soil properties, yet relatively little research has been conducted to understand the effect of soil layering on penetrometer readings. This paper focuses on the penetration of a probe within layered soils and investigates the layered soil effects on both penetration resistance and soil deformation. A series of centrifuge tests was performed in layered configurations of silica sand with varying relative density in a 180° axisymmetric model container. The tests allowed for the use of a half-probe for observation of the induced soil deformation through a poly(methyl methacrylate) window as well as a full-probe for measurement of penetration resistance within the central area of the container. The variations of penetration resistance and soil deformation characteristics as they relate to penetration depth, soil density and soil layering are examined. The results of deformation are also compared with previous experimental data to examine the effect of the axisymmetric condition. The effects of soil layering on both resistance and soil deformation are shown to be dependent on the relative properties between soil layers.

In this paper the face stability of shallow tunnels in partially saturated soil is investigated by means of centrifuge testing and numerical analysis. A loess soil is chosen for this study and its properties are characterised by laboratory testing. Centrifuge model tests are carried out at different soil densities and with different depths of water table. The numerical analyses are performed with an advanced constitutive model for partially saturated soil. The constitutive model is implemented in a finite-element code using the Bishop's effective stress. The model parameters are calibrated based on drained triaxial compression tests on partially saturated soil. The wetting collapse of the tunnel face is simulated by imposing a zero pore pressure condition at the ground surface. The numerical results are compared with the centrifuge tests.

Structure–soil–structure interaction between adjacent structures, which may occur in densely populated urban areas, has received little attention compared to the soil–structure interaction of single isolated structures. Additionally, recent earthquakes in/near such areas (e.g. the Christchurch series, 2010–2011) have shown that large motions can be followed by strong aftershocks. In this paper, the seismic behaviour of isolated structures and pairs of adjacent structures under a sequence of strong ground motions has been investigated using a combination of centrifuge and finite-element modelling. The latter utilised an advanced constitutive model that can be parameterised from routine test data, making it suitable for use in routine design. The finite-element models were shown to accurately simulate the centrifuge-measured response (in terms of surface ground motion and structural sway, settlement and rotation) even after multiple strong aftershocks, so long as the buildings' initial conditions were reproduced accurately. For the case of a building structure with a close neighbour, structural drift and co-seismic settlement could be reduced or increased as a result of structure–soil–structure interaction, depending chiefly on the properties of the adjacent structure. This suggests that careful arrangement of adjacent structures and specification of their properties could be used to control the effects of structure–soil–structure interaction. In all cases where adjacent structures were present, permanent rotation (structural tilt) was observed to increase significantly, demonstrating the importance of considering structure–soil–structure interaction in assessing the seismic performance of structures.

During earthquakes, saturated sandy soils may generate significant excess pore pressures and approach a state of liquefaction. Structures founded on shallow foundations above such soils may consequently undergo large settlements. Recent case history analysis has shown that the stress imposed by the foundation is a key factor in the estimation of such settlements. However, the case history data showed that although increasing bearing pressure caused an increase in settlements as expected, this was only true up to a point, and that very heavy structures appeared to settle less than some lighter structures. This work aims to investigate these counter-intuitive results by means of controlled experimental testing using a geotechnical centrifuge. Results of the centrifuge tests show that the trend derived from case histories is correct and that liquefaction-induced settlements peak for a given bearing stress (90 kPa for the models tested) and reduce for greater applied stresses. Further, by analysis of excess pore pressure distributions beneath the foundations it is shown that the main factor inhibiting pore pressure generation beneath the footings is not so much the confining pressure as the in-situ static shear stress around the edge of the foundation. This is supported by element test data from the literature. When this initial static shear stress is so high that the applied cyclic shear stress cannot exceed it (i.e. the direction of shear stress does not reverse) then pore pressure generation is greatly reduced, thus causing the observed reduction in expected settlements.

This paper reports a set of centrifuge tests to investigate the increase in bearing capacity of a skirted circular foundation that results from preloading with consolidation. Results are compared with coupled numerical analyses using the modified Cam Clay model. Fundamental mechanisms are revealed from the stress paths followed during preloading and consolidation at the element level, accounting for both the change in mean normal effective stress and in the hardening parameter. A formulation is developed to predict the gain in bearing capacity as a function of the level of preloading applied on the foundation and the degree of consolidation achieved. The formulation provides good agreement with the centrifuge results.

Spudcan foundations of various geometries are commonly used in the field for supporting offshore oil and gas drilling (jack-up) rigs. This paper investigates the effect of the spudcan and its base geometry on the penetration and extraction resistance in soft normally and lightly overconsolidated clay. A series of centrifuge tests was undertaken on model spudcans of three different geometries typically used in the field. The spudcans were extracted from penetrations of 2·64∼3 (area equivalent) diameters and after an operation period of ∼19 months. Additional large-deformation finite-element analyses allowed for exploring a range of spudcan base geometries. Corresponding field data were accumulated. The results from this study have confirmed the accuracy of the assumption of a circular conical-based footing and the use of area equivalent diameter for assessing spudcan penetration resistance, as suggested by ISO guidelines 19905-1. Simulation of a flat-based circular plate resulted in 19% higher resistance compared to that of conical-based footings. For spudcans with cutouts and with a peripheral skirt, maximum penetration resistances were, respectively, 14 and 12% lower, and maximum extraction resistances were 20 and 14% lower compared to those for a general spudcan. The corresponding required jack-up leg length was 9 and 7% higher. The effect of peripheral skirt length was shown to be minimal (at least for length relative to spudcan diameter ≤ 0·42).

Challenges associated with dynamically installed anchors include prediction of the anchor embedment depth, which dictates the anchor's holding capacity. This is particularly true for calcareous sediments, as very little performance data exist for this anchor type in these soils. This paper reports results from a series of model tests undertaken to provide insight into the behaviour of a torpedo anchor during dynamic installation and monotonic pullout in lightly overconsolidated calcareous silt. The tests were carried out in a beam centrifuge, varying the drop height and consequently the impact velocity, and the consolidation period prior to anchor pullout. The mudline load inclination was also varied to encompass various mooring configurations. The centrifuge model test data were used to calibrate: (a) an analytical dynamic embedment model, based on conventional bearing and frictional resistance factors but with strain-rate-dependent undrained shear strength for the soil; and (b) an analytical quasi-static vertical pullout capacity model, accounting for reverse end bearing and frictional resistance. A total energy based expression, appropriate for calcareous silts, was proposed for predicting anchor embedment depth for a given anchor geometry, mass and impact velocity. For assessing anchor vertical holding capacity, a piezocone based direct design approach was also proposed, deriving anchor end bearing and frictional resistance from cone tip resistance and sleeve friction, respectively. Anchor capacity under inclined loading was presented as failure envelopes expressed in terms of dimensionless vertical and horizontal components of anchor net resistance, which agreed well with a finite-element based envelope developed for embedded foundations. The regain of anchor capacity was found to be in good agreement with predictions based on the cavity expansion framework.

The effects of tunnel construction on existing single piles have been extensively investigated, but the influence of twin tunnel advancement on an existing pile group is seldom reported in the literature. In this study, a series of three-dimensional (3D) centrifuge model tests and numerical simulations using an advanced hypoplastic soil model were carried out to investigate the response of an existing 2 × 2 pile group to piggyback (i.e. vertically aligned) twin tunnelling at various depths. Soil parameters for 3D numerical back-analyses were obtained from stress-path triaxial tests. In each twin tunnelling simulation in a centrifuge, the first tunnel was advanced three-dimensionally by controlling volume loss in-flight near the mid-depth of the pile shaft, before the second tunnel was constructed either next to the toe of the pile group (test ST), below but to one side of the pile group (test SB) or directly beneath the pile toe (test SU). The piggyback twin tunnelling in test ST resulted in the largest transverse tilting (of 0·2%) but the smallest settlement of the pile group under a working load. This is because the second tunnelling caused significant non-uniform change in vertical effective stress underneath the four piles in the group. On the contrary, the tunnelling directly beneath the pile group (i.e. test SU) caused the smallest tilting but the largest settlement of the pile group (4·6% of pile diameter) and substantial mobilisation of shaft resistance. This is attributable to the most significant and uniform loss of toe resistance of each pile in the group resulting from stress relief from the second tunnelling. Two distinct load transfer mechanisms can be identified in the pile group, namely downward load transfer in test ST and upward load transfer in tests SB and SU.

The commonly used British Standard constant head triaxial permeability test for testing of fine-grained soils is relatively time consuming. A reduction in the required time for soil permeability testing would provide potential cost savings to the construction industry, particularly in the construction quality assurance of landfill clay liners. The purpose of this paper is to evaluate an alternative approach of measuring permeability of fine-grained soils benefiting from accelerated time scaling for seepage flow when testing specimens in elevated gravity conditions provided by a centrifuge. As part of the investigation, an apparatus was designed and produced to measure water flow through soil samples under conditions of elevated gravitational acceleration using a small desktop laboratory centrifuge. A membrane was used to hydrostatically confine the test sample. A miniature data acquisition system was designed and incorporated in the apparatus to monitor and record changes in head and flow throughout the tests. Under enhanced gravity in the centrifuge, the flow through the sample was under ‘variable head' conditions as opposed to ‘constant head' conditions as in the classic constant head permeability tests conducted at 1g. A mathematical model was developed for analysis of Darcy's coefficient of permeability under conditions of elevated gravitational acceleration and verified using the results obtained. The test data compare well with the results on analogous samples obtained using the classical British Standard constant head permeability tests.

Spudcan foundations of mobile jack-up rigs are penetrated into the seabed under seawater ballast preload, which is shed prior to rig operations commencing. During pauses in the installation process and during operation, soil beneath the spudcan foundations stiffens and strengthens due to consolidation. On the application of further loading or during spudcan extraction, this causes increased resistance, which in extremis can result in punch-though type failure. This note reports results from a series of experiments with particle image velocimetry measurements that were performed in a drum centrifuge to facilitate observation of the effects of a load-hold period on the soil movements around a model spudcan during subsequent further loading. The results show that the dimensionless load-hold period dominates the enhancement in the penetration resistance, due to significantly more soil being mobilised following a long load-hold period. These observations might be useful to (a) predict the enhancement in bearing capacity factor due to a load-hold period during installation or operation and (b) predict the footing extraction resistance during jack-up re-deployment.

Shallow foundations supporting bridge piers, building frames, shear walls and monuments are often subjected to extreme lateral loading such as wind in offshore environments, or strong seismic shaking. Under such loading conditions, foundations may experience a host of non-linear phenomena: sliding on and uplifting from the supporting soil or even soil failure in the form of development of ultimate bearing capacity mechanisms. This type of response is accompanied by residual settlement and rotation of the supported structural system. Nevertheless, inelastic foundation performance can provide potential benefits to the overall seismic integrity of the structure. Thanks to such non-linearities, energy dissipation at or below the foundation level may eventually limit the seismic demand on structural elements. Several theoretical and experimental studies have provided encouraging evidence to this effect. This paper has a dual objective: first, to study the behaviour of shallow foundations under vertical and lateral monotonic loading and under lateral slow cyclic loading of progressively increasing amplitude; second, to explore the differences in foundation response between reduced-scale 1g and centrifuge 50g model testing. Emphasis is placed on interpreting their discrepancies by unveiling the role of scale effects. The role of soil densification due to multiple loading cycles with uplifting is also highlighted.

Increasing demands for oil and gas exploration in deep water with soft seabed conditions are resulting in the size and weight of subsea shallow foundations stretching the capabilities of installation technologies. One innovation to reduce foundation footprints involves designing foundations to move in a tolerable manner to absorb applied loads rather than being engineered to resist these loads and remain stationary. Critical design considerations are the evolution of foundation capacity and the mode of foundation displacement. The foundation should be designed to slide with acceptable settlement and rotation to prevent overstressing the joints with connected pipelines. This paper presents observations from centrifuge model tests of a mat foundation designed to slide under applied loading. The foundation is subjected to a simulated lifetime of operation, with many cycles of sliding and intervening periods of consolidation. The results provide insights to assist design, including a remarkable rise in the lateral foundation resistance over the sliding events, through repeated episodes of shearing and reconsolidation, and quantification of the accumulated settlements and rotations. The foundation is shown to translate with minimal rotation. The settlement between sliding events is more significant. This is due to the tendency for soft clay to contract on shearing, as excess pore pressures generated during sliding subsequently dissipate. The sliding-induced consolidation settlements control the tolerability of the performance of the mobile foundation.