Centrifuge model study on spudcan footprint interaction 7

In this chapter, an extensiveseries of centrifuge model tests results were analyzed focusing on thefollowing: i the effect of leg flexural rigidity, spudcan diameter and preloadpressure

on this problem was studied in detail in Chapters 4 and 5 As the deployment

of a jack-up rig at various sites depends on the rig availability, it is common tohave different rigs to be deployed to a fixed platform for workover or drillingadditional wells The jack-up configuration is hence an important variable inthe interaction between a footprint and a spudcan In this chapter, an extensiveseries of centrifuge model tests results were analyzed focusing on thefollowing: i) the effect of leg flexural rigidity, spudcan diameter and preloadpressure on spudcan installation at 0.5 times spudcan diameter offset from thefootprint centre; and ii) spudcan-footprint interaction at various offsetdistances Dimensional analyses are conducted to generalise this complexproblem

6.2 Footprint condition prior to future spudcan installation

It is unquestionable that the footprint features affect its interaction with thespudcan installation in the vicinity of the footprint Typically, a footprint hastwo main features: an uneven seabed and the soil shear strength variation In

Chapter 4, it has been established that the shear strength variation of soilbeneath the footprint dominates the interaction in soft to firm clays Shortlyafter a footprint is formed, the soil shear strength varies across the footprinthaving lower strength within the footing diameter and higher strength furtherfrom the footprint centre This soil condition may cause a non-centric spudcaninstallation to have a tendency of sliding towards the footprint centre Studies

inChapter 5 revealed that the degree of the soil variation changes with time asthe excess pore pressures generated by the previous spudcan installationdissipate The behaviour of shear strength changes with time is dependent onthe initial stress-history of the soil However, when the time involved, either inoperational period and elapsed time after a footprint is formed, is sufficientlyshort (negligible excess pore pressure dissipation), the soil condition beneaththe footprint for both NC and OC clays is found to be similar Previousfindings in Chapter 5 revealed that, above the initial spudcan penetration depth(de) in OC clay, the re-installed spudcan experiences higher moment andhorizontal forces in the short-term than the long-term This is because the soilthat is previously heavily remoulded (within spudcan area) regains somestrength in the long-term and these results in a less non-uniform soil profileacross the footprint

All the tests presented in this chapter were conducted on clay sampleswith an over-consolidated profile Only the short-term cases with nil

operational period (which yields higher horizontal force and moment) areconsidered A generalised short-term soil condition of a footprint is producedbased on five (5) tests (the test details are summarised in Table 6.1) Threetests were conducted on Malaysia kaolin clay, whereas the other two testswere conducted on UWA kaolin clay It should be noted that the coefficient ofconsolidation for both clays and the spudcan used are not the same Hence, theadjusted time factor,  is used to characterize the consolidation status of thesoil after the extraction of spudcan (Table 6.1) Figure 6.1 shows thegeneralised short-term soil condition of a footprint formed in soft to firm claysfor  < 0.002 and  < 0.2, where 1 and 2 are the adjusted time factors forsoil consolidation during the operational period and during the elapsed timeafter a footprint is formed, respectively.1is the operation duration of previousrig which depends on the type of work involved, whereas2is the elapsed timeafter footprints formed to a time of a new rig installation Hence, 1and2 areindependent from each other The footprint generally consists of a crater withthe lowest point of about 0.2D deep and a highly non-uniform soil beneath thecrater surface The degree of soil disturbance is classified by the shear strengthratio, Rsu, which is the ratio of the footprint su to the undisturbed su at the sameelevation.

The strength ratio of less than 0.5 is categorised as heavily remoulded, which

is basically confined within the spudcan area (Rd/Df ≤ 0.5) Dfis the diameter

of the spudcan used to form the footprint The soil with strength ratio of 0.5 to0.7 is classified as moderately remoulded, whereas a ratio range of 0.7 to 0.9 is

classified as less remoulded The extent of radial soil disturbance varies withdepth As discussed in Section 5.3.1.1, the soil failure mechanism changesfrom shallow (or general) failure to deep (or localised) failure when thespudcan penetration depth increases This defines the extent of soil disturbance

as reflected in the su measurements across a footprint At depths of up to 0.5Df,the radial disturbance is found to extend to Rd/Df of 1.5 Below this depth, themajor radial disturbance (Rsu < 0.7) is found to be confined within Rd/Df =0.75 In term of vertical extent, it is found that the soil is heavily remoulded to

a depth of up to 1.1de, where deis the penetration depth to spudcan base level.Below this depth, a minor soil strength reduction (0.7 < Rsu < 0.9) extending

up to 0.3Df is observed In short, the above describes the soil condition prior tothe future spudcan installation for all tests which will be presented later in thischapter

One of the important variables in this study is spudcan diameter, D.Each complete test involved two spudcan penetrations in which the firstpenetration was to form the footprint and the second penetration was toinvestigate the spudcan-footprint interaction Df denotes the spudcan diameterfor the first penetration, whereas Ds denotes the spudcan diameter for thesecond penetration For all the tests presented in Sections 6.3 and 6.4, thesame spudcan was used to perform the first penetration (to create a footprint)and the second penetration Hence, D (= Df = Ds) is used to denote the spudcandiameter for both penetrations A schematic diagram of the typical testarrangement is shown inFig 6.2

6.3 Effect of jack-up rig configuration

The main interest of the study in this section is to investigate the behaviour of

a spudcan installation partially overlapping with a footprint The footprintproblem is essential a soil-structure interaction problem Both the footprintcondition and the structural configuration can affect the manner of theinteraction Hence, in this problem, two conditions need to be considered:i) the footprint characteristics as discussed inSection 6.2;

ii) the jack-up configuration such as the leg flexural rigidity, spudcan

size and rig preload

In an ideal condition, a footprint should be axisymmetric in plane When across-section through the footprint centre to the spudcan centre is considered,the interaction is symmetrical along the longitudinal axis The probable soilfailure mechanism during the spudcan re-penetration has been discussed in

Chapter 4 As the leg-hull connection is modelled as fully rigid where nolateral displacement and rotational movement are allowed, the spudcan-footprint interaction is evaluated in term of three major ‘resultant’ loadcomponents (vertical force, V, horizontal force, H and moment, M) acting atspudcan level These forces were measured by the strain gauges instrumented

on the spudcan leg (seeAppendix A for detail)

6.3.1 Leg flexural rigidity

It has been suggested that leg flexural rigidity will influence the interactionbetween a spudcan and a footprint (Stewart and Finnie, 2001; Foo et al 2003;Dean and Serra, 2004) Stewart and Finnie (2001) pointed out that if the jack-

up leg is fully flexible, the spudcan will tend to slide into the adjacent crater

On the other hand, if the leg is fully rigid, the spudcan will penetrate vertically

into the soil with large lateral force build-up Foo et al (2003) postulated thatthe infinitely stiff leg would take the most bending moment as compared toone that was fully flexible.

To investigate the effect of leg stiffness on spudcan-footprintinteraction; two tests, namely P1 and P2, were conducted on soil samples thathaving similar strength profiles The model leg used in test P1 (Leg 1) has aflexural rigidity, EI, of 1.22×1012Nm2 which is 15.6 times stiffer than the EI

of the model leg used (Leg 2) in test P2 A 10 m diameter spudcan was usedand the penetration depth required to achieve 460 kPa preload pressure was8.7 m and 8.65 m for tests P1 and P2, respectively (Fig 6.3a) The test detailsand results are summarised inTables 6.2 and 6.3, respectively.Figs 6.3b and

c show the H and M profiles for Leg 1 and Leg 2, respectively For both legs,the induced H increases with penetration to a maximum of about 1.1 MN (Leg1) and 1.08 MN (Leg 2) On the other hand, a larger M was obtained for Leg 2

as compared to that of Leg 1 indicating the leg flexural rigidity has a moresignificant effect on the M than the H As it is not feasible to install thedisplacement transducer on the spudcan (to avoid influence to the spudcanmeasurements), the tip displacement (at L.R.P.) for both legs is computedusingeq (A.17) in Appendix A.The tip displacement plot is presented in Fig.6.3d It is apparent that Leg 2 deflects with a much greater amplitudecompared to Leg 1 Hence, the stiffer the leg is, the less deflection takes placeunder a similar loading condition It is postulated that the horizontal force may

be reverse of the horizontal displacement as the force induced decreases withincrease in displacement (Stewart and Finnie, 2001; Jardine et al., 2001) Inthis case, the effect of tip displacement on the induced H is not obvious as

there is only a small deviation in H throughout the penetration depth up to de.This may be due to the limitation of the experimental set-up where nohorizontal displacement is allowed by the rigid connection between the leg andthe vertical actuator The tip displacement involved is potentially lower thanthe movement of a jack-up unit as a system and hence, the effect of horizontaldisplacement on this set-up is discounted Similar observation was made by

Cassidy et al (2009)as they found no distinguished difference in the H profilefor legs having 4 different stiffness values using a similar experimental set-up(rigid connection) However, these findings may only be applicable to therange of leg stiffnesses and forces investigated Beyond these stiffness andloading ranges, the penetration response may be different It is also worthnoting that the comparison in leg rigidity stiffness was made for the sameboundary condition of fixed connection between the leg and the loadingactuator If the connection is not fully fixed, the induced horizontal forcewould be different

6.3.2 Effect of spudcan diameter

Early rigs had 8 to 12 individual legs and each leg was supported by relativelysmall spudcan compared to modern-day rigs Modern-day rigs are designed tohave fewer legs, typically three legs, which lead to a trend of using largerspudcans (Poulos, 1988) According to the evolution of individual footingspresented by McClelland et al (1982), spudcan sizes can vary from 4.8 m (forOffshore No 52 made in 1955) to 20.1 m (for Marathon Gorilla made in1982) In this section, the effect of spudcan diameter will be studied Twospudcan sizes, namely 6 m and 10 m, were selected and four tests wereconducted The test details are summarized inTable 6.2

6.3.2.1 Tests P2 (D = 10 m) versus Test P3 (D = 6 m)

Tests P2 and P3 were conducted in soil samples having similar undisturbedshear strength profiles The same leg (Leg 2) was used in tests P2 and P3 Thespudcan was re-installed at an offset of 0.5Df from the footprint centre It isworth to highlight that both spudcans have the same base angle There are twoscenarios: i) the same preload level for different spudcan sizes, the largerspudcan will yield lower preload pressure; and ii) the same preload pressurefor different spudcan size (e.g the rig with smaller spudcans will have morelegs than that with larger spudcans) In this study, the scenario ii) isconsidered The VHM plots for tests P2 and P3 are shown inFigs 6.3 and 6.4.The bearing response of a foundation is influenced by the degree of non-homogeneity, kD/sum, where k is the rate of increase in shear strength withdepth in kPa/m, D is the foundation diameter and sum is the shear strength atmudline Although the undisturbed soil conditions for both tests are similar,owing to different footing diameters, the 6 m spudcan required less penetrationdepth than the 10 m spudcan in order to achieve the same penetrationresistance It is observed that the penetration depth required for the 6 mspudcan was about 1.5 m shallower than the 10 m spudcan to achieve preloadpressure of 460 kPa For test P3, H increases with depth to a maximum value

of 0.71 MN at a depth of 6 m Similarly, M increases with depth to amaximum value 2.36 MNm at a depth of 6.3 m At the same depth, themagnitude of the VHM for 6 m spudcan is lower compared to those obtainedfor 10 m spudcan as a smaller footing area is involved

Two dimensionless parameters, namely load inclination angle  andnormalised load eccentricity e/Ds are used to evaluate the resultant soil

reaction The angleindicates how much the resultant soil reaction is inclinedrelative to the vertical and it is a ratio of H to V at the corresponding depth.The e/Ds, on the other hand, represents the eccentricity e of the vertical soilreaction from the spudcan centre The derivations of these two parameters arepresented in Appendix A.Figs 6.5a & b show the load inclination,  and thenormalized eccentricity, e/Ds for tests P2 and P3 The correspondingpenetration depth, d is normalised with de A high can be either due to a high

H or a low V In shallow depth penetration or low depth ratio, d/de= 0 – 0.2,because of the existence of the crater and the partial support of the spudcan(attributed to low V), a high  results For depth ratio of 0.2 – 0.8, an almostconstant is obtained for both tests This implies that H increases at a similarrate as V On the other hand,  decreases at depth d/de ≥ 0.8 As the initialshear strength profile increases linearly with depth, V for re-penetration alsoincreases with depth The decreasing below d/de of 0.8 is attributed by thedecreasing H This indicates that less sliding force is produced when thepenetration is beyond the heavily remoulded zone The higher is obtained forthe 6 m spudcan compared to the 10 m spudcan

Similar to , high e/Ds is observed at d/de of up to 0.4, which isattributed by the partial support of the spudcan that results in a high M and alow V After this depth, the e/Ds of both tests is almost constant with a value

of about 0.025 before it decreases to nil

6.3.2.2 Test P4 (D = 10 m) versus Test P5 (D = 6 m)

Similar to tests P2 and P3, the effect of spudcan size using 6 m and 10 mspudcans is examined in tests P4 and P5 but in relatively softer soil samples

The results of VHM are presented inFig 6.6 The 6 m spudcan requires 9.0 mpenetration (measured from the mudline to the spudcan base level), whereasthe 10 m spudcan requires 11.66 m to achieve preload pressure of 460 kPa forthe initial penetration Details of the tests are tabulated inTable 6.2 For bothtests, the H profile increases with penetration depth to a maximum value of 1.3

MN for the 10 m spudcan and 0.685 MN for the 6 m spudcan A typical

‘double-humped’ M profile is obtained where the upper hump is due to thepartially supported spudcan whereas the lower hump (and also the maximumM) is due to the soil strength variation The test results are summarised in

6.3.3 Effect of preload pressure

For earlier rigs, the average bearing pressure lies within a range of 200 to 350kPa (Le Tirant, 1979) For modern rigs, the average vertical bearing pressurescan be in excess of 400 kPa for a fully embedded spudcan (Randolph et al.,2005) Some rigs have considerable preload pressure in the range from 575 to

960 kPa (Poulos, 1988) It can be seen that the preload pressure is a variablethat depends on the rig type Two tests, namely P6 and P7, were conducted to

investigate the effect of preload pressure (for the initial penetration) onspudcan-footprint interaction The same 6 m spudcan was used on the samesoil strength profile that can be expressed as su = 28 + 5z (in kPa) Thespudcan re-penetration was conducted at an offset distance of 0.5Df from thefootprint centre The test details are listed in Table 6.2 As the soil strengthprofiles for both tests are almost the same, the preload pressure of 580 kParequired to penetrate 2.6 m deeper (to a depth of 7.86 m) than the 460 kPapreload pressure (de = 5.27 m), seeFigs 6.8a & b This has an implication onthe footprint condition, which a deeper remoulded region is created for thehigher preload pressure case During the spudcan re-penetration at 0.5D, theinduced H and M for test P7 (580 kPa) are significantly higher than thatinduced in test P6 (460 kPa), see Figs 6.8c & d The test results aresummarised in Table 6.3 In term of load inclination, as shown in Fig 6.9a,higher obtained for 580 kPa case than 460 kPa case indicates an increase intendency of sliding for higher preload case (or higher de) On the other hand,the e/D curves for both tests are almost identical This finding suggests that, inthe same soil profile, the eccentricity is less affected by de.

6.4 Dimensional analyses

It is important to estimate the potential forces induced on the structure duringspudcan installation in the vicinity of footprints for assessment of structureintegrity and stability It has been established that the induced H and M areaffected by spudcan diameter D, preload pressure (or depth of initialpenetration, de), radial offset distance Rd, soil undrained shear strength suandadjusted time factor,  As the soil shearing extent is a function of spudcan

diameter, and the soil shear strength profile is almost linearly increasing withdepth, the undisturbed suat a depth of 1D is used in the normalisation Asdiscussed in Section 6.3.3, the effect of preload pressure is reflected in thepenetration depth, de The Hmaxcan be generalised as follow:

Table 6.3 For all tests studied, the footprint was formed with operationalperiod of less than 1 day or 1 < 0.002 and the spudcan re-penetration wasperformed at 0.5D offset from the footprint centre shortly after the footprintwas formed with 2 < 0.2 As all tests were conducted in soil samples withlightly over-consolidated profile, the undrained shear strength may beidealised as approximately linearly increasing with depth Earlier studies (e.g

Martin, 1994 and Hu et al., 1999) revealed the soil bearing response of a

foundation is affected by the degree of homogeneity involving dimensional ratio kD/sum Similarly, for footprint problem that is essentially abearing capacity problem, the induced H and M are deemed to be affected bythe non-homogeneity degree of the subsoil particularly for the portion of thespudcan outside the remoulded region For all tests, the non-homogeneitydegree, kD/sum of the soil falls within 0.45 to 5.22 The relationship between

non-Hmax/suD2 and de/D corresponding to Rd/D = 0.5, 1 < 0.002 and 2 < 0.2 isplotted inFig 6.10 It is observed that the Hmax/suD2increases with de/D Therelationship is fitted reasonably well with an exponential function for de/D of

up to 1.5 as follow:

)06.1exp(

Fig 6.11 shows the relationship between Mmax/suD3and de/D corresponding to

Rd/D = 0.5, 1 < 0.002 and 2 < 0.2 The Mmax/suD3 generally increases with

de/D and converges to a constant value of 0.21 from de/D = 1.2 to 1.5 Theresults can be curve-fitted by the following:

193.0)ln(

6.4.2 Depth of occurrence of Hmax and Mmax

The corresponding depth dh and dm(the depth measured from the mudline tothe spudcan L.R.P.), at where the Hmax and the Mmax occurred, respectively,are studied here The normalised depth dh and dmwith spudcan diameter (dh/Dand dm/D) plotted against de/D are shown inFigs 6.12 and6.13, respectively.The dh/D is found to increase almost linearly with de/D following therelationship as follow:

18.091

6.4.3 Load inclination angle, h and normalised load eccentricity,

em/D

Fig 6.14 shows the load inclination angle,h (at where Hmax occurs) is plottedagainst de/D for all sixteen (16) tests There is no clear relationship betweenhand de/D For most tests,h falls within 2° to 4°

Fig 6.15 shows the normalised load eccentricity, em/D (at where Mmaxoccurs) versus de/D For all tests (except tests T4 and T5), the em/D is fairlyscattered distributed within a wide range of 0.02 to 0.06 for de/D of up to 0.9.However, at higher depth ratio de/D of 1.1 to 1.5, the em/D converges to avalue of 0.033 For tests T4 and T5, em/D is exceptionally high being 0.1 and0.125, respectively This is because both tests were conducted in relativelystiff clay with the shear strength at mudline, sum of 38 kPa (for T4) and 55 kPa(for T5), the Mmax occurred within the crater depth that yield a low V valuewith high M resulting in high e/D (= M/VD) ratio

6.5 Effect of offset distance

Till 2005, there are more than 500 rigs operating worldwide Jack-up rigsconsist of multi-leg and each leg supported by a spudcan For modern rigs,three-leg rigs are most common Most rigs are different in configurations such

as spudcan size and leg spacings InSection 6.5.1, the effect of offset distancebetween a spudcan and a footprint of the same spudcan size is investigated.Assuming that only the leg spacing is different, the same spudcan was used tocreate the footprint and to perform the spudcan re-installation at various offsetdistances, Rd In Section 6.5.2, the effect of offset radial distance is

investigated by taking the spudcan and footprint of different diameters intoconsideration Details of the tests are summarised inTable 6.4.

6.5.1 Spudcan of the same diameter

Six (6) tests were conducted at NUS to investigate the effect of offset distance

on the spudcan re-installation A 6 m spudcan was used to form a footprint and

to perform the re-penetration immediately after the footprint was formed, at

Rd/Df = 0, 0.25, 0.5, 0.75, 1.0 and 1.5 The test arrangements are shown inFig.6.16 with the tests denoting as tests OA1 – OA6, respectively The soilsamples used for all tests were prepared in a similar way Fig 6.17 shows theundisturbed su profile of the soil samples used The su for all tests are boundedwithin su = 25 + 5z and su = 30 + 5z (in kPa).The initial penetrations to createthe footprint for all tests were terminated when the preload pressure achieved

460 kPa with less than 1 day operational period (1< 0.01) The spudcan penetration was performed shortly after the footprint was formed with

where V is the measured vertical load and A is the largest area of the spudcan

Figs 6.18a – f show the preload pressure for the initial penetration and the penetration at various offset distances, namely 0, 0.25, 0.5, 0.75, 1 and 1.5

re-times footprint diameter, Df from the footprint centre In general, the displacement response for the re-penetration is softer than the initialpenetration The load displacement response is stiffer for re-penetration furtherfrom the footprint and is almost similar to the initial penetration response forthe re-penetration at 1.5Df Fig 6.19 shows the percentage reduction in Vduring the re-penetration for all six tests The general trend of V reduction ateach offset distance is that the largest percentage reduction occurs near themudline (d/de = 0+) and the percentage decreases with increasing d/de Theexistence of crater (no soil) and heavily remoulded soil (Rsu < 0.5) at nearsurface may attribute to this high percentage reduction in V (see Fig 6.1 forgeneral footprint condition) In term of offset distance, the percentagereduction in V decreases with increase in offset distance and to nil at offsetdistance, Rd = 1.5Df (test OA6) This is because the spudcan re-penetration at1.5Df is not directly overlapped with the remoulded soil zone (Fig 6.1).

load-6.5.1.2 Bearing capacity factor

The average bearing pressure, q (or herein denotes as preload pressure), isexpressed as follow:

uo

N

where Ncois the non-dimensional bearing capacity for undisturbed soil under

an undrained penetration and suois the corresponding undrained shear strength

at the spudcan load reference point (L.R.P.) It is worth noting that Nco is afunction of spudcan cone angle, cone roughness, embedded depth d, and rate

of increase in strength with depth (Houlsby and Martin, 2003) The spudcanused is made from aluminium alloy with a cone angle of 158º For all sixcases, the non-homogeneity degree, kDf/sum is fairly close, ranges from 1 to

1.2 The non-dimensional bearing capacity factor, Nco for undisturbed soilunder an undrained penetration can be defined as:

uo uo

q or As

is For tests OA3 – OA4, the Ncf converges to Nco after d/D =1.0 which isabout 0.2D below the initial penetration depth For test OA6, the Ncfvalue isclose to Nco

6.5.1.3 Induced H and M

Figs 6.22 and 6.23 show the induced H and M during spudcan re-penetration

at various offset distances In general, the H profile increases with penetrationdepth until it reaches a maximum value Amongst all the cases, Rd/Df = 0.75

gives the highest H of 0.72 MN, whereas Rd/Df = 0 (exactly on the footprint)gives the lowest H of 0.11 MN Similarly, the highest M of 2.3 MNm occursfor the re-penetration at 0.75Df, whereas the lowest M of 0.3 MNm occurs at0Df offset Theoretically, the spudcan re-penetration at 0Df offset shouldexperience no horizontal force and moment owing to axial symmetry of soildisturbance during initial penetration However, the initial condition of themodel seabed may not be perfectly level and that may subsequently result in aless than perfect axisymmetric response In short, considerable H and M areinduced on the re-penetrating spudcan at Rd/Df = 0.5 to 1.0 The re-penetration

at exactly the footprint centre exhibits the lowest vertical load displacementresponse and smallest induced H and M compared to the re-penetration atother offset distances

6.5.1.4 Load inclination angle,  and normalised load eccentricity, e/D s

Fig 6.24 shows the magnitudes of load inclination angle for all six tests There-penetration at 0.75Df gives the highest load inclination angle of 5.3º,whereas the re-penetration at 0Df gives the lowest angle of less than 1º Interm of normalised load eccentricity as shown inFig 6.25, almost similar e/Dsprofiles for the re-penetration at Rd/Df = 0.25 and 0.5 are observed, and asimilar trend applies for Rd/Df = 0.75 and 1.0 The e/Ds for the re-penetration

at 0Df and 1.5Df is nearly zero due to the low M The test results aresummarised inTable 6.5

6.5.1.5 Dimensionless forms: H max /s u D 2 and M max /s u D 3

It has been established inSection 6.4.1 that the induced Hmax and Mmax can benormalised in dimensionless forms of Hmax/suDs2and Mmax/suDs3 and they are

both a function of de/Df for a fixed offset distance of 0.5Df In this section, theinduced Hmaxand Mmax in the dimensionless forms (Hmax/suDs2and Mmax/suDs3)

at various offset distances are studied based on three series of centrifuge testresults (a total of sixteen (16) tests) Series 1 consists of 6 centrifuge testsconducted in NUS and the results were presented inSection 6.5.1.3 Series 2and 3 are extracted from Stewart and Finnie (2001) and Cassidy et al (2009),respectively The test details are summarised inTable 6.4 The test results andthe analyses output are presented in Table 6.5 The non-homogeneity degree

kDf/sum of the soil sample used in all three series tests falls within a narrowrange (between 1 and 4.85) Fig 6.26 shows the relationship between

Hmax/suDs2and de/Df for Rd/Df up to 1.5 The data points from each test serieswere curve-fitted with a 2nd order polynomial curve (shown in dotted lines)indicating the approximate trend of the Hmax/suDs2 for three different de/Dfratios As discussed in Section 6.4.1, the Hmax/suDs2 increases with de/Df for

Rd/Df from 0 to 1.5 All three sets of results show a consistent trend of Hmaxover offset distances For de/Df = 0.39 to 0.88, relatively high Hmax/suDs2valuesare obtained for the repenetration at Rd/Df of 0.5 to 1.0 with a peak at 0.75.The Hmax/suDs2decreases to 0.05 when Rd/Dfis approaching zero

Fig 6.27 shows the relationship of between Mmax/suDs3 and de/D for

Rd/D up to 1.5 For Mmax/suDs3, only ten (10) test results are available as the Mresults were not reported in Stewart and Finnie (2001) Owing to limited data,

no clear relationship between Mmax/suDs3and de/Df at various Rd/Df can beobtained However, generally, higher Mmax/suDs3is observed for higher de/Dfratio for Rd/Df = 0.25 to 1.5 Based on the NUS data set, similar to Hmax/suDs2,

relatively high Mmax is observed for re-penetration at 0.5 to 1Df with a peakvalue occurring at 0.75Df.

In spudcan-footprint interaction, the H and M induced on the spudcanare due to the non-uniformity in bearing resistance and/or soil sliding failure.The larger difference in soil resistance between the two halves of the spudcan(the spudcan is cut into half passing through the centre and perpendicular tothe longitudinal axis), the larger the forces induced As discussed in Section6.2, the major radial soil disturbance is found to be confined within Rd/Df =0.75 Outside this region, the soil is less remoulded and beyond 1.5Df, the soilexhibits an almost undisturbed shear strength profile If a spudcan is installedcompletely inside this major soil remoulded zone at offset distance Rd/Df of 0

to 0.25, or completely outside this zone at Rd/Df > 1.25, relatively low H and

M are induced compared to those partially overlapped with this region (0.25 <

Rd/Df< 1.25), as the soil resistance underneath the spudcan is less varied

6.5.1.6 Effect of d e /D f on load inclination angle,  h and normalised load

eccentricity, e/D s

Fig 6.28 shows the magnitudes of h (at where Hmax occurred) at various

Rd/Df for all three de/Df ratios No clear relationship between h and de/Df isobserved This observation is consistent with the earlier finding whichsuggests that h is not dependent on de/Df, as discussed in Section 6.4.3 Interm of offset distance, relatively highh occurs at offset distance between 0.5and 1Df with a peak angle of about 4.5° at 0.75Df The value of h is less asthe offset distance is approaching the footprint centre or 1.5Df At the footprintcentre, an inclination angle of 0.5° is obtained

Fig 6.29 shows the em/Ds (at where Mmax occurs) at various Rd/Df for

de/Df = 0.39 and 0.88, respectively Similar to h, no clear relationshipbetween em/Ds and de/Df is observed For Rd/Df, higher em/Ds is developed forthe re-penetration from 0.5 to 1Df with a highest value at 0.11 occurring at0.75Df

6.5.2 Spudcans of different sizes

Insection 6.5.1, the same spudcan was used to perform the tests, which gives adiameter ratio Df/Ds of 1, where Df is the diameter of the spudcan used tocreate the footprint and Ds is the diameter of the spudcan used to perform there-penetration In this section, six (6) tests were performed to investigate theeffect of spudcan diameter ratio on the interaction For all six tests, a 8 mspudcan was used to create a footprint A 6 m spudcan and a 10 m spudcanwere subsequently used to perform the spudcan re-penetration at differentpositions The soil samples used have similar undisturbed su profiles as thoseused in tests OA1 – OA6 (seeFig 6.17 for the undisturbed su) The test detailsand a summary of test results can be found inTables 6.4 and 6.5,respectively

6.5.2.1 8 m footprint with 6 m spudcan re-penetration

Four (4) test results (namely tests OA7 – OA10) are presented in this section.The Rd/Df for tests OA7 – OA10 is 0.125, 0.39, 0.65 and 0.875, respectively.The test arrangements are shown inFigs 6.30a – d For test OA7, the spudcan

is positioned in such a way that it completely overlaps with the footprint withmaximum eccentricity (edge-touched-edge) From the previous study, thisposition should be the most critical position for the fully overlapping case

Figs 6.31a – d show the preload pressure during 8 m spudcan initial

penetration and 6 m spudcan re-penetration The load displacement responsefor the re-penetration is generally weaker than the initial penetration andfurther away from the footprint centre, the smaller the reduction in the verticalreaction This observation is consistent with the findings reported in

Section 6.5.1.1

Fig 6.32 shows the induced H during re-penetration for tests OA7 –OA10 When the spudcan penetrates at an offset distance from the footprintcentre, the horizontal load increases with penetration depth and reaches a peakvalue before gradually decreasing towards zero The spudcan penetration at anoffset distance, Rd of 0.65Df experiences a significantly H larger than all othercases At this position, the maximum H occurs at d/de of 0.9 with a value of1.19 MN For the spudcan penetration at 0.125Df (test OA10), the induced H

is considerably lower with the highest value of 0.15 MN compared to the otherthree tests

Fig 6.33 shows the induced M profiles during the spudcan penetration at various offset distances Similar to the induced H, the spudcanwhich penetrates at 0.65Df experiences largest M compared to the other threecases For 0.65Df, the induced M has a ‘double-humped’ profile The upperhump occurs slightly below the mudline indicating the effect of the profile ofthe crater, whereas the lower hump occurs near de (which gives a maximumvalue of 3.2 MNm) indicating the effect from the soil property variation Theinduced M for the spudcan penetrates at 0.39Df (test OA8) increases withdepth to a maximum value of 2.55 MNm at d/de = 1.1 For test OA10, themaximum M of 1.54 MNm occurs right after the spudcan is fully in contact

re-with the seabed (d/de = 0+) and the M decreases to a value of about 1.3 MNmfrom d/de = 0.2 to 1.

6.5.2.2 8 m footprint with 10 m spudcan re-penetration

Two (2) test results (tests OA11 – OA12) are presented in this section The

Rd/Df for tests OA11 and OA12 is 0.125 and 0.83, respectively The testarrangements are shown in Figs 6.30e – f Figs 6.34 shows the loadpenetration response during the initial penetration and the re-penetration fortests OA11 and OA12 The vertical soil reaction for the re-penetration in testOA11 is much weaker than test OA12 as the former has a larger areaoverlapped with the footprint The induced H and M during spudcanpenetration at 0.125 and 0.83Df are presented in Figs 6.35 and 6.36 Theinduced H is significantly higher for the spudcan penetration at 0.83Dfcompared to 0.125Df Surprisingly, the spudcan experiences a similarmaximum M of 9.2 MNm for both tests but at different elevations, d/de = 0.17for test OA12 and d/de = 0.8 for test OA11 For test OA11, although the 10 mspudcan fully overlaps with the footprint, 36% of the full spudcan areapenetrates in relatively less remoulded soil As the soil has an intact shearstrength profile which increases with depth and so does the difference in theremoulded soil strength inside and outside the overlapped area Thisdifferential bearing pressure causes high moments to be developed at lowerdepths

6.5.2.3 Effect of D f /D s

To investigate the effect of diameter ratio, Df/Ds, on the spudcan-footprintinteraction, twelve (12) test results (tests conducted in NUS) were compared in

dimensionless forms: Hmax/suDs2 and Mmax/suDs3 For these tests, de/D rangesfrom 0.7 to 1.0.

Tests OA1 – OA6 have Df/Ds ratio of 1, tests OA7 – OA10 have a ratio

of 1.33, and tests OA11 – OA12 have a ratio 0.8 Fig 6.37 shows the plot of

Hmax/suDs2 versus Rd/Df for the three Df/Ds ratios A fitted-curve (shown indotted-line) indicates the comparison relationship of the Hmax/suDs2 and the

Rd/Df for Df/Ds of 1 In general, higher Hmax/suDs2 is obtained for cases withhigher Df/Ds ratio (the bigger footprint than the spudcan) An exceptional high

Hmax/suDs2 ratio of 0.57 occurs at Rd/Df = 0.65 is observed for Df/Ds of 1.33.For Df/Ds of 0.8, both data points fall below the fitted-curve for Df/Ds of 1.However, the probable more critical Hmax/suDs2 for Df/Ds of 0.8 may not becaptured as only two data points are available

Fig 6.38 shows the plot of Mmax/suDs3at various offset distances for

Df/Ds of 0.8 to 1.33 There is no clear trend showing high Df/Dsyields high

Mmax/suDs3 However, similar to normalised Hmax, for Df/Ds of 1.33, thespudcan penetration at offset distance of 0.65Df also gives a high Mmax/suDs3ratio of 0.26 In general, higher Mmax/suDs3 is observed for the spudcanpenetration takes place at offset distances from 0.39 to 1.0Df

6.5.2.4 Practical implications

SNAME (2002a) recommends the use of an identical rig and positioning therig in exactly the same position as the previous unit In this situation, theadverse effect of the footprints on the rig installation can be minimised (see

Figs 6.26 and 6.27 for zero offset distance) In a case where no identical rig isavailable, one should consider the probable detrimental forces induced on therig (that may endanger the structure integrity and stability) for the un-matched

spudcan imprint installations Indeed, spudcan-footprint interaction is not somuch dependent on whether the rigs are identical but more of a case where itcannot be installed at the same spudcan centre If so, the offset distancebetween footprint and spudcan becomes crucial in this problem Hence, acarefully planned rig selection is deemed important to minimise the adverseimpacts on the rig during the installation The established relationship of theoffset distance and the forces and also the studies on the size effect mayprovide some guidelines for a more justifiable rig selection and rigpositioning/heading The critical offset distances should be avoided to reducethe likelihood of spudcan sliding and the risk of structure over-stressing Theprobable induced forces can be estimated from the provided charts The effect

of these forces on the structure stability and integrity can be assessed by mean

of numerical analyses

6.6 Concluding remarks

The manner of a spudcan interact with a footprint is not only dependent on thefootprint configuration, but also the jack-up configuration and the offsetdistance First part of this chapter reported the results of a series of centrifugemodel tests that had been undertaken to investigate the three primaryparameters: leg flexural rigidity, spudcan diameter and leg preload pressure ofthe first penetration The measured loads show no distinguished difference inthe H profile for leg stiffness within a range of 7.82×1010Nm2 to 1.22×1012

Nm2 The restrained connection (between the model leg and the verticalactuator) used in the model set-up may attribute to this result as the effect ofhorizontal displacement on the soil reaction is discounted In term of spudcansize effect (the same diameter spudcan was used for both initial and re-

penetration), higher forces are generated for the larger diameter spudcan Thisclearly demonstrates that the induced forces are a function of spudcandiameter Larger H and M are induced on the spudcan that interacts with thefootprint formed with higher preload pressure as deeper remoulded region iscreated.

Charts showing the relationship between the non-dimensional forms,

Hmax/suDs2 and Mmax/suDs3and de/Dffor the spudcan re-penetration at 0.5Dfaredeveloped It is found that Hmax/suDs2 and Mmax/suDs3 increase with de/Df.However, no significant influence of de/Df on the h and em/Ds is found Interm of offset distance, relatively high H and M are induced on the spudcan forspudcan installation within 0.5 to 1Df offset from the footprint centre Thisfinding suggests that a future rig installation should avoid these offsetdistances in order to reduce the risk of sliding and/or structural over-stressing.When considering spudcan diameter ratio, Df/Ds (for ratio = 0.88, 1.0 and1.33), higher Hmax/suDs2 is observed for cases with higher Df/Ds ratios Theexperimental results have successfully identified some critical parameters such

as spudcan size, preload pressure, offset distance and spudcan diameter ratio,which affect the spudcan-footprint interaction and need to be considered forfuture rig deployment

* z is the depth from the mudline (in meter)

** Classified based on the average undrained shear strength (from the mudline to de) in accordance with BS 8004:1986

*** Tests done in NUS used Malaysia kaolin clay, whereas test done in UWA used UWA kaolin clay Refer to Tables 3.5 and 3.6 for theengineering properties

Adjusted time factor, cvt/RTest No Undisturbed

su*(kPa)

Consistency** D

(m) (m)de de/D

Test done in(***)

EI for Leg 1 = 1.22×1012 Nm2

EI for Leg 2 = 7.82×1010Nm2

Leg length for legs 1 and 2 = 28 m

de– the initial penetration depth (from mudline to spudcan base level)

z – the depth from mudline in meter

Test Name su profile

(kPa)

D(m) Preload Pressure(kPa) (m)de Leg Test description

Leg stiffnessSize effect