BPCLAIRRIDGE:INDEPENDENTFOUNDATIONASSURANCE FORTHECAPACITYOFDRIVENPILESINVERYHARDSOILS K Hampson and TG Evans* BP Exploration Operating Company Sunbury UK (*formerly) RJ Jardine Imperial College London UK P Moran Lloyd’s Register Energy Bath UK B Mackenzie and MJ Rattley Fugro GB Marine Limited Wallingford UK Abstract The second phase of development in BP's Clair Field Clair Ridge prises two bridge-linked piled steel jacket structures: one for Drilling and Production and a second for Quarters and Utilities. The jackets were installed successfully in 2013 in 140 m of water about 6.5 km north-north-east of Clair Phase 1 platform. The soil conditions at Clair Ridge are similarly extremebut present a higher degree of variability both in layering and installing driven piles. This paper describes the approach taken by BP's foundation assurance team to provide prehensive validation of long-term axial pile capacity. The effects of pile slotting cyclic loading and group action were considered referring to the soilsmechanical properties revealed by prehensive geotechnical site investigations. Validation method calibration by driving data back-analysis is also discussed. 1.Introduction 1.1Background The Clair Field is located about 75km west of the Shetland Islands UK. The first development (Clair Phase 1) was sanctioned by BP and its co-venturers Chevron ConocoPhillips and Shell in 2001 and production began in 2005 via a single fixed piled steel platform and associated oil and gas export fa- cilities. Evans et al. (2011) present various aspects of the development planning. A second phase of production Clair Ridge prised two bridge- linked piled steel structures: a Drilling and Produc- tion (DP) platform and a Quarters and Utilities (QU) platform and was sanctioned in 201l. A drilling Figure 1: The Clair Ridge DP and QU plarforms template and two instrumented 72-inch diameter docking piles were pre-installed at the DP platform location in mid-2011 and these installations were The piled foundations for the Clair Ridge DP plat- adopted as driving trials for the main platform piles. form prised five 2.74 m (108 inch) diameter 100 mm uniform wall thickness piles per corner leg. The platform jackets were installed in 2013 in a wa- driven to penetrations from 25 m to 40.5 m. The QU ter depth of approximately 140 m. The topsides were added in mid-2015 (QU) and mid-2016 (DP) and are platform is founded on three 2.59 m (102 inch) di- shown in Figure 1. First oil production is expected ameter 100 mm uniform wall thickness piles per corner driven to between 25 m and 27 m. late in 2017. The detailed design and fabrication of the steel jackets and foundations were performed by Two piles at each of the jacket legs were instrument- Kvaerner Jacket Technology (KJT) and Kvaerner ed with strain gauges and accelerometers which Verdal (KV) respectively. The foundation design were monitored continuously during driving. Re- was supported by the Norwegian Geotechnical Insti- drive tests were performed at target penetrations on tute (NGI). The jackets and topsides were installed two instrumented piles from each jacket. by Heerema Marine Contractors (HMC).
Clair Ridge Dinection of ioe move Inoreased sand and coerse during ice advanee granuler material Hneu buunp Direction of ioe Grounded ice sheet ant by advanging ice lodgement ti ution g Sea level by advan jBwst moees so: fugeae rombaseofadvang loe shieet.. seting fro foating ie Suspension Calve cebel Cottles and boulders Quatemary sediments sent through Progack deposits from calved iceberg Sequ Figure 2: Conceptual summary of glacial processes on the West of Shetland Shelf 1.2 Site Conditions platform locations still posed challenges for the de- The Clair Field is underlain by soils that were depos- sign and installation of driven piled foundations: ited pressed and repeatedly sheared by hun- dreds of metres of ice during successive glaciations. ●The ISO 19902:2007 methods for estimating Waxing and waning of glacial ice sheets has largely static pile capacity are based empirically on the controlled the spatial and temporal distribution of API pile load test database reported by Olson particular sediment packages. The depositional envi- (1984) and assume that the embedded length to ronments and the associated stress history have had diameter (L/D) ratios of offshore piles are suffi- a major effect on the soil properties. Of particular in- ciently high (typically >15) for axial capacities terest are the glacial deposits of the Otter Bank se- to be unaffected by lateral and moment loads; quence which occurs extensively on the outer part The Clair tills are much stronger than those in- of the West of Shetland Shelf. These soils prise cluded in the API pile test database. In addition interbedded hard clays and very dense sands with the L/D ratios of the DP and QU jacket piles (9 gravel to boulder-size igneous rock inclusions.Fig- to 14) fall below the range represented in the da- ure 2 shows a conceptual summary of some key gla- tabase and arguably below the range at which cial processes on the West of Shetland Shelf. axial and lateral response interactions may be negelected. In effect there were no validated Shallow geophysical and geotechnical surveys were methods for estimating the pile’ axial capacities; performed at Clair Ridge in 2009. The soils were ●The Clair Ridge deposits were too laterally vari- found to be similar to those at the Clair Phase I plat- able (see Figure 2) to support a design approach form site where extremely hard and dense tills were based on average soil properties over the struc- encountered (Aldridge et al. 201l; Jardine et al. ture’s footprint. A leg specific geotechnical data 2011); however greater levels of lateral variability based approach was required; and thicker sand units were noted. While Aldridge et al. (2011) report easier than predicted continuous pile driving at Clair Phase 1.3 Design Challenges 1 significant set-up was also observed during Despite invaluable experience from Clair Phase 1 driving pauses. There was potential for hard the soils cobbles and boulders at the Clair Ridge driving and possible pile refusal in the event of
The ISO procedures used to estimate the base near to the target depths; and case axial pile capacities would be conservative ●Cobbles and boulder-size inclusions posed po- for Clair Ridge because it does not fully capture tential risks to pile installation that might result in (1) hard driving (2) excessive pile damage for clays with high yield stresses observed in and/or (3) pile refusal. open-ended tubular steel pile test databases; ●Methods that are based on sounder soil mechan- The Clair Ridge Project design team applied a *best- endeavours’ approach to the assessment of axial pile ics principles and/or are a better statistical fit to capacities similar to that adopted during Clair Phase the subset of pile test data that most closely 1 (Evans et al. 2011). The process involved using the match the Clair Ridge pile geometries and-soil industry standard ISO procedures for developing the conditions would give more reliable and less base case designs and validating those solutions us- conservative characteristic pile capacities than ing other robust physically reasonable approaches. the ISO method; An independent foundation assurance team (IFAT) ● Characteristic pile capacities may be reduced by post-peak softening of unit skin friction along was appointed by the Project to validate the DP and QU jacket piled foundation solutions.The IFAT the pile shafts and should therefore be consid- team prised specialists from BP Imperial Col- ered in design; and lege Fugro and Senergy. This paper describes the ●Project-life operational factors such as local and technical approach taken by the team with a particu- general scour gapping/slotting under cyclic lat- lar focus on the prehensive validation of long- eral loading pile group interaction and axial cy- term axial pile capacity. clic loading may reduce the basic characteristic capacities of the Clair Ridge piles. While these The team was also asked to carry out independent operational factors should be addressed it may predictions of pile driveability to develop pile instal- be overly conservative to apply them all to lation acceptance plans and to monitor the pile in- characteristic pile capacities derived by the po- stallations. However detailed descriptions of these tentially overconservative ISO method. activities are outside the scope of this paper. 2.3Validation Process Figure 3 illustrates the validation process adopted by 2. IFAT Design Assurance the IFAT team. The characteristic axial pressive 2.1 Strategy and tensile pile capacities represented the expected Given the design challenges associated with the Clair Ridge platform foundations a key objective of capacities that a single isolated pile would develop if loaded quasi-statically with no other factors taken in the IFAT work was to confirm that the base case de- to account. Operational factors were those as ex- sign had sufficient theoretical reserve capacities to pected to arise during the 40-year life of the plat- acmodate minor installation problems without forms and which may reduce pile characteristic ca- necessarily requiring remediation. The design ro- pacities: (1) general scour of near-surface sands (2) bustness criterion set by the Clair Ridge Project was gapping/post-holing and other damage in the hard that the tolerable utilisation factors (UFs) for maxi- clays due to lateral cyclic loading (3) pile group in- mum loaded single piles and maximum loaded pile teraction and (4) axial cyclic loading. Operational static axial pressive and tensile capacities were pression. then obtained by multiplying the corresponding 2.2 Basis for Approach characteristic pile capacities by the operational fac- tors in the exact sequence shown on Figure 3. The approach taken by IFAT was based on the fol- lowing convictions concerning the Clair Ridge soils 2.4 Validation Methods and the design of open-ended tubular driven steel Two different validation methods were adopted to piles in such soils: estimate the long-term characteristic axial capacities The hard clay tills encountered at the platform of the DP and QU platform piles. A third level of locations are similar to those encountered at the validation was obtained by calibrating the two pre- Clair Phase 1 platform; dictive methods with site-specific pile driving data. : The soils are not cemented and their high Validation Method 1 (VM1) adopted the ICP ef- strengths are due to mechanical processes only; fective stress method for driven piles in sands ●The soils are sufficiently variable to require leg- and clays (Jardine et al. 2005) which has been specific pile designs;
calibrated for range of soils including those Extensive series of offshore and onshore laboratory with glacial origins; tests were performed on samples recovered from the Validation Method 2 (VM2) was based on the investigation. These series included advanced stress ISO method but considered the statistical bias path triaxial cyclic simple shear and interface shear of the method for open-ended low aspect ratio testing and were similar to that reported by Aldridge piles in heavily overconsolidated clays; and et al. (2011) and Jardine et al. (2011). All parameters ●Validation Method 3 (VM3) was a calibration of required for input to the ICP method and cyclic the VM1 and VM2 predictive procedures by analysis were carefully measured by prehensive parison of the long-term characteristic static laboratory testing. pile capacities estimated for pre- installed 72 inch template docking piles with those extrapo- 3.2 Integrated Ground Model lated from driving records using signal- IFAT's work was informed by a shallow geological matching techniques. and geotechnical engineering ground model for Clair Ridge that was developed following a detailed as- sessment of the available geophysical geochrono- Derive Geotechnical logical geological and geotechnical data for the Ground Model Parmeters Clair Field. 3.3 Design Soil Profiles and Parameters The Clair Ridge soils were found to be highly later- ally variable. Soil profiling and parameter selection Validation Method Selection Charcterisic Pile Respons were therefore performed on leg-by-leg basis. Gen- Single Ple Capacity Aralysis erally this process was straightforward and based on the available data. In some cases a holistic approach based on seismostratigraphic unitisation from the engineering ground model was required. This was Operational Pile Operational particularly useful in cases where only short CPTu Capocity Reduction Factors strokes could be achieved in the hard soil conditions. Figure 4 summarises some key geotechnical parame- ters for one of the leg locations investigated. Operational Capacity = Chracteristic Capacity x Operational Reduction Factoes Figure 3:Axil pile capacity validation process Location-specific cone factors were derived for clay layers to account for regional depositional and post- depositional effects. Notably high Nke values be- 3. Geotechnical Parameters tween 18 and 33 were required to find Su values 3.1Site Investigation Combined piezocone (CPTu) and deep boreholes from net cone resistance. The soil:steel interface friction angles measured according to ICP method were performed at each corner of the DP and QU procedures were also found to be relatively high platforms along with one further deep borehole at the drilling template location giving nine deep with ute ranging from 26° to 28° in the clay units and a limiting value of 28.8° being assigned to the boreholes in total. High quality samples were recov- ered from each sampling borehole. Static pile bear- dense sand units. ing capacity analyses using ISO 19902:2007 were performed during the course of the investigation to 3.4 Stress History and Yield Stress Ratio ensure that each borehole depth exceeded the rec- The stress history of the soils at Clair Ridge required careful attention. Whilst the exact nature and extent ommended pile penetration by ~3 diameters plus 6 of previous glacial events was uncertain it is clear m or to a minimum of 45 m below seafloor (BSF). these formations have experienced significant load- Based on this approach a single borehole was ter- ing and unloading phases shear deformation and minated at 60 m BSF with investigation at all other possible weathering. To assess the effect of these platform corners being terminated at depths from 45 processes the relationship between yield stress ratio m to 50 m BSF. Boreholes were performed using (YSR) vertical effective stress and undrained shear Eight further shallow boreholes were performed strength was investigated within a rational frame- work as outlined by Jardine et al. (2005). (one per leg location) to facilitate mudmat design in the variable soil profiles.
Existing oedometer based methods for prediction of estimate characteristic pile capacities using both yield within dense glacial tills often produce poor VM1 (effective stress method) and VM2 (modified quality correlation or a poorly defined yield point. total stress method). The approach also allowed a more realistic assessment of soil strength and YSR was introduced. The holistic approach adopted for profile with reference to previous geological loading prediction ofverticaleffective yield stresses oy and stress history of the clay layers. The Q. values was calibrated directly against laboratory strength used for sand layers were interpreted from CPTu da- data (Figure 4a) for each specific clay layer. The ap- ta alone. This was also carried out within a holistic parent degrees of preload were considered across framework for each seismostratigraphic unit. The each jacket’s footprint. An iterative method was then application (where possible) of a standardised set of used to calibrate a final YSR= oy/o profile parameters across VM1 and VM2 eliminated subjec- against depth from CPT and laboratory strength data tivity in the dataset and allowed a more rational ap- at each leg location (Figure 4b). proach to method parisons. Ceae Resstance g [MPa] Uadralned Shear Strength sg [BPa] 4. Characteristic Axial Pile Capacities 001608009060010 N001000 15002000 2500 4.1Validation Method 1 Strew Hatory. 4 VM1 adopted the ICP method for driven piles to provide leg-specific estimates of characteristic axial pressive and tensile capacities based on the ge- otechnical parameter selection process described above. E All ICP pile calculations indicated that the DP and QU platform piles would behave in an unplugged manner under static loading at final pile penetration. UUY 088 In view of the erratic qe data and soil layer variabil- 35 ity the q profiles used for estimating unit pile base resistance were chosen more conservatively than 08 those adopted for calculating pile shaft capacity. This approach is remended by Jardine et al. (2005; 2011) and discussed in detail by Jardine et al. Prior Preload pPa] Yield Stress Ratis [-] (2015) for similar circumstances and leads to a low- 10 102030405060 er risk of piles terminating in a weaker-than- expected layer. Although not significant for the gla- cial clay units the effect of strain softening was con- sidered implicit in VM1 since & values were as- 10 - signed based on interface ring shear testing to residual states. 4.2Validation Method 2 The primary objective of VM2 was to establish whether any conservatism is inherent in the ISO method when considering the subset of the API pile test data which most closely match the Clair Ridge 33 conditions. This analysis was employed to establish statistical bias factors for direct input into pile shaft 49 8 00) - capacity calculations. 49 69 - Figure 4: Variation of (α) measured q and Sy and (b) iterated The pile test database subset considered 21 high prior preload and YSR for one DP plarform leg location quality tests on open-ended piles in overconsolidated clays from the test data collated for API (Olson 1984) and by Imperial College (Chow 1997). The 3.5ValidationMethodParameters approach takes advantage of positive statistical bias- The unified approach to parameter selection allowed a consistent set of design parameters to be applied to L/D* ratios where D* is the equivalent diameter of
独立基础保证驱动齿轮在非常硬的土壤中的能力.pdf
