Centrifuge model study on spudcan footprint interaction 6

The primary objectives of the experiments are to investigate: i the soil shear strength changes with elapsed time after the formation of a footprint due to spudcan installation and extra

in firm/stiff clays On the other hand, the soil strength variation in the footprint

is dominant for spudcan-footprint interaction in soft clays More importantly,the degree of soil strength variation changes with time as excess porepressures generated by the previous spudcan installation dissipate (Gan et al.,

2007 and Leung et al., 2007) The soil undergoes re-consolidation during theoperational period and after the extraction of the spudcan This has significantimplications on spudcan-footprint interaction

This chapter presents experimental data on the soil characteristicsbeneath and around a footprint at various times and its influence on an off-centred spudcan re-installation The centrifuge tests presented in this chapterwere conducted on the University of Western Australia (UWA) drumcentrifuge It is worth highlighting that the tests presented in Chapter 4 wereconducted in NUS whereas in this chapter, the tests were conducted in UWA

It is noted that the spudcan, soil and preload pressure were different for bothsets of experiment The effect of these parameters will be studied andgeneralized by dimensionless analysis in Chapter 6 The schematic model set-

up of test conducted in UWA is shown in Fig 5.1(a) The soil shear strengthprofiles at various positions from the footprint centre were evaluated by ballpenetrometer tests The effect of the footprint on the new spudcan installationwas evaluated in terms of vertical and horizontal forces and rotational momentacting at the spudcan reference point with the sign convention as shown inFig.5.1(b) The primary objectives of the experiments are to investigate:

i) the soil shear strength changes with elapsed time after the

formation of a footprint due to spudcan installation and extraction,ii) effects of jack-up operational period on soil strength

characteristics, and

iii) effects of time on spudcan-footprint interaction

To accomplish these objectives, a total of 10 experiments were conducted inboth normally consolidated and over-consolidated clays The raw data wasreported inGan et al (2008a)

5.2 Test programme

5.2.1 Selection of jack-up operational period and elapsed time

after a footprint is formed

In general, the operational period depends on the type of work the rig isengaged in Depending on the task involved, the operational duration of amobile jack-up rig in the field can vary from weeks to as long as 5 years.Typically, a jack-up rig operates at a site from a week to 1 year for exploringoil and gas and up to 2 years for drilling production wells at a new platform

On the other hand, the elapsed time between visits of jack-up rig depends onthe production targets In general, a future rig will visit an established site forworkover after 3 to 5 years Occasionally, a rig is required to return quickly to

a site for production enhancement On other occasions, the well may be leftundisturbed without jack-up intervention for a period of over 5 years

As such, two operational periods (see Fig 5.2) that were considered to boundthe problem were selected in the present study:

a) Case OP 0 - a very short (<1 day) operational period where the

spudcan was extracted immediately upon achieving the desiredpreload of 27 MN;

b) Case OP 2 – upon achieving the desired preload level, the preload

level was reduced to 0.75 times the desired preload level of 27 MNand held for 2 years

The ball penetrometer tests were performed in and surrounding the footprintcreated by OP 0 and OP 2 at 0.25, 0.5, 0.75, 1, 1.25 and 1.5D (where D is thespudcan diameter) from the footprint centre at prototype elapsed times of 1, 3,

5 and about 100 years The elapsed time of 100 years was selected to ensurethat all the generated excess pore pressures have fully dissipated Thelocations of ball penetrometer tests are presented inFig 5.3

5.2.2 Sample preparation

The tests were conducted on the drum centrifuge at the University of WesternAustralia A detailed description of the drum centrifuge can be found inStewart et al (1998) A total of ten (10) tests were performed in a full drumsample at 200g Five tests were conducted in normally consolidated (NC) clay

and five were in over-consolidated (OC) clay Australian kaolin clay was usedand its key properties are shown inTable 3.6.

For the preparation of NC clay, a 12 mm thick sand layer was firstplaced at the base of the drum channel to provide a drainage path for the claysample The clay slurry was then sprayed into the channel while thecentrifuge was spinning at 20g The slurry was then allowed to consolidateovernight at 200g before the second top-up of slurry was placed The completeclay sample was subsequently consolidated at 200g for two days when at least90% consolidation was achieved At this stage, the sample was approximately

160 mm thick with a normally consolidated strength profile The undrainedshear strength profile of the sample was evaluated by performing several T-barand ball penetrometer tests, as presented in Fig 5.4a The undisturbed sucan

be expressed as su = 1.15z, where su is in kPa and z is the depth from mudline

in meter

Upon completion of the tests for NC clay, a 30 mm thick sand layerwas sprayed on top of the sample The drum was then ramped up to 300g forabout 2 days Once the pore-pressures stabilized, the water was drained offand the drum centrifuge was stopped The sand layer was then carefullyscraped away The sample was then ramped to 200g and left to re-consolidatefor another two days After the pore pressures stabilized, an OC sample ofapproximately 150 mm depth The sample had an over-consolidation ratio,OCR of 12 at 1 m and 2 at 20 m below the mudline based on an assumption of

a constant vertical effective stress, v’ of 6.5 kN/m2 The undisturbed shearstrength profile of the sample was evaluated by ball penetrometer tests and

presented in Fig 5.4(b) The undisturbed su(in kPa) can be approximated by

su= 5 + 1.8z

The good repeatability of the test results at different spots indicates theuniformity of the drum sample The measured undrained shear strength profilefor both NC and OC clay samples are fitted with the empirical relationshipexpressed as below (Ladd and Foott, 1974; Jamiolkowski et al., 1985;Koutsoftas and Ladd, 1985):

It is found that empirical su predicted with s = 0.185 and m = 0.8 bestfits the measured undrained shear strength for both clays For NC clay, theOCR value is 1 throughout the entire clay thickness This empirical formula isconsistent with those adopted by Teh (2007) and Lee (2009), who worked onthe same clay material

5.2.3 Model set-up and instrumentation

The spudcan has a prototype diameter of 14.5 m A schematic cross-section ofthe model spudcan is shown in Fig 5.1b The model leg is a thin walledcircular hollow steel section with prototype flexural stiffness value of 3.2×1012

Nm2 and prototype length of 38 m The model leg is instrumented with two (2)levels of bending gauges and one (1) level of axial gauge Hence, the spudcan-

footprint interaction was evaluated in term of forces and moment acting at thespudcan load reference point (see Fig 5.1b) The formulae used to computethese forces were derived and are presented in Appendix A Similar to theexperiments conducted in NUS, a fully rigid connection between the modelleg and the vertical actuator was adopted.

5.2.4 Shear strength measurement devices - Penetrometers

The shear strength profile was measured from the penetration resistance of aball penetrometer Plasticity solution for the flow around a sphere (Randolph

et al., 2000) provides the basis for obtaining the shear strength from themeasured penetration resistance In the present study, the miniature ballpenetrometer had a diameter of 5 mm with a shaft diameter of 2.5 mm A moredetailed description of this ultra-small ball penetrometer that was mould fromepoxy was discussed in Section 3.5.2 As discussed in Section 4.2.3, the ballpenetrometer was found to give a comparable but more stable measurementthan the T-bar For this 5 mm ball, the area ratio of the ball to the spudcan issufficiently small (less than 0.005) for fairer measurements The ball wasinstalled at a penetration rate of 3 mm/s It is established that this rate issufficiently fast to cause the penetration in undrained condition based on thevelocity group parameter proposed by Finnie (1993) A resistance factor of10.5 was adopted for the T-bar in kaolin clay as recommended byStewart andRandolph (1994), whereas a resistance factor of 13.5 was adopted for the ball

as used by Lee (2009) As discussed inSection 3.5.2, the Nball factor is slightlyhigher than 10.5 for ball made from aluminium alloy that was recommended

by Watson et al (1998) This slightly higher Nballfactor may be due to a higherinterface friction ratio of the epoxy and clay that subsequently give higher

penetration resistance For undisturbed NC clay, the su profile measured usingthe ball penetrometer shows a comparable suprofile to that obtained using the

20 mm (length) x 5 mm (diameter) T-bar (seeFig 5.4(a))

5.3 Effect of time on soil characteristics of a footprint

The soil is remoulded during spudcan penetration and extraction The extent ofremoulding is dependent on the penetration depth of the spudcan which, inturn, corresponds to the soil strength and desired preload pressure Under asimilar preload pressure, deeper spudcan penetration is expected in a weakersoil

In all cases, the spudcan was penetrated into the model ground to amaximum preload pressure, qo of 160 kPa This was followed by an immediateextraction (less than 1 day in prototype time) for case OP 0 or holding at0.75qo (120 kPa) for 2 year for case OP 2 The corresponded loaddisplacement responses for the simulation of OP 0 and OP 2 are shown inFigure 5.5

Details of all the tests presented in this chapter are summarised inTable 5.1 All the test results presented hereinafter are in prototype units usingappropriate scaling laws (Taylor, 1995) from the model units, unless otherwisestated It is worth noting that the time in the model test is scaled up by N2based on the time scale for consolidation, where N is the ratio of centrifugalacceleration to gravitational acceleration

5.3.1 Normally consolidated clay

A footprint was formed after the spudcan was completely lifted out of themodel ground The extent of soil disturbance beneath the footprint was then

evaluated by performing ball penetrometer tests To evaluate the changes insoil shear strength, the shear strength ratio, Rsu defined as suof the footprint to

su of the undisturbed soil at the same position is adopted

By comparing the shear strength profile at various radial distancesfrom the footprint centre at an elapsed time of 1 year, it can be seen that thesoil within the spudcan area (see Figs 5.7(a) and (b)) has been substantiallyremoulded where the shear strength is roughly half of the undisturbed shearstrength The remoulding effects at 0.75D and further from the footprint centre

(see Figs 5.7(c) to (f)) are comparatively less significant Hence, it can beconcluded that the degree of soil disturbance beneath a footprint decreaseswith increase in radial distance, Rd In terms of disturbance in relation todepth, within the spudcan area (0.25D and 0.5D), the strength reduction is notobserved at a depth of 0.2D below the spudcan base level, de de is thepenetration depth at the desired preload pressure, do (measured from themudline to the L.R.P.) plus the thickness of the spudcan conical base, (seeFig.5.1) In fact, the soil at these depths gradually gains strength with time asindicated by Rsubeing greater than 1 For the 100-year long-term case, the soilshear strength increases to a maximum value of about twice the undisturbedshear strength Interestingly, though no significant remoulded effect isobserved after 1 year at 0.75D, the soil has gained considerable strength after

100 years No significant change in shear strength at 1.5D is observed in boththe short-term and long-term In summary, the soil within the spudcan area(0.25D and 0.5D), which experiences greater remoulding as reflected in lower

Rsu values compared to those located beyond 0.75D, gains strength in the term The soil at positions 1.25D and 1.5D, which locates outside the heavilyremoulded zone, exhibits insignificant change in shear strength in both theshort-term and long-term This finding suggests that the soil beyond 1.5D fromthe footprint centre remains more or less intact with no significant disturbancedue to spudcan activity

long-5.3.1.1 Comparison between soil conditions of a footprint formed with a

very short (<1 day) and 2 year operation periods

A similar test procedure was adopted to conduct test NC 2 except that thespudcan was held at 0.75qo for 2 years prior to the extraction, where qo is the

maximum preload pressure, 160 kPa The measured su profiles are presented inFigs 5.8(a) – (f) and the corresponded Rsu values against depth ratio arepresented inFigs 5.9(a) – (f) The results show a similar trend of soil strengthchanges as that observed in test NC 1 That is the su profile is weaker than theundisturbed soil in the short-term (1 – 5 years) and becomes stronger in thelong-term (100 years) However the soil below de exhibits higher shearstrengths as compared to those observed in test NC 1.

To investigate the effect of operational period on the soil strengthchanges, a comparison of the soil condition between tests NC 1 and NC 2 wasmade by mapping Rsu with respect to depth and radial positions The Rsucontour maps for both tests are presented inFigs 5.10(a) – (d)for an elapsedtime of 1 year and 100 years, respectively The extent of soil disturbance can

be explained by examining the soil failure patterns during a spudcaninstallation in soft clay Purwana (2007) found that the major lateral soildisplacement extended to about 1.3D from the spudcan centre at penetrationdepth, d = 0.1D, and the lateral extent gradually decreased to 0.75D at

d = 0.75D where deep failure mechanism was initiated This observationagrees with the earlier finding by Hossain et al (2004) who found that the soilflow altered from shallow failure to deep failure mechanism at d/D = 0.75.During spudcan extraction, after the separation between the underlying soiland the uplifting spudcan occurred, the soil on top of the spudcan flowedlocally around the spudcan edge with a fan zone of 0.25D (Purwana, 2007).Soon after the spudcan was extracted from the model ground, an abrupt soilcircular slide occurred, where the near surface soil at the sides of the spudcan

slid into the cavity, forming a several meters deep bowl-shape depression(Purwana, 2007).

The above four stages of soil failure mechanisms during spudcanpenetration and extraction can be expressed as a series of schematic diagrams,shown inFig 5.11 For the OP 0 and 1 year case (seeFig 5.10(a)), the lateralextent of soil disturbance at shallow depth of up to 0.5D is relatively wide with

a Rd/D of greater than 1.5D that corresponds well with the soil failuremechanism at shallow depth After d/D = 0.5, the apparent soil disturbance isconfined within Rd /D = 0.75 that also corresponds well with deep failuremechanism On the other hand, for OP 2 and 1 year elapsed time case (seeFig.5.10(c)), the soil having higher shear strength than the undisturbed soil isobserved at and below de within spudcan area (Rd/D = 0 to 0.5).A higher pullout force required for OP 2 during the extraction is also evidence of thestronger underlying soil (Fig 5.5(b)) No significant difference in Rsu profilefor Rd/D ≥ 0.75 is observed as compared to case OP 0 Similarly, for the 100-year case, a higher shear strength gain is observed for OP 2 than OP 0,particularly below d/de= 1 and confined within the spudcan area (see Figs.5.10(b) and (d)) The difference in Rsu profile from 0.75D to 1.75D for bothcases is less significant

The above comparison reveals that the maintained pressure during theoperational period has a greater effect on the soil within the spudcan area thanthe soil outside the spudcan area The phenomenon of soil gain-in-strengthwith time has been examined by monitoring the pore pressures changesthroughout the entire test These results will be further discussed in thesubsequent section

5.3.1.2 Pore pressure monitoring in NC clay

In order to investigate the changes in pore pressure during spudcan activity,five (5) pore pressure transducers, namely P1 – P5, were pre-embedded intothe beam channel around the site of test NC 5 prior to the clay slurryplacement P4 and P5 were placed within the spudcan area, at 0.25D from thespudcan centre, whereas P1 – P3 were placed at radial distances of 1.3D to1.45D from the spudcan centre at three different elevations The position of P1– P5 are shown in Fig 5.1(a) The changes in excess pore pressure, ue (inexcess to hydrostatic pressure) of the entire simulation including the spudcanpenetration, 2-year operational period at 0.75qo, and followed by the spudcanextraction, were monitored The measured ue and the corresponding preloadpressure against the model time are shown in Fig 5.12 The generated ueatvarious penetration stages are shown in Figs 5.13(a) and (b) It is observedthat P4 and P5 indicate relatively high excess pore pressure compared to thoseinstalled outside the spudcan area This is because the soils within the spudcanarea experience high compressive and shear stresses from the penetratingspudcan Unlike P4 and P5, ue measured at P1, P2 and P3 may be due only toshear stresses induced by the penetrating spudcan, hence, resulting in lowermagnitudes This finding explains why the soils (within and near the spudcan)which experience greater stresses would gain more strength compared to thosefurther from the spudcan in the long-term, due to greater positive ue beinggenerated

In short, positive excess pore pressure is generated during the spudcanpenetration as the soil is being compressed and sheared With time, the excesspore pressure dissipates and the soil gains in strength The effect of

remoulding and re-consolidation on soil strength can be explained usingcritical state soil mechanics, as illustrated in Fig 5.14 During spudcanpenetration, water can be gradually entrained as the spudcan remoulds the soil.Two components of excess pore pressures are generated by increase in totalstress (compressive) and shear stress respectively During pore pressuredissipation, the effective stress tends towards the applied total stress and thistotal stress is higher if the preload pressure is present When soilreconsolidates, the water content under a given effective stress lies on anunloading-reloading line (URL), and the dissipation of positive pore pressuretakes the soil down follow URL to a lower moisture content (or lower specificvolume than the initial state) and results in a higher undrained soil strength It

is obvious that the dissipation of the excess pore pressures in the soil beneaththe spudcan is directly related to the duration of operation, size of the spudcanand the soil permeability As discussed earlier, the operational period varieswith the type of work the rig is engaged in The coefficient of consolidationcan vary widely for different clays, for instance, 40 m2/year for kaolin clayused in NUS and 2 m2/year for kaolin clay used in UWA Hence, it may bemore appropriate to describe the degree of pore pressure dissipation in term of

an adjusted time factor than the time Zhou (2006) used the adjusted timefactor,  for the soil consolidation during the waiting time (or operationalperiod defined in this study) in the numerical modelling of spudcan extraction.This factor, which was often used for soil consolidation in surface footing(McNamee and Gibson, 1960a,b; Schiffman et al., 1969), has a form of

where cis the adjusted coefficient of consolidation, t is the consolidation timeand R is the radius of the circular footing For simplicity, in this study, thevalue of c is assumed to be the coefficient of consolidation, c v of the clay.Two adjusted time factors, 1 and 2 are considered for soil consolidationduring the operational period and during the elapsed time after a footprint wasformed, respectively Fig 5.15(a) shows the dissipation of excess porepressure during the operational period About 25-35% of ue dissipates at theend of the 2-year operation period or at ~ 0.075 The soil gains strength asthe ue dissipates This explains why the 1-year case for OP 2 exhibits a slightlyhigher su value compared to the 1 year case for OP 0.

During spudcan extraction, a relatively minor negative excess porepressure is generated for soil outside the spudcan area, as measured by P1 – P3(see Fig 5.13(a)) For soil beneath the spudcan base, substantial suction isbuilt up during the pull out as indicated by the negative ue values (see Fig.5.13(b)) However, after the separation between the underlying soil with theextracting spudcan occurs, ue returns to positive with a magnitude slightlylower than that at the onset of extraction As a result, a net positive excess porepressure is left in the soil after the spudcan is completely pulled out from themodel ground Fig 5.15 (b) shows the degree of dissipation of net ue duringthe elapsed time after the footprint was formed In this test, the net ue has fullydissipated after 30 years or  ~ 1.13 for P1 – P5 This explains the soilsbeneath the footprint become stronger than the (intact) undisturbed groundafter 100 years for both tests with different operational periods

be observed beyond 1.5D In the long-term (elapsed time of 100 years), wherethere is time for the remoulded soil to re-consolidate, the soil generally regainssome strength The amount of strength gain is dependent on the location of thesoil relative to the footprint The soil above de, which has been excessivelydisturbed during the initial spudcan penetration and extraction, the soil regainssome strength with time but the strength is still smaller than that of theundisturbed soil The soil below de and within the spudcan area, whichexperiences an increase in total stress (couple of compressive and shearstresses) during the initial spudcan installation, is found to be slightly strongerthan the intact undisturbed soil with a maximum Rsuvalue of 1.3.

For test OC2, a very similar soil condition as observed in test OC1 isnoted except for the soil below de and within spudcan area is found to beslightly higher than those observed in test OC1 for both short and long-terms.For the 1-year case, the major soil disturbance zone is similar for both OP 0and OP 2, which is confined within de in depth and 1.25D in radial distance.However, the soil in OP 2 exhibits a slightly higher strength than OP 0 asindicated by the upward shifting of Rsu = 1 line Similarly, for the 100 year-case, the greater soil strength is obtained in OP 2 than OP 0 This isparticularly the case for the soil within spudcan area and below de in case OP 2than OP 0 Likewise for test NC2 (case OP 2), the soil undergoesconsolidation during the operational period may attribute to this higher Rsuvalues.

5.3.2.1 Pore pressure monitoring in OC clay

Four (4) pore pressure transducers, namely O1 – O4 were installed within thesample for test OC5 to investigate the changes in pore pressure during spudcanactivity Fig 5.21 shows the position of O1 – O4 PPT O4 was placed withinthe spudcan area, at 0.25D from the spudcan centre, whereas O1 – O3 wereplaced at radial distances of 1.3D to 1.42D from the spudcan centre at threedifferent elevations The measured ue and the corresponding preload pressureagainst the model time at various stages of spudcan activity are shown inFig 5.22 The generated ue at various penetration stages are shown inFigs 5.23(a) and (b) It is observed that O4 indicates relatively high excesspore pressure compared to those installed outside the spudcan area (O1 – O3).This is because the soils within the spudcan area experience high compressiveand shear stresses from the penetrating spudcan The ue pore pressure

dissipates with operational time At the end of 2 year operation or 1 of about0.076, the 25 – 40% ue generated by spudcan preloading has dissipated and thesoil gains some strength (Fig 5.24(a)) This agrees with the earlier findingwhere the soil condition for OP 2 is slightly stronger than that for OP 0.

Compared to NC clay, a relatively negligible net positive ue remained

in the soil after the spudcan was completely pulled out from the ground.Fig 5.24(b) shows the degree of dissipation of net ue during the elapsed timeafter the footprint was formed In this test, the net ue has fully dissipated after

30 years or  ~ 1.13 for P2 – P5 This explains why the soils beneath andaround the footprint became stronger in the long-term than in the short-term.However, unlike NC clay, the soil condition of the footprint formed in OCclay in the long-term is still much weaker than the intact undisturbed soil

5.3.2.2 Comparison between shear induced pore pressure in NC and OC

clays

PPTs P1 – P3 (test NC5) and O1 – O3 (test OC5) are all located outside thespudcan area at radial distance of 1.3 – 1.45D from the spudcan centre Theexcess pore pressures are generated due to the shear stresses induced duringspudcan penetration.Fig 5.25 shows the shear induced pore pressures duringspudcan penetration in NC and OC clay The (dp-d)/dp = 1 represents theL.R.P of the spudcan is at the mudline; (dp-d)/dp = 0 represents the spudcan is

at PPT elevation; and (dp-d)/dp < 0 indicates the spudcan penetrates below thePPT elevation dpis the depth measured from the mudline to the PPT level, d isthe spudcan penetration depth measured from the mudline to the L.R.P and q

is the corresponding preload pressure In general, ue increases with q when thespudcan is approaching the PPTs elevation For NC clay, the ue/q increases as

the spudcan approaching the PPT elevation and reaches a maximummagnitude when the spudcan is about at the PPT elevation Higher shearinduced ue is observed in NC clay than OC clay This clearly indicates the role

of stress history in the soil shearing behaviour In NC clay (OCR = 1), the soiltends to compress under shearing resulting in higher shear induced porepressure compared to that in OC clay and this leads to a higher shear strengthgain in the long-term after the generated pore pressure dissipates

5.4 Effect of stress history on the soil strength changes with time

It is possible to interpret the effect of stress history on the soil propertychanges in both NC and OC clays by comparing the pattern of soil propertychanges with time from the test results presented in the earlier sections On theone hand, the NC clay was prepared in such a way that it was consolidatedunder its own effective weight at 200g Upon completion of the primaryconsolidation, the clay has a pre-consolidation stress, p’ equals to the initialeffective vertical stress, vo’ that gives OCR = 1 throughout the entire claydepth.On the other hand, after a pre-consolidation with 30 mm thick sand at300g, the OC clay has an OCR profile with a value of 12 at 1 m and 2 at 20 mbelow the mudline based on a constantvo’ of 6.5 kN/m3 For OP 0 and 1-yearcase (see Fig 5.10(a) and Fig 5.20(a)), the extent of soil disturbance isreasonably comparable in both clays It appears that when the time involved isrelatively short with negligible dissipation of ue in both operational period andelapsed time after a footprint is formed, the extent of soil disturbance within afootprint is well defined by the soil failure patterns at different depths duringspudcan penetration and extraction However, when the time is long enough

for a significant degree of ue dissipation to take place, the stress history of theclay has a considerable effect on the behaviour of soil property changes In NCclay, it is found that the soil is generally stronger than the undisturbed soilespecially within the spudcan area and below de This is because the stressesgenerated during spudcan activity have altered the stress history of NC clay to

an over-consolidated profile in the long-term and upon completion of the clayre-consolidation under the new v’ In contrast, the change in soil property asthe excess pore pressure dissipates in the OC clay is more complicated Thestructure of OC clay above de may be destroyed during the spudcaninstallation and extraction activity Moreover, only relatively insignificant uewas generated during the spudcan activity, after it has re-consolidated underthe newv’, the soil is unable to restore to its initial strength However, for thesoil below de, where the soil experiences an increase in total stress by thespudcan preload pressure (which is higher than the initial p’) becomes moreover-consolidated than the initial stress condition after completion of v’ re-consolidation Hence, a stronger soil profile than the undisturbed soil results

5.5 Effect of time on spudcan-footprint interaction

In this section, the influence of footprint conditions on the new spudcaninstallation is investigated A spudcan footprint consists of two maincharacteristics: an uneven seabed profile and a highly non-uniform shearstrength profile In the earlier sections, it is found that the soil condition of afootprint changes with time as the excess pore pressure dissipates and the soilstrength changes with time is dependent on the initial pre-consolidation stress.All these factors affect the manner in which a spudcan interacts with a

footprint Four scenarios were identified: rig revisits shortly (e.g 1- 3 years)and a very long period after footprints were formed under two operationalconditions (OP 0 and OP 2) for both NC and OC clays The same spudcan wasused to perform re-installation at 0.5D offset from the footprint centre Theinteraction between the new spudcan installation and the footprint wasevaluated in term of vertical load, V, horizontal force, H and moment, Macting at the spudcan reference point (see Fig 5.1(b)) The structure preload

on the spudcan was measured by the axial strain gauge, whereas the inducedmoment and horizontal force were evaluated through extrapolation from thebending moment profile obtained from the bending gauges instrumented onthe model leg As the model leg was rigidly fixed to the vertical actuator nolateral displacement at the connection was possible Therefore, the likelihood

of sliding is indicated by H while the tendency of overturning (or thelikelihood of over-stressing structures if the movement is resisted) is indicated

by M If these forces are adequately resisted by the structure and the rig as asystem, where an additional lateral restraint can be obtained by pre-installingthe other leg(s), the spudcan can be installed with minimal positioningproblem If these forces are structurally unsustainable, the rig installation maysuffer structure damage and/or unsatisfactory rig positioning and movements.Catastrophic failure may occur if the installation is close to the fixed platform

Previous findings reported by Stewart & Finnie (2001) showed thatrelatively high adverse forces were developed when a spudcan was installed atoffset distance from 0.5 to 1D from the original installation site or thefootprint centre This agrees with Cassidy et al (2009) who found that theworst location for reinstallation to be at an offset of 0.5D from the initial

spudcan installation Hence, the 0.5D centre-to-centre used in the testsreported in this study is deemed to be within the critical distance The testdetails are presented in Table 5.1 and the test results are summarised inTable 5.2.

5.5.1 Normally consolidated clay

Footprints formed in NC clay consist of a concave surface profile with alowest level of 3 m or 0.2D below the initial mudline and a highly non-uniform shear strength profile that changes with time Owing to sometechnical problems, there was a delay in testing the spudcan re-penetration forcase OP 0 The re-penetration was unable to be conducted at 1 year (prototypetime) after the footprint was formed, but was carried out at an elapsed time of

3 years For a better illustration of the footprint conditions prior to spudcan penetration at 0.5D offset from the footprint centre, su contour maps of fourcases were produced and are presented in Figs 5.26(a) – (d) The footprintprofiles shown were approximated using a laser device scanning across themodel ground surface during test NC 5 In the short-term (e.g 1 – 3 year, seeFigs 5.26(a) and (c)), it is observed that above d/de= 1, the contours of suslope towards the footprint centre indicating a weaker soil zone confinedwithin Rd/D = 0.75 due to soil remoulding process In the long-term, the soilwithin this zone becomes stronger than those outside this zone (Rd/D > 0.75).This is reflected in the su contours being gentler in gradient above d/de = 0.5and the contours sloping away from the centre d/de > 0.5 – 1.4 (see Figs.5.26(b) and (d)) It is worth noting that a localized ‘crust’ is observed below de(with a thickness of up to 0.4D thick) within Rd/D = 0.5 for OP 0 and OP 2cases

re-Figs 5.27(a) – (d) show the measured vertical load, V of the initialpenetration and re-penetration at 0.5D for all four cases In the short-term, theload-displacement response for re-penetration is softer than the initialpenetration by a reduction of up to 20% Consistent with the observation thatthe soil gains strength as excess pore pressure dissipates, a stiffer load-displacement response is observed for the long-term re-penetration case Nosignificant difference in V is observed for both OP 0 and OP 2 This may bedue to the duration of 2 years for the operational period is only permiting(dissipation of excess pore pressure of about 20-30% due to low permeability).

In addition, the operational period has an effect on the soil property changesfor the soil within the spudcan area As an offset of 0.5D was adopted in thesecases, the influence of operational period on V is deemed to be lower

Figs 5.28(a) – (b) show the induced horizontal force, H duringspudcan re-penetration at 0.5D centre-to-centre under the four scenarios.According to the sign convention illustrated inFig 5.1(b), positive H denotes

a likelihood of spudcan sliding towards the footprint centre, whereas negative

H denotes a likelihood of sliding away from the footprint centre For term penetration cases, a positive H profile is observed above depth ratio d/de

short-= 0.6 and the H becomes negative with deeper penetration with a maximumvalue of -0.64 MN for OP 0 and -0.95 MN for OP 2 occurring near d/de = 1.Despite the gain in soil strength in the long-term, the H profile for the long-term case is surprisingly less critical than the short-term case The aboveresults imply that the spudcan may have the tendency of sliding towards thefootprint centre at the beginning of the installation, but when the stronger claybelow d/de starts to influence the interaction, the spudcan tends to slide away

from the footprint centre The tendency of sliding may decrease with increase

in elapsed time before the visit of a future rig

The induced M for all four cases are shown in Figs 5.29(a) – (b).Positive M denotes a likelihood of spudcan rotation with the rotating directionagainst the footprint centre as shown inFig 5.1(b) For the short-term case, apositive M profile is observed up to a depth of 0.9D before it becomesnegative with a maximum value of -13 MNm for OP 0 and -15.4 MNm for OP

2 For the long-term case, in contrast, the M profile is found to bepredominantly negative with a maximum value of -17.9 MNm for OP 0 and -22.4 MNm except for a small ‘hump’ on the positive side (< 4 MNm)indicating the effect from the crater This agrees with the finding stated inChapter 4, which the crater was found to be less influential on the spudcan-footprint interaction in soft clay In a uniform ground, applied moment acting

on a shallow foundation is often considered as an eccentricity of the appliedvertical load However, in a spudcan-footprint interaction problem, therotational moment is induced by the non-uniformity of the ground bearingresistance, where the resultant bearing resistance tends to rotate the spudcanfrom the stronger soil side to the weaker soil side Hence, the M profiles areaffected by the soil variation of the footprint The footprint soil conditionchanges with time (from disturbed soil in the short-term to a stronger than theundisturbed soil in the long-term) and the degree of soil disturbance and gain-in-strength vary across the depth and radial distance, as illustrated in the su

maps(see Figs 5.26(a) – (d)) This explains why relatively high M occurrs atdepth below de for the re-installation in the long-term than in the short-termwhere the variation in su across the spudcan is greater

5.5.2 Over-consolidated clay

Footprints formed in the OC clay in the present study consist of a concavesurface with a lowest level of 2.3 m or 0.16D below the initial mudline Thisprofile was measured by a laser scanning device during tests OC 1 and 2.Similar to NC clay, four cases were studied and the corresponding su contourmaps for all cases prior to the spudcan re-penetration are presented inFigs 5.30(a) – (d) For the short-term cases (Figs 5.30(a) and (b)), as themajor disturbance is confined within Rd/D = 0.75, the su contours essentiallyslant towards the footprint centre following a profile similar to the craterprofile In the long-term, a relatively gentle su profile is observed as a result ofreconsolidation of the remoulded soil The soils within the spudcan area whichwere heavily remoulded have gained significantly more strength than the soilsaway from the spudcan, and this has resulted in a smaller differential in su It isalso worth noting also that the soil beneath deand confined within Rd/D = 0.5shows slightly higher shear strength than the undisturbed su When the samespudcan penetrates at an offset distance of 0.5D, a weaker load-displacementresponse than that of the initial penetration is observed due to the weaker suprofiles The V profiles during initial penetration and re-installation at 0.5Dfor the four cases are presented inFigs 5.31(a) – (d) The reduction in V is up

to 40 % for the short-term case and 30 % for the long-term case The OP 2case gives slightly higher V response than OP 1 cases as the soil for OP 2 isstronger

H and M profiles of the re-penetration of the four cases are shown inFigs 5.32(a) – (b) and 5.33(a) – (b) respectively In general, both H and Mshow a similar trend for the short and long-term cases Above d/de = 1,

relatively high H and M values are observed for the short-term (1 year) case.For d/de = 1 to 1.4, the long-term (100 years) case is, in contrast, noted to bemore critical In the short-term, H increases with penetration depth to amaximum of 0.9 MN for OP 0 and 0.8 MN for OP 2, indicating a likelihood ofthe spudcan sliding towards the footprint centre The long-term cases show ahigher tendency of sliding away from the footprint centre occurring at aroundd/de = 1 with a maximum negative H of -0.4 MN for OP 0 and -0.6 MN for OP

2 Similar to H, M increases with penetration depth (up to d/de= 0.8) in term to a maximum value of 15 MNm for OP 0 and 11 MNm for OP 2,indicating the variation in bearing resistance where the stronger soil is locatedoutside the initial spudcan area Below d/de= 0.8, the long-term case shows anegative M profile to a maximum value of -12 MNm for OP 0 and -13 MNmfor OP 2 The trend of spudcan rotation agrees well with the soil variationsshown in the su contour maps

short-5.6 Practical implications – Spudcan-footprint interaction in

NC and OC clays

If a new spudcan is installed partially overlapping with an old footprint in NCclay, the spudcan will when compared to the initial penetration eitherexperience weaker load-displacement response in the short-term or strongerresponse in the long-term depending on the elapsed time between the new andthe previous spudcan installation (or more appropriately the degree of uedissipation) The spudcan will experience relatively high H and M below theprevious rig’s penetration depth The practical implication is that if theseforces are structurally unsustainable for a rig to have a safe installation at adesired position, one may consider a rig with lower required preload pressure

where the rig installation can be terminated at a sufficient distance above de.Hence, an accurate estimation of V for the re-penetration is deemed essential.

In OC clay, for penetration depth above de, a spudcan which isinstalled partially overlapping with the footprint will experience weaker load-displacement response regardless of the length of elapsed time The spudcanwill likely experience high H and M above de If these forces are found to behazardous to the new rig installation, one shall consider avoiding the footprints

or use an identical rig to re-install at the same position of the previous rig asrecommended in SNAME (2002) If the spudcan-footprint interaction isunavoidable due to site constraints, one may consider to carefully position thenew rig in such a way that the critical positions (0.5 – 1.0 times spudcandiameter offset from the centre of the initial site) can be avoided (the effect ofoffset distance will be discussed in Chapter 6) If that fails other footprintmitigation measures would need to be adopted However, only limited studies

on the spudcan-footprint interaction mitigation methods (e.g Jardine et al.,

2001 & 2002, and Dean and Serra, 2004) and leg inclination alerting systemusing Rack Phase Difference (RPD) monitoring (see for example Foo et al.,

2003) are available in the public domain The effectiveness of those methods

is still in question Hence, it is recommended that further studies should becarried out on the spudcan-footprint interaction mitigation methods withconsideration of the identified footprint characteristics and the soil failuremechanism as reported in Chapter 4 and the time effect on the soil propertychanges

In reality, a footprint condition can be altered by many factors such asthe spudcan extraction methods, soil volume trapped within trussed leg, soil

stratification, physical properties of soil (e.g coefficient of consolidation,sensitivity, and stress history), regional sedimentary process etc It is thusrecommended to produce su contour maps for a site, which the behaviour ofspudcan re-penetration can be more accurately estimated based on the actualsoil condition The question arises is that the cost and time involved may betoo high to make it unrealistic to conduct soil tests as many as what were done

in the model tests A carefully planned offshore site investigation is deemedessential The earlier findings revealed that a zone of soil experienced moreextensive changes in su within 1.5 times spudcan diameter or 0.75D from thefootprint centre This finding suggests that the site investigation shouldconcentrate within this zone and to have at least 2 tests located at eachfootprint (e.g 0 – 0.75D from the footprint centre)

both the short and long-term The soil further away from 1.5D (from thefootprint centre) is likely to remain intact with no significant disturbance fromthe spudcan activity The soil underneath the spudcan base undergoesconsolidation during the operational period and formed a ‘localized crust’beneath the spudcan base This ‘crust’ layer has an adverse effect on thespudcan re-installation partially overlapping with it as larger moment acting

on the spudcan due to considerable variation in bearing resistance at this level.The strength of this crust increases with the duration of jack-up operationalperiod This consequently produces a higher moment than expected

The initial soil condition of the footprint in OC clay is found to besimilar to that observed in NC clay It is then postulated that if the timeinvolved is sufficiently short with negligible dissipation of ue in bothoperational period and elapsed time after a footprint is formed, the extent ofsoil disturbance within a footprint corresponds well to the soil failuremechanisms during the spudcan penetration and extraction However, whenthere is a considerable length of time involved, such as prolonged operationalperiods or the new rig installation takes place many years later, the soilcondition of a footprint in OC clay is clearly different from that observed in

NC clay Compared to NC clay, relatively less shear induced pore pressurewas generated surrounded the footprint in OC clay Hence, the OC clay in theheavily remoulded zone (confining within 1.5D in width), though gainingsome strength in the long-term, the shear strength is still considerably lowerthan that of the intact undisturbed soil

The effect of time on spudcan re-installation at 0.5D centre-to-centrewas investigated In NC clay, the spudcan re-installation experiences a weaker

vertical load response in the short-term and a stronger response in the term Larger horizontal force is noted for the re-installation in the short-term,whereas a larger moment is observed for the re-installation in the long-term.Maximum H and M, for the range of conditions tested occur at a depth close tothe initial penetration depth, de The tests with simulation of 2-year operationalperiod give higher maximum H and M, suggesting that the operational period

long-of previous rig plays a critical role on the safety long-of new spudcan installation.The above observations can be generalized to an important implication on theplanning of new rig installation: a lower preload pressure rig may be a betterchoice for a site consisting of soft normally consolidated clay This is toensure that the rig penetration can be terminated at a sufficient depth above theidentified critical soil zone

For OC clay, on the other hand, the spudcan experiences a relativelyhigh H and M at depth above de in the short-term and below dein the long-term This suggests that the adverse effect of a footprint varies with time as thesoil condition changes In order to improve the safety of a new rig installation,carefully selecting a suitable rig and positioning it in a safe distance free fromfootprint interference is then recommended Where this is not possible,suitable footprint preventive measures with sufficient consideration on theidentified footprint characteristics shall be performed to minimize thedetrimental forces from footprints However, studies on the footprintpreventive measures are still lacking Hence, a more systematic study on thepreventative measures accounting for all identified footprint characteristics isrecommended