Spudcan extraction from deep embedment in soft clayApplied Ocean Research
Spudcan extraction from deep embedment in soft clay
Manuscript submitted to Applied Ocean Research
Omid Kohan (corresponding author)
PhD Candidate
Centre for Offshore Foundation Systems and ARC Centre of Excellence for Geotechnical Science and Engineering
University of Western Australia
Perth, WA 6009
Australia
Christophe Gaudin
Professorial Fellow
Centre for Offshore Foundation Systems and ARCCentre of Excellence for Geotechnical Science and Engineering
University of Western Australia
Perth, WA 6009
Australia
Mark J. Cassidy
Winthrop Professor
Centre for Offshore Foundation Systemsand ARC Centre of Excellence for Geotechnical Science and Engineering
University of Western Australia
Perth, WA 6009
Australia
Britta Bienen
Associate Professor
Centre for Offshore Foundation Systems and ARC Centre of Excellence for Geotechnical Science and Engineering
University of Western Australia
Perth, WA 6009
Australia
Words: 4632 (excluding abstract and references)
Figures: 16
Tables: 4
Abstract
After drilling is completed, spudcanfootings of mobile jack-up rigsare extracted from the seabed before the jack-up is manoeuvred to a new location. In some instances, the extraction may prove to be difficult and time consuming,especially when the spudcans are deeply embedded, because the pull-out capacity of the rig is less than the extraction resistance ofthe spudcans. In soft soil, the extraction resistance may be significantly augmented by the development of suction at the spudcan invert.To investigate this phenomenon, a deeply embedded 30 mm diameter model spudcan was extracted in a series of physical model experiments conducted at an acceleration of 200g in a geotechnical beam centrifuge. The spudcan,instrumented with two pore pressure transducers,one at the top and one at the bottom face, was extracted fromnormally consolidated clay and under undrained conditions.Eight tests are reported exhibitingembedmentsranging from 1.5 to3 spudcan diameters and varying operation periods. The excess pore pressure and maximum breakout force measuredreveal insights into the magnitude of the suction forces at the spudcan invert, which were observed to increase with the embedment depth. No change in failure mechanism was observed between 1.5 and 3 spudcan diameters depth.
Key words: Spudcan; centrifuge modelling; soft clay; suction; extraction.
Introduction
Self-elevated mobile jack-up units (Figure 1)play an important role in offshore drilling in shallow waters, up to approximately 150m depth. The inverted conical footings of jack-ups,which are known as spudcans and can be in excess of 20 m in diameter in a modern jack-up (Cassidy et al., 2009), can be penetrated in a wide range of soil conditions. In softer soils, spudcans require largepenetration before meeting sufficient bearing capacity to withstand the jack-up’s self-weight and the expected operational loads.Penetrationof up to two or threespudcan diameters may be necessary before reaching equilibrium during the preloading process (Endleyet al., 1981; Menzies and Roper, 2008).
When a jack-up rig is removed from a site and redeployed, its spudcans must be extracted from the seabed. To overcome the soil resistance, the hull is floated, and lowered beyond neutral draft. However, tolerances on the maximum allowable overdraft within the marine operations manual restrict the maximum extraction pull to between 30 and 50% of the maximum compressive loadthat can be applied during installation (Purwana et al. 2009). In softsoils fordeep spudcan penetration (> 1.5 times the spudcan diameter) and long operation periods,the buoyancy of the hull may not be sufficient to extract the spudcan. It is reported that spudcan extraction from penetration depths of one or two spudcan diameters can require one or two weeks, and in some extreme cases, up to ten weeks (InSafe JIP, 2008). The spudcan extraction process, especially from soft clay,may therefore bea time-consumingprocess. Withaverage jack up day rate in the range US$60,000 to US$160,000(depending on the water depth),this has significant financial impact.
Figure 2 shows failure mechanisms during initiation of undrained spudcan extraction, as developed by Gaudin et al. (2011) from observations from Particle Image Velocimetry (PIV)analysis of physical tests by Purwanaet al. (2006) and numerical analysis by Zhou et al. (2009). In the first stage of the undrained extraction of the spudcan, the main soil resistance is comprised of the weight of the soil above the spudcan, the resistance along a shear plane generated above the spudcan and negative excess pore pressure,namelysuction,that is developed at the spudcan base in undrained extraction. In fact, the extraction mechanism is a combination of an uplift mechanism of the soil at the top of the spudcan and reverse end bearing at the spudcan invert due to suction.The contribution of both theses mechanisms is influenced by the duration of the jack-up operation. At the top of the spudcan,Purwana et al. (2009) measured via T-bar tests a reduction of 67% of the shear strength immediately after spudcan installation, followed by an increase of 30% (raising the shear strength to 87% of the undisturbed undrained shear strength) after an operation period of 400 days.Similarly, the gain in shear strength underneath the spudcan after the same operation period time was evaluated as 1.70 times the undisturbed strength by Purwana et al. (2009) from numerical analysis.Both outcomes imply an increase in effective stresseswithinthe soil underneathand at the top of the spudcan, resulting from dissipation of the excess pore pressuresgenerated duringthe penetration process, albeit at a different rate and magnitude. The phenomena governing the changes in effective stresses in the soil are complex and for the soil at the top, potentially include changes in total stresses due to arching.
In a second stage, the extraction resistance typically reaches a peak followed by a dramatic reduction in resistance. The failure mechanism is then replaced by a localised flow around mechanism, still associated withan uplift mechanism of the soil above the spudcan (Gaudin et al., 2011).
From the observed failure mechanism described, it may be inferred that suction forces contribute significantly to thepeak undrained extraction resistance. The importance of base suction generated during spudcan extraction was first revealed by a series of centrifuge tests performed at an acceleration of 100g and simulating the installation and extraction of spudcans from uniform soft clay with an undrained shear strength in the range of 12-40 kPa (Craig and Chua,1990).Results indicated that the magnitudeof suction wasrelated to the compressive loading history and the associated embedment ratio prior to extraction. However, issues such as the operational period that the jack-up is installed for were not studied by and these form an important component of the testing programme discussed in this paper.
Purwanaet al. (2005) experimentally investigated the effect of operation period and operating load magnitude level on spudcan extraction. Results demonstrated thatthe extraction resistance increases with the operation period. In contrast, the level of jack-up operating load(i.e. the load maintained during the operation period) has an insignificant effect on spudcan extraction in comparison with the time that a jack-up is under operation.It is noteworthy thatPurwanaet al.(2005) investigatedspudcan extraction from embedment up to 1.5 spudcan diameters.To the authors’ knowledge, the deepest spudcan penetration reported is 78 meters in the Gulf of Mexico, corresponding to an embedment ratio of 5.6 (Menzies and Lopez, 2011), although this is exceptional and penetrations up to a maximum of three spudcan diameters are more common (Menzies and Roper, 2008).
The objective of the present study is to extend the database of Purwanaet al. (2005) to embedment up to 3 times the spudcan diameter, to notably investigate if a change of mechanism at deeper embedment may affect the suction generation at the spudcan invert. For this purpose, aseries of centrifuge tests were performed,featuring penetration and extraction after varying operating period of a model spudcan penetrated at embedment ratio between 1.5 and 3.
Vertical loads and pore pressures at the top and bottom of the spudcan during the installation, operation period and extraction of the spudcan were monitored, and the results are reportedand discussed.
Soil preparation and characterisation
Commercial Kaolin clay with characteristicsprovided in Table 1was used to create a softsoil sample in the beam centrifugeat the University of Western Australia (Stewart and Randolph, 1991; Gaudin et al., 2011). The mixture of Kaolin and water at a moisture level of twice the liquid limit formeda de-aired clay slurry, which was then poured into a rectangular strongbox over a 15 mm thick drainage sand layer.
Thereafter, the sample was consolidated under self-weight in the centrifuge at an acceleration of 200g for a period of approximately five days.Over the consolidation time, settlement of the sample was measured, and at the end, the final height of the soil specimenwas approximately 180mm.
A 15mm diameter miniature piezoball penetrometer (as also used by Mahmoodzadeh et al., 2011)was used to derive the undrained shear strength profile of the sample with a bearing capacity factor of 10.5 (Low et al., 2011; Leeet al., 2012; Leeet al., 2013).The test was performedin flight at a rate of 1mm/s to ensure that undrained conditions were measured (Chung et al., 2006).Theaverage shear strength gradientwas approximately 1.1kPa/m (Figure 3).
Experimental programme and procedure
Amodelspudcan with diameter D of 30 mm was fabricatedto investigate the extraction of deeply embedded spudcans(Figure 4).The spudcan was manufactured fromaluminium alloy 6061-T6 and was connected to a two-dimensional actuator via a load cell.The model spudcan was instrumented withtwo pore pressure transducers (one at the top face and one at the base) that were installed at approximately half the distance between the centre andthe edge of the spudcan. The cross-section of the pore pressure transducers at the top and base of the spudcan is illustrated in Figure 5.
Eight tests wereperformedat an acceleration of 200g in a beam geotechnical centrifuge (Randolph et al., 1991).Tests one to fourwere designed to investigatethe effect of the embedment depth on spudcan extraction, whereas testsfiveto eightwere performed to investigate the effect of the duration of operation time on spudcan extraction.In the first fourtests, the spudcan installation depth was varied from 1.5to 3times the spudcan diameter.In these tests, spudcan extraction occurred after two years operating load (in prototype scale).In the remainder of the tests,the operation period varied from immediate extraction to three years, and the spudcan embedment ratio was 1.5D.Details of the test programme are provided inTable 2.
Spudcan penetration and extraction was undertaken at a penetration rate v of 0.3 mm/s, resulting in a normalised velocity V=vD/cvgreater than 30 (assuming a coefficient of consolidation cvof 3.99 m2/y, at a stress level consistent with the spudcan embedment, see Table 2). This ensured that spudcan installation and extraction occurred under undrained conditions(Finnie and Randolph, 1994), mimicking in–situ conditions. In the field, successful spudcan extraction may require between 6 hours and 30 hours. Considering spudcan diameters in the range 10 to 20 m and coefficient of consolidation in the range 0.1 to 100 m2/year, normalised extraction velocity in–situ are typically greater than 30.
The same test procedure was used for all cases and consisted of three stages.In the first stage, spudcan penetration was performed in-flight in displacement-control mode. The embedment depth rangedfromapproximately 8.8m to 18.1m(prototype scale) corresponding to an embedment ratio of 1.5 to 3, respectively.In the second stage, the jack-up operation period was simulated by holding a constant vertical load of approximately 85% of the maximum installation loadfor upto three yearsin prototype scale.For operating period of 2 years and above, pore pressure measurements at the spudcan invert indicated that at least 85% of consolidation was achieved. Finally, in the third stage, spudcan extraction was performed at a constant rate of 0.3mm/s.
For all stages, the vertical force on the spudcan (corresponding to the penetration resistance, the applied load, and the extraction resistance for the three stages of testing, respectively)and pore pressures at the top and the invert of the spudcan were monitored.
Experimental Results
Installation resistance
The development of penetration resistance Qp, excess pore pressure(with respect to the hydrostatic pressure) at the spudcan invert ui and at the spudcan top ut, arepresentedin Figure 6,Figure 7,and Figure 8, respectively, for the installation, operation and extraction stages.
Figure 9presentsthe normalised net vertical load Qp/(A.su) where Qp is the net penetration resistance measured by the load cell, A the projected area of the spudcan and su the undisturbed shear strength at the spudcan embedment, againstthe normalised embedment H/D, where H is the penetration depth and D the spudcan diameter.Note that the spudcan embedment is defined at the lowest full diameter of the shoulder of the spudcan. This provides insight into the net bearing capacity factors during penetration.
During installation, excess pore pressures, both at the top and the invert of the spudcan,increase linearly with depth.Tests performed by Purwana et al. (2005) on a larger spudcan, instrumented with both total and pore pressure transducer at the top and invert of the spudcan, demonstrated that excess pore pressures where equal to the change in total pressures during penetration, indicating no change in effective stresses and so a fully undrained process. Based on the same assumption, the penetrating pressure,comprising of the applied pressure qp = Qp/A and the excess pore pressure at the top of the spudcanut,ins,is compared to the resisting pressure ui,ins corresponding to the excess pore pressure at the spudcan invert in Figure 10. Values at the end of the installation phase presented in Table 3. The agreement is reasonably good throughout the full penetration process, confirming the observations fromPurwana et al. (2005), and demonstrating the undrained response of the soil.
This result is however surprising. The phenomena governing the changes in pore pressures at the invert and at the top of the spudcan are complex and involve changes in both effective and total stresses. At the spudcan invert, the soil is essentially sheared so an element of soil underneath the spudcan is expected to experience areduction in effective stresses, reflecting the remoulding of the soil, as well as an increase in pore pressures. The magnitude of the reduction in effective stresses is difficult to assess and is likely to vary along the spudcan. At the top of the spudcan, the phenomenon is even more complex. Pore pressures at the top of the spudcan are likely induced from the shearing of the soil (which is flowing from underneath the spudcan), but also from a cavity expansion mechanism associated with the cylindrical leg of the spudcan, and a reduction in total stresses due to arching and potential silo effect along the column of soil on the top of the spudcan. Similarly to the invert of the spudcan, changes in effective stresses are expected, although they were not observed by Purwana et al. (2005), and are not suggested by Figure 10.
Indeed, accurate assessment of the contribution of the various components to the penetration resistance is difficult as both the top and invert pore pressure measurements are local measurements extrapolated over the entire surface. Purwana et al. (2005), using a larger model with several pore pressure transducers, showed that the excess pore pressures at the spudcan invert increased towards the centre of the spudcan. In addition, the pore pressures were measured at the soil spudcaninterface(rather then in the soil body) and do not necessarily reflect changes within the soil underneath and at the top of the spudcan.
While spudcan penetration is a complex problem, it is noteworthy that it can be elegantly captured by only two parameters, a bearing factor Nc and the undrained shear strength su, as demonstrated in Figure 9.Immediate back-flow on the top of the spudcan was observed visually during testing. This confirms the analysis made by Hossain et al. (2005), indicating that deep failure mechanism, characterised by symmetrical flow-around, occurs at a relatively shallow embedment for soft soils. Indeed, the normalised net vertical load development in Figure 9 exhibits a constant value from an embedment ratio of about 0.7. Bearing factors calculated from the experimental measurements are compared in Figure 9 with large deformation finite element (LDFE) analysis in ideal Tresca soil and Tresca soil modified to account for strain softening and strain rate effects (Hossain and Randolph, 2009). The centrifuge results lean towards the modified numerical solution, i.e. yielding a bearing factor in the range 9-10.4, indicating that undrained conditions are prevalent within the soil and that significant strain softening takes place.
Operating period
Following penetration, 85% of the maximum penetration load (except for Test 2.0D2.0Y in which the holding load was 100% of the installation load due to a temporary technical problem in the centrifuge)was maintained on the spudcan for operating times ranging from 0 to 3 years prototype(see Table 3). This stage resulted in the consolidation of the soil underneath (and to a reduced degree at the top of) the spudcanand additional spudcan settlement as summarised inTable 3.During the operating period, excess pore pressure at the top and bottom of the spudcandissipated, as shown in Figure 11, which presents the development of the degree of consolidation with the time factorTv = tcv/D2,where t is the time since the beginning of the operationalperiod and cvhas been assumed to be the virgin coefficient of consolidation (estimated as a function of the stress level, see Table 2).
It is noteworthy thatdegrees of consolidation ranging from 85% to90% wereachievedat the spudcan base at the end of the operation period for all tests, whereas at the top of the spudcan, the degree of consolidation of about 40% to 60%was inferred(Figure 11).The lower degree of consolidation at the top is best explained by a reduction of the coefficient of consolidation by potentially one order of magnitude. Such a large reduction may be explained partially by the lower stress level experienced by the highly remoulded soil at the top of the spudcan, but also by a significantly higher modulus of compressibility. It is however important to recognise, as for the installation process, that the pore pressure measurements are undertaken at one single point and do not necessarily reflect the behaviour of the entire mass of soil at the bottom and at the top of the spudcan.
Spudcan extraction - Increasing embedment depth and constant operating period – Tests 1 to 4
As previously reported by Purwanaet al.(2005), Bienen et al. (2009), and Kohan et al. (2013), negative excess pore pressures(or suction) generated during extraction reach a peak at the point of maximum extraction resistance, also termed breakout point.In the present case, maximum suction was measured slightly after the breakout point, after displacements ranging from 0.02D to 0.06D.There is no explanation for this behaviour, except potential delay in the pore pressure measurements resulting from poor saturation of the transducer porous stone.Accordingly, the analysis assumes that both peak suction and peak extraction resistance occur simultaneously.