Centrifuge model study on spudcan footprint interaction 5

The footprint characteristics are dependenton factors such as spudcan shape and size, soil type and strength, previousspudcan penetration, and the elapsed time after previous spudcan ext

Chapter 4

Footprint Characteristics and Their

Influence on Spudcan-footprint Interaction

An extensive series of centrifuge model tests was carried out to investigate thespudcan-footprint interaction problem in clays of various strengths The firstpart of this chapter focuses on identifying the footprint characteristics, whereasthe second part is to investigate the effects of these footprint characteristics onthe new spudcan installation The spudcan-footprint interaction is evaluated interms of horizontal force and moment profiles acting on the spudcan duringinstallation Half spudcan tests were also conducted to observe the soil failuremechanism during the initial penetration, extraction and the spudcan re-penetration at 0.5 times spudcan diameter offset from the initial site Theresults will be presented in this chapter

The above footprint characteristics can significantly influence the manner inwhich a spudcan interacts with it The footprint characteristics are dependent

on factors such as spudcan shape and size, soil type and strength, previousspudcan penetration, and the elapsed time after previous spudcan extraction.Spudcan-footprint interaction can occur in a wide range of soil conditions(Osborne et al., 2006) In this chapter, the footprint characteristics and themechanism of interaction between spudcan and existing footprint in clay withdifferent shear strength profiles were investigated using centrifuge modellingtechnique Details of the centrifuge test results and practical implications ofthe findings are reported in this paper

4.2 Test programme

4.2.1 Formation of a spudcan footprint

All the centrifuge model tests were carried out at 100g in the NationalUniversity of Singapore geotechnical centrifuge The tests were conductedusing Malaysia kaolin clay with the physical properties as shown in Table 3.5

Details of the experimental model set-up and soil sample preparation can befound inChapter 3 As the present study aims to investigate the soil responses

in the spudcan-footprint interaction in detail, only a single leg with fully rigidconnection was modelled A 100 mm diameter model spudcan was used Thespudcan has an 11o base angle and an 80o conical tip A schematic of thespudcan and the model leg (Leg 1) with instrumentation details are given in

prototype length of 28 m and a flexural stiffness of 1.18×1012 Nm2 at 100g.This flexural stiffness value is similar to Jack-up 116-C (Global Maritime,

2003) The model leg is rigidly connected to the spudcan and the other end is

rigidly connected to the loading actuator The leg was instrumented with 2levels of axial strain gauges and 3 levels of bending gauges During a test, theloads on the spudcan were measured from the strain gauges at 1 secondinterval.

To form a footprint, the spudcan was first penetrated into the modelground at the desired location at a penetration rate of 1 mm/s Based on thevelocity group parameter proposed by Finnie (1993), undrained condition forthe penetration process was preserved at this penetration rate The spudcancontinued penetrating until the desired preload pressure was achieved Therequired preload level in the field is dependent on several factors such asnumber of footings, size of footing, environmental loadings and soil bearingresistance For earlier rigs, the average bearing pressures on soil beneath thespudcan lie between ranges of 200 to 350 kPa (Le Tirant, 1979) For a modernjack-up, maximum leg loads can be higher than 140 MN that produces averagevertical bearing pressures in excess of 400 kPa for a fully embedded spudcan(Randolph et al., 2005) Some rigs have considerably higher preload pressuresranging from 575 to 960 kPa (Poulos, 1988) In general, there has been anincrease in the maximum vertical installation bearing stresses from a range of

200 to 400 kPa, to around 400 to greater than 600 kPa for the popular rigclasses (Osborne et al., 2006) The maximum preload pressure used was 460kPa in all the tests This is deemed to fall within the normal preload pressure.The spudcan was then withdrawn immediately at a rate of 1 mm/s and afootprint was created This is to simulate the scenario where a rig is engagedfor a very short operational period Depending on the type of work engaged,the operational period may range from a week to up to 2 year Time is an

important variable in this problem This is particularly the case for footprintsformed in fine grained soils, where a re-consolidation process can occur Theeffect of time on spudcan-footprint interaction will be studied inChapter 5.

4.2.2 Evaluation of spudcan-footprint interaction

As the leg-hull connection was modelled as fully rigid, where no lateraldisplacement and rotational movement are allowed, the spudcan-footprintinteraction is evaluated in term of three major ‘resultant’ load components(vertical force, V, horizontal force, H and moment, M) acting at the spudcanload reference point (L.R.P.) These forces are computed based on the straingauge measurements on the leg using the formulae shown inAppendix A

4.2.3 Evaluation of soil condition – Ball penetrometer test

For the tests presented in this chapter, a miniature ball penetrometer consisting

of an 11.9 mm diameter ball connected to a narrow shaft was used to evaluatethe undrained shear strength profile of the clay prior to spudcan penetrationand after the formation of the footprint After several trials, the ballpenetrometer was found to provide comparable but more stable readings thanthe commonly used T-bar for the measurements of shear strength profile of afootprint This is because the soil profile within a footprint is highly variable.This results in bending of the T-bar owing to asymmetrical resistance, and thisaffects its measurements In this situation, the axisymmetric ball penetrometerhas an advantage over the T-bar as it is less susceptible to bending

The ball was installed at a penetration rate of 3 mm/s Such a rate issufficiently fast to maintain the penetration process under undrained conditionbased on the velocity group parameter proposed by Finnie (1993).The

undrained shear strength profile of the soil can be obtained by measuring thepenetration resistance of the ball while penetrating into the clay (Randolph et

capacity factor for the ball penetrometer, Nball, is taken as 10.5

4.2.4 Experiment procedure

After the clay sample achieved at least 90% degree of consolidation, ballpenetrometer test was conducted in-flight at a position far from the intendedfootprint location to measure the original undisturbed shear strength profile ofthe clay The movable platform was shifted to different locations such that ballpenetrometer tests could be conducted at various offset distances from thecentre of the footprint, see Fig 4.1 Upon completion of shear strengthmeasurements, spudcan re-penetration was then immediately performed at arate of 1 mm/s at the desired location Throughout the spudcan installation,extraction and re-installation, all the strain gauges were monitored at 1 secondintervals

4.3 Experimental results and discussions

4.3.1 Shear strength profiles and spudcan penetration depths

Four centrifuge model tests, namely tests CS_1 to CS_4, were conducted inclay with various shear strength profiles In all tests, the spudcan used has aprototype diameter D of 10 m The ball penetrometer test results obtained prior

to spudcan installation reveal that the undrained shear strength, su, increasesalmost linearly with depth for the four clay samples (Fig 4.2) A summary ofthe suprofile presented in the form of (sum+kz) where sum is the shear strength

at mudline in kPa, k is the gradient of shear strength profile in kPa/m and z is

the depth from the mudline in m, is given in Table 4.1 Among the four claysamples, test CS_1 has the strongest shear strength profile representing firmclay while test CS_4 has the weakest soil strength profile representing softclay.

Unless otherwise stated, all the test results presented hereinafter are inprototype units, which are derived using appropriate scaling laws (Taylor,

1995) from the model units In the present study, a systematic sign convention

as recommended by Butterfield et al (1997) is adopted where downwardvertical force, V, is positive, horizontal force towards footprint, H, is positiveand clock-wise moment, M, is positive, seeFig 4.3 for these forces relative tothe footprint position All the load components (V, H and M) and penetrationdepth presented in this report are those acting at the spudcan load referencepoint shown in Fig 4.3 The preload pressure, q for spudcan initial and re-penetration can be obtained by the following relationship:

where V is the measured vertical load and A is the largest bearing area of thespudcan, which is 78.5 m2 On the other hand, though M and H at thereference point cannot be measured directly from the experiment but they areevaluated through extrapolation of the bending moments measured by bendinggauges instrumented at the spudcan legs using the equations derived in

vertical soil bearing resistance reached 460 kPa For the same spudcan, thefinal spudcan penetration depth is dependent on the soil strength The preloadpressure-displacement curves during spudcan penetration and extraction for

tests CS_1 to CS_4 are shown inFig 4.4 It should be noted that a penetration

of 0 m refers to the widest spudcan section is right at the mudline The finalpenetration depth of the initial spudcan installation, do is about 2 m, 5 m, 8 mand 13 m for tests CS_1, CS_2, CS_3 and CS_4, respectively

4.3.2 Characteristics and physical profile of footprint

Spudcan extraction was performed immediately (within 5 seconds in modeltime) after spudcan installation A miniature camera was mounted on thecontainer to capture the movement of the surface soil during spudcaninstallation and re-installation Unfortunately, the footprint characteristics forTest CS_3 could not be captured due to problem with the camera Fig 4.5a

reveals that the footprint formed in test CS_1 having an initial penetrationdepth do of 2 m is about one spudcan diameter wide with an almost verticalcylindrical circumference and a relatively less disturbed base A number oftension cracks were observed around the circumference of the footprint Fig.4.5b shows that the footprint from test CS_2 (do = 5 m) consists of a bowl-shaped depression with its deepest point of about 2.3 m at the centre The size

of the circular depression of heavily disturbed soil is about 1.6 spudcandiameters

For the footprint formed in test CS_4 (Fig 4.5c) with do = 13 m, itsdiameter is about 2 spudcan diameters and its deepest point at the centre isalmost 2 m deep Soil heave along the footprint periphery is clearly visible.Within the depression, the soil is heavily remoulded and the highly irregularand ‘lumpy-like’ soil surface indicates that the soil is rather soft

4.3.3 Soil failure mechanism during spudcan penetration and

extraction

The observations in Section 4.3.2 showed that the seabed profile of a footprintvaries with the penetration depth under the same preload pressure Toinvestigate how the spudcan penetration and extraction alter the footprintprofile, two half-spudcan tests, namely tests CS_1A and CS_2A wereconducted to observe the soil movement on soil having similar shear strengthprofile as tests CS_1 and CS_2 Unlike tests CS_1 and CS_2 which employed

a reduced-scale full spudcan, only one half of the full spudcan was employed

in tests CS_1A and CS_2A with the face of the half-cut spudcan placed behind

a transparent Perspex window of a rectangular model container The claysample was textured with black flocks and beads and a digital camera wasplaced in front of the Perspex window to capture high resolution photographsthroughout the tests The photographs were analysed using Particle ImageVelocimetry technique employing GeoPIV8 software (White & Take, 2002;White et al 2003) The detail of the half-spudcan test set-up was presented in

The soil movement patterns at various selected stages of penetration andextraction shown are in the forms of velocity fields and the correspondednormalized velocity contours The velocity field represents the increment ofresultant displacements taken from a pair of sequential images at 2 sec interval.The normalized velocity contours indicate the ratio of the soil displacements tothe spudcan movement, which was mainly in the vertical motion The majordisplacement field is defined as soil displacements equal to or greater than10% (or ratio 0.1 as shown in velocity contour plots) with respect to the

spudcan movement The results presented have a physical dimension in aprototype unit of meter.

4.3.3.1 Spudcan penetration in undisturbed ground – test CS_2A

The soil movement during the spudcan penetration at various penetrationdepths are presented in a pair of velocity vector and velocity contour plots andshown in Figs 4.6 and 4.7 Only the results obtained from test CS_2A isreported in this section.The surface soil movements of the entire penetrationprocess were captured in the full spudcan test (test CS_2) and are presented in

Figs 4.10a – d

When the largest surface of the spudcan completely touched theseabed, the soil was pushed downward and outward The extent of the majorsoil displacement was about 0.8D (vertical) from the spudcan load referencelevel and and 1D (lateral) from the spudcan centre As the spudcan continuespenetrating to about 1 m, some minor crack lines began to appear on the soilsurface (seeFig 4.10b) When the spudcan continued the penetration below 1

m, soil beneath the spudcan was continuously being pushed outward anddownward with substantial volume of soil being brought down together withthe advancing spudcan The major vertical soil displacement extended to about1D while the radial lateral extent was still confined within 0.8 – 1D A stableopen cavity on top of the spudcan was formed without significant soilbackflow over the spudcan penetration depth At this stage, more crack lineswere developed on the surface and the crack width enlarges with thepenetrating spudcan (see fig 4.10c & d) As in an undrained condition, nochange in volumetric strain, the soil surrounded the spudcan’s circumference

volume of the cavity and the spudcan The heaving motion induced tensileforce on the soil and resulted in the development of tension cracks Untilpenetration depth of 5 m as the desired preload pressure of 460 kPa wasachieved, a cavity on top of the spudcan remained stable with substantialcracks formed (seeFigs 4.9 and 4.10d).

4.3.3.2 Spudcan extraction – test CS_2A

At the initial stage of extraction, soil beneath the spudcan experiencessubstantial rebound and moved upwards with the extracting spudcan (see Fig.4.11) The tension cracks width decreased with the uplifting movement of thespudcan, as observed in Fig 4.15a Substantial upward movement of theunderlying soil continued until the onset of spudcan breakout (Fig 4.12) atdepth of 2.1 m or a spudcan uplift movement of 2.9 m At this stage, the cracklines were closed up as observed in Fig 4.15b After spudcan breakout,uplifting of the underlying soil was less dominant with more soil from thesides moving inward After another small uplift displacement of 0.l m, theseparation of soil from the spudcan base was observed, see Fig 4.13 Fromthis stage onwards, the soil movement only involved soil flow from the sides

to beneath the base of spudcan Fig 4.15c shows a circular crack of about1.6D formed at the soil surface when the spudcan is at 1.9 m Shortly after thespudcan was fully extracted from the model ground (see Fig 4.14 and 4.15d),

an abrupt circular slide was observed to occur where the soils at the sides slidtowards the depression with the creation of a 2.3 m deep bowl-shapeddepression After spudcan extraction, soil upheaval was clearly observedsurrounded the footprint circumference, as shown inFig 4.14

4.3.3.3 Effect of spudcan extraction on footprint profile

Hossain et al (2005) reported that the soil back flow during spudcanpenetration is due to soil failure triggered by the penetrating spudcan ratherthan due to the collapse of soil around the vertical sides of the depression Forsoil with relatively higher shear strength, a stable open cavity can be formed

on top of the spudcan without any significant soil back flow to the finalpenetration depth As shown in Fig.4.11 (test CS_2A) and Fig 4.16a (testCS_1A), the sides of the depression above the spudcan at the final penetrationdepth remained stable with no significant soil back flow observed in the tests.For test CS_1A (do = 2 m), Fig 4.16b reveals that at the onset of spudcanbreakout at about 0.5 m of spudcan uplift movement, the extent of theunderlying soil moving upward was reduced and soils from the sides movedlaterally towards the base of the uplifting spudcan Immediate after spudcanbreakout, a gap was observed between the spudcan base and the soil When thespudcan is fully extracted from the model ground (Fig 14.6c), the adjacentsoil remained stable, leaving a cylindrical depression with almost vertical sides.The findings suggest that in relatively firm soil where a stable open cavityformed on top of the spudcan at the final penetration, the surface geometric of

a footprint is dependent upon the stability of the depression during spudcanextraction

Purwana et al (2007) reported the results of extraction of halfspudcan tests in soft clay In soft clay, as massive soil backflow involvedduring the spudcan penetration, only a limited height of open cavity remained

on top of the buried spudcan During the spudcan extraction, the underlyingsoil moved upwards with the uplifting spudcan and the top soil flowed from

the top to the spudcan base As the spudcan penetration depth of over onespudcan diameter is relatively deep, the lateral extent of the soil distortionduring the collapse of the depression is wider at about two times spudcandiameters It can hence be established that the footprints formed for tests CS_3

& CS_4 would have similar physical features but wider compared to thatobserved in CS_2

4.3.4 Soil condition within and around a footprint

Immediately after the spudcan extraction, the undrained shear strength suprofile of the footprint at several locations from the footprint centre (see

spudcan tests The overall process took approximately 6 minutes model time

to complete Previous studies by Gan et al 2007 and Leung et al 2007

revealed that the remoulded soil within a footprint gained strength with timewhen the excess pore pressure generated by previous installation dissipated.This implies that the su profile presented in this chapter represents the weakerstrength than the long-term strength where the re-consolidation of the soiltakes place

The su profiles of the footprints and the corresponding undisturbed suprofiles are shown in Figs 4.17 to 4.19 for CS_1, CS_2 and CS_4,respectively The overall process for ball tests took approximately 0.1 year(prototype time for soil consolidation) to complete It is evident that theamount of soil disturbance decreases with increase in radial distance from thefootprint centre and the disturbance becomes less significant beyond onespudcan diameter This observation is consistent with the findings byStewart(2005), Gan et al (2007) and Leung et al (2007)

To further evaluate the shear strength variation within a footprint, theshear strength contours for tests CS_1, CS_2 & CS_4 are plotted inFigs 4.20

remoulded and undisturbed su, respectively with dotted lines indicatingextrapolated su contour lines It is worth noting that the su profile basicallyfollows the physical profile of the footprint depression observed from the half-spudcan tests As an example, the almost zero su values measured at 0.2D fromthe footprint centre essentially confirms the extent of the depression

In general, the soil experiences substantial reduction in su up to a distance

of 0.8D to 1.0D from the footprint centre, with the reduction decreases withincreasing radial distance During initial spudcan penetration, the soil belowthe spudcan is being remoulded and pushed down by the advancing spudcan.This attributes to the relatively lower su within the spudcan (i.e 0.5D fromfootprint center) as compared to the soil outside the spudcan area at the sameelevation A surprisingly higher su relative to undisturbed soil at a depth of

10 m and within radial distance of 5 m for test CS_2 is observed (Fig 4.21).This is probably attributed to the uplifting of the underlying soil (stronger soil

as the shear strength profile increases with depth) during spudcan extraction asthe soil experiences substantial rebounding and suction develops due to stressrelief at the spudcan base (Purwana et al., 2005) The ‘slanting’ profile of sucontour indicates the possibility of soil sliding towards the footprint

It is evident that spudcan-footprint interaction very much depends on theshear strength of the soil and the physical profile of the footprint due toprevious spudcan extraction The next section examines the effects of physical

geometry of footprint and soil shear strength profile on the spudcan penetration.

re-4.3.5 Spudcan-footprint interaction

To investigate the most critical spudcan-footprint interaction situation, thesame spudcan was used to create footprint and to conduct the re-penetration at0.5D from the footprint centre for tests CS_1 to CS_4

The profiles of preload pressure q, horizontal force H and moment Mduring spudcan installation and re-installation are shown in Fig 4.23 for thefour tests and a summary of the results is presented inTable 4.1 It is observedthat the maximum H and M develop at relatively shallower depth for testsCS_1 and CS_2 with firmer soil profile and at greater depth for tests CS_3 andCS_4 with softer soil profile

4.3.5.1 Penetration in firm clay (d o = 2 m) – test CS_1

The H & M increase considerably once the spudcan is in contact with the soiland soon reach their maximum magnitudes of 1.27 MN (Hmax) and 19.66MNm (Mmax) at a re-penetration depth of 1.2 m (Fig 4.23a) When thespudcan first penetrates into the ground, it was only partially supported on thedisturbed ground (seeFig 4.24a) Such eccentric bearing resistance causes thespudcan to rotate and slide into footprint as registered by the measured M & Hvalues When the spudcan advances beyond 1.2 m depth, the ‘suspended’ part

of spudcan comes into contact with the soil resulting in smaller eccentricbearing resistance acting on the spudcan and hence both H and M reduce (see

(denotes as CS_1B) was conducted where the crater was pre-formed on the

undisturbed sample An identical spudcan was buried into the clay to 2 mdepth and the clay was pre-consolidated at 1-g gradually to the level same astest CS_1 The small variation in crater depth in both tests is due to the soilbeneath the footprint in test CS_1 rebounded after the removal of the spudcan.This has resulted in a shallower crater compared to the pre-formed crater Theundisturbed su for both tests are presented in Fig 4.25a The load displacementresponse for test CS_1B is slightly stronger than test CS_1 as the soil in testCS_1B has undisturbed strength (Fig 4.25b) The induced H and M are higherfor test CS_1B than test CS_1 as the greater differential in bearing resistance

in the pre-formed crater as the surrounded soils are undisturbed (Figs 4.25c

variation in the soils surrounded the crater does not result in higher H and Mduring the interaction This finding suggests that initially the depression andthen the partial support of the spudcan dominate the spudcan-footprintinteraction in surface penetration case in firm clay

4.3.5.2 Penetration in soft to firm clay (d o = 5 m) – test CS_2

first in contact with the soil and increases to a maximum value of 13 MNm at

a re-penetration depth of 1 m at which part of the spudcan is not rested on thesoil Although the induced H increases with depth to a maximum value of 1.2

MN at depth of 2.6 m, the value of H stays at about 1 MN to a depth of 5 mbefore it decreases As the deepest point of the depression is around 2.1 m, thehigh H value after this depth is probably attributed to the soil strengthdeviation This aspect will be further examined later Under this scenario, new

spudcan installation is adversely affected by both the depression and the soilstrength variation.

4.3.5.3 Penetration in soft to firm clay (d o = 8 m) –test CS_3

below the depression The induced H increases with penetration depth andreaches its maximum value of 1.11 MN at 6 m depth On the other hand, theinduced M exhibits ‘double humps’ with the first peak M of 6 MNm occurring

at 1 m depth This is an indicator of the effect of the depression on the spudcanre-penetration The second peak occurs at a depth of 7.3 m with a higher peak

M of 7.95 MNm

4.3.5.4 Penetration in soft clay (d o = 13 m) – test CS_4

The induced H and M profiles shown in Fig 4.23d are similar to thoseobserved in test CS_3 with higher H and M values recorded at greater depth.Before reaching the penetration depth of the previous spudcan installation, asofter load-displacement response is noted with an average reduction ofpreload pressure of about 85 kPa Beneath 13 m depth, the load-displacementresponses for spudcan penetration and re-penetration are similar H increasesalmost linearly with depth and reaches a maximum value of 1.15 MN at 9 mdepth and then decreases to a nil magnitude at 14 m depth Likewise, the Mvalue also increases with depth and reaches a maximum value of 9.4 MNm at

11 m depth This observation suggests that the soil shear strength variationdominates spudcan-footprint interaction in soft clay

4.4 Mechanisms of spudcan-footprint interaction – Soil

response

To observe the soil failure mechanism during spudcan re-penetration, soilmovement during spudcan re-penetration at 0.5D was captured in the samehalf spudcan test of CS_2A and the displacement vector plots of soilmovement at various depths are shown inFigs 4.26 – 4.30 Each set of figuresshow the soil movement during the spudcan-footprint interaction at fivedifferent depths, namely 0 m, 1 m, 2.6 m, 5 m and 7 m The displacement fieldassociated with each stage is presented in term of velocity field, normalizedvelocity contour, lateral and vertical velocity contours The velocity contour isthe resultant velocity of the lateral and vertical velocities For the signconvention, the downward vertical movement and inward lateral movement(towards the footprint) with respect to the spudcan are taken as positive

When the largest bearing area of the spudcan was at the seabed level(depth = 0 m), part of the spudcan is practically resting on no soil (see

pushed outwards and downwards that attributes to an eccentric soil reaction

No significant soil sliding failure is observed At 1 m penetration, largercontact area between the spudcan and the soil resulted in a larger soildisplacement extent compared to 0 m penetration (Fig 4.27a & b) As shown

spudcan with larger soil bulb on the right hand side (refers to the figure) Thisasymmetric lateral soil movement indicates the onset of the sliding of soilmass towards center of footprint In term of the vertical soil movement, afairly symmetric movement pattern was observed (see Fig 4.27d) Due to the

strength further from footprint center, the mobilized soil bearing resistanceacting on the spudcan was highly non-uniform across the spudcan base andsubsequently resulted in an eccentric soil reaction This eccentric soil reactioncoupled with the soil sliding failure towards footprint generated an increasing

H and M profiles from 0 to 1 m

When the spudcan penetrated to 2.6 m (where the occurrence of Hmax), amore substantial soil mass sliding towards the footprint center was observed

slide towards the footprint center As the movement of the spudcan wasrestricted as a rigid connection was adopted, the tendency of sliding wastransmitted into horizontal force This is reflected in the increasing H profile to

a maximum value of 1.2 MN at 2.6 m as indicated inFig 4.23b At this stage,the vertical contour was still fairly symmetrical with major soil displacementextended from the spudcan base to 0.8D

At the initial penetration depth (do = 5 m), the soil velocity contour wasfairly symmetrical (seeFig 4.29b) The contours profile for lateral andvertical velocity at both the left and right hand sides appeared to be similar, asshown in Figs 4.29c and d When the spudcan penetrates beyond the initialpenetration depth to 7 m (seeFigs 4.30a & d), no further sliding failure wasobserved The soil movement underneath the advancing spudcan appeared to

be more symmetrical This corresponded to decreasing H & M profiles withdepth after do, as shown inFig 4.23b

Based on the above observation, a simplified soil failure mechanism atdifferent penetration depths is postulated The footprint condition prior to thespudcan installation is shown in Fig 4.31a For penetration depth above the

crater depth, H and M are caused by the non-uniform soil bearing resistancewhere the resultant soil reaction inclines at an angle of  to vertical and aneccentricity of e from the spudcan centre This inclined eccentric soil reactiontends to push and rotate the spudcan towards the footprint centre, as indicated

the crater, the soil sliding failure is initiated that drives the spudcan towardsthe footprint centre At the same time, the resistance from the soil outside thesliding plane tends to push the spudcan towards the footprint centre, as shown

the present model tests, the tendency of horizontal movement and rotation aretransmitted as horizontal forces (Hbearing and Hslide) and rotational moments(Mbearing and Mslide) acting at the spudcan A combination of these two forceshas the effect of magnifying the magnitude of induced horizontal force (asboth forces tend to move the spudcan towards the footprint centre) whilediminishing the magnitude of induced rotational moment (caused by rotation

in opposite direction)

As all the soil samples used in the model tests have an undisturbed shearstrength profile increasing with depth, this explains why the measured Hincreases with penetration depth and decreases only in a zone beneath theheavily remoulded soil This also explains why the measured M does notalways increase with depth, in particular for deep penetration cases.Theoretically, if these forces are fully resisted by the jack-up rig (where thehorizontal restraint can be obtained by re-oriented rig in such a way that thepre-installation of the other legs onto undisturbed ground is possible asrecommended by SNAME (2002a)) and the forces can be structurally resisted

by the rig, the jack-up rig can be safely installed into the ground with lesssliding concerns Otherwise footprint mitigation measures need to be taken toensure a safe spudcan re-installation.

4.4.1 Influence of footprint characteristics to off-centred spudcan

installation and potential footprint mitigation methods

When a spudcan penetrates off-centre in an old footprint, the spudcan is onlypartially supported by the ground due to the depression In reality, a craterexists in every footprint and the soil beneath it is highly non-uniform.Questions that arise are; i) how would a depression affect the new spudcaninstallation, and ii) which of the factors: physical profile of depression orvariation of soil strength is dominant in affecting the new spudcan installation

As presented earlier, the physical profile of a depression depends on theundisturbed soil shear strength and spudcan penetration

Similar to a slope stability problem, both the depth and gradient of thesloping surface of a depression play significant roles in the interaction betweenthe spudcan and footprint in shallow penetration depths In addition, the soilstrength variation within the shear zone of the new spudcan is deemed to besignificant in spudcan-footprint interaction In other words, the physicalprofile of the depression is significant in the scenario where the footprintdepression profile is steep and deep, a scenario common in firm to stiff soils.This has resulted in the high H and M recorded at shallow depths (within thedepression), as occurred in test CS_1 To minimize the effect of footprint onnew spudcan installation, one may consider modifying the seabed profile such

as evening out the depression, or infilling the depression with suitableimported material, in order to provide a more level contact between the

spudcan and the seabed Modifying the spudcan configuration such as skirtedspudcan to transfer the bearing plane to a depth below the depression may also

be considered However, the effectiveness of these mitigation methods clearlydeserves further study

Levelling the seabed profile by means of infilling or excavation isunlikely to be effective in mitigating the spudcan-footprint interaction problem

if the critical H and M are due to the variation of soil strength that occur farbelow the crater depth such as in soft clay In such cases, one may consider ofcreating additional ‘disturbance’ to weaken the soil in the stronger part(outside the previous spudcan imprint or locations further from footprintcentre) such as Swiss-cheesing or soil boring to produce a less deviated soilshear strength profile for the new spudcan installation

4.5 Concluding remarks

Footprint depression is formed due to extraction of previously installedspudcan The soils within a footprint are heavily remoulded with a highlyvariable shear strength and physical profile Under the same preload pressure,the penetration depth of the spudcan would depend on the soil shear strength.The spudcan penetration depth and the subsequent spudcan extraction arefound to significantly affect the physical profile of the footprint In softer claywith relatively deep spudcan penetration, a bowl-shaped spudcan imprint ismore likely to be formed after spudcan extraction In firmer clays withrelatively shallow spudcan penetration, a nearly vertical sided cylindrical withfairly firm base indentation is more likely to be formed The su beneath thefootprint immediately after its formation is lower than those of undisturbed

footprint centre Moreover, the su contours basically follow the physicalprofile of the depression explaining why the tendency of the spudcan to slidetowards the footprint centre.

The physical profile of the depression is recognized to be the dominantfactor in spudcan-footprint interaction in firmer soils On the other hand, thesoil strength variation is found to be more significant in softer soils Theability to identify the dominant factor in the interaction is beneficial in order todevise an effective mitigation measure to overcome the problem of thespudcan sliding into the footprint SNAME (2002a) recommended to infill thefootprint with imported materials in cases where spudcan-footprint interactioncannot be avoided The mechanism of interaction identified suggests that suchmethod may not be effective in soft clays in which the footprint consists of adeep heavily remoulded soil Simplified failure mechanisms on spudcan-footprint interaction at different depths are proposed in the present study Themechanism identify may help future studies to investigate the effectiveness ofdifferent mitigation methods such as infilling with sand and gravel (Jardine etal., 2001), stomping and Swiss-cheesing on overcoming spudcan-footprintinteraction problems

1 z is the penetration depth from mudline (in m);

2 Spudcan diameter, D = 10 m;

3 Preload pressure = 460 kPa

* classified based on the average undrained shear strength (from the mudline to the final penetration depth) in accordance with BS 8004:1986

** measured from the mudline to the crater tip

#Pre-formed crater was used (different from all other tests where the footprint created by performing spudcan penetration andextraction)

Consistency* Penetration

depth,

Craterdepth**

Fig 4.1 Schematic diagram shows test arrangement

Fig 4.2 Undisturbed undrained shear strength, su profiles of the tests