ANINTEGRATEDPILEFOUNDATIONREASSESSMENTTO SUPPORTLIFEEXTENSIONANDNEWBUILDACTIVITIESFOR AMATURENORTHSEAOILFIELDPROJECT Argiolas R. Geotechnical Engineer Granherne Limited a KBR Company Jardine R. Imperial College London Abstract This paper considers integrated foundation reassessment and site investigation planning for a North Sea life extension and expansion project. A quantitative review of existing borehole and iterative back-analysis of recorded blow count data aided the planning of new investigations and a more reliable assessment of the new and existing platforms’ foundations. The existing development prises an FPSO and two steel jackets with skirt piles driven into predominantly very dense sands. Three of four piles driven for one jacket refused on driving and when back-analysed in conjunction with the Imperial College Pile (ICP) capacity method the driving data indicated characteristic cone penetration test (CPT) values higher than the maxima that could be recorded with 1990s survey equipment. Subsequent investigations with higher capacity CPT cones confirmed the postulated higher q values and allowed pile design parameter profiles to be updated and applied while also addressing the potential for cyclic degradation under storm loading and pile shaft capacity increase over time. The study demonstrated a good overall degree of redundancy in the jacket foundations' reserve capacities. 1.Introduction The oil field development is located in just over 100 m water depth in the Central North Sea. The WPP B2 70.3m initial development which included a Floating North 包 83101.5m Production Storage and Offloading Facility (FPSO) and Wellhead Protector Platform (WPP) led to first oil in the 1990s. Further developments have taken place since first oil including a Bridge-Linked Platform (BLP); new platform developments are now planned. BLP B6C 76.5m 2.Borehole and platform layout Nine boreholes B1 to B9 were drilled between 1993 回 and 1996 to depths between 19.3 and 101.5 m at the 19.30 locations indicated on Fig. 1 relative to the WPP and BLP jacket legs. Ot 80 m Eight 96" (2.438 m) Outer Diameter (OD) piles were Figure 1: Platform and borehole layout driven in pairs (5 m apart) to 57.5 m at each leg of the WPP jacket with a Menck MHU3000 hammer. Four 3.Geological Ground Model similar piles were driven one at each BLP jacket leg. The site investigations executed from 1993 to 1996 with an IHC S2300. The BLP piles had a target included surface/shallow-subsurface geohysics and penetration of 65.3 m but three of the four were 2D high resolution seismic geophysical surveys over terminated after hard driving at shallower depths of the platform subsea structure and pipeline route 64.8 64.45 and 62.7 m for legs B2 B4 and D4 areas.Boreholes were advanced with downhole respectively. pushed tube sampling and in situ CPT techniques.
The ground model developed from the integrated site although Jardine et al (2015) remend treating investigations presented in Fig. 2 prises: such values cautiously. ●In the same way the short CPT strokes that ●Unit I: a thin layer of Holocene loose sand. resulted from halting the CPT strokes at 'maxed- ●Unit Il: a Forth formation channel infill running out’ qc values may have led to missing some thin southwest to northeast prising very soft clay sub-layers with low q values. approximately 8 m thick at the platform sites ●The onshore laboratory testing involved a only increasing to 4 m at the subsea template areas. limitedhigh quality triaxial testing andno ●Unit IIl part of the Coal Pit formation which interface shear tests. Information on clay index prises soft to very stiff sandy clay. Unit III only appears in the channel infill area. Units IVand V:Older Coal Pit formation 5. Back-analysis of exsiting piles’ driving records sequences of Firm to Hard sandy Clay and 5.1 Background and back-analysis method Medium Dense to Very Dense Sand layers whose An iterative back-analysis was undertaken of the pile thicknesses range from 10 to 20 m and maximum driving records to produce Soil Resistance on Driving depth extends to over 100 m at the platform sites. (SRD) profiles that could be pared with ICP capacity assessments. The process started with initial X' best-estimate’ local resistance estimates that were informed by judgement and lower-bound guidance charts for interface shear angles 8. Updating was then JACKETS made on the basis of 1-D wave equation analyses. It FPSO Template C is important to recognise the potential limitations of applying such tools for back-analysis which include: North Manifold B ●The 1-D wave equation analysis offers only a simplified model of pile driving. 400 300m SRD back-analysis is subject to significant X X(NW) JACKETS Manifold B X(SE) deliver unique inverse analyses of the soil input parameter profiles. ●The pile shaft resistances recorded during driving are likely to fall below those available in static tests.Post-installation pore pressure equalisation Figure 2: Ground model: map and section X'-X and other ageing processes lead to marked shaft set-up in sands and also gains in most clays. 4. Review of existing site investigation data ●Pile tip driving resistances may fall far below those The jacket piles develop most of their axial capacity expected under static loading conditions. in the dense sand layers of Units IV and V. Calculations performed with CPT methods such as Despite the above limitations the back-analysis gave the ICP (Jardine et al 2005) indicate that the axial insights into field behaviour particularly the BLP piles that had refused which proved valuable in capacities are sensitive to local variations in these Units’thicknesses positions and states. guiding new site investigations and enabling a more reliable capacity assessment. SRD was assessed using However the 1993 and 1996 site investigations did not provide all the information required to undertake the Alm and Hamre SRD “friction fatigue’ method an ICP capacity assessment reliably. In particular: (1998) matching the measured blow count data with predictions from GRLWEAP a 1-D wave equation analysis program that simulates pile response to The CPT cones employed in the 1990s did not driving.The approach considers the initial static and offer sufficient capacity to provide full q profiles final residual shaft resistance values linked by a in the densest sands (1996 qc measurements in friction-fatigue’ function and applies soil properties Fig. 6). The operators generally halted the CPT which include undrained shear strength Φ' angle strokes when qc exceeded 50 MPa although some CPT q and f measurements and side and tip measurements were made with qe up to 75 MPa. damping values of 0.25 and 0.50 s/m respectively. Measurements at other North Sea sites have shown Initial ‘forward predictions’ were based on boreholes q can exceed 100 MPa in very dense sands chosen for each pile's static capacity assessment.
Input parameters were modified interactively to SRD in MN improve the match between predicted and recorded blow count records. 2 Pile BlowcountEnd Bls/250 B2-51 1361 Time (hm) The back-analysis results are summarised in Figs. 3 8 6 82-52 00:51 B4 51 9E 00:55 and 4 for the WPP and BLP piles respectively. The 10 84-52 2460 01:20 ET0 WPP traces are evidently more dispersed than the D2-51 D2-52 1000 1283 15193 00:33 BLP equivalents. In addition the second pile from D4-51 2052 00:44 D4-52 2215 01:05 each pair driven at the four WPP legs manifested a 20 o:O higher SRD than the first. This is due to the second pile penetrating into sand layers where the lateral 26 stresses have already been raised by driving the first E30 pile. The second pile acts in turn to impose additonal 32 radial stresses on the first pile's shaft leading to a 34 OR 38 36 positive group action effect (in the sand layers only) as demonstrated experimentally by Chow (1997). MO) 42 40 5.2 WPP results 48 The total blow counts for WPP Leg D2 were 52 approximately half those recorded for D4 and B4 while the Leg B2 counts fell between the these limits. The backanalysed SRD profiles reflect these significant variations between piles. Considering the 57.5 m final penetration six piles terminate in a 66 petent probably very dense sand layer. Five B2-S1 B4-S1 **D2-S2 D2-S1 D4-S1 SRDs fall in the 65-74 MN range while one reaches B2-S2 =B4-S2 == D4-S2 an 84 MN final End of Driving (EoD) resistance. Figure 3: WPP 8-pile back-analysis results The six WPP piles at the B2 B4 and D4 legs show 5.3BLPresults broadly similar SRDs to the BLP piles driven to the As shown in Fig. 4 all piles show SRD increasing same depth. However the two Leg D2 piles show with depth and rising markedly at penetrations greater significantly lower SRDs than the other six over their than 56 m. Three of the four piles indicate EoD SRD full depth with EoD values of 50 to 55 MN. Leg D2 values exceeding 100MN close to the maximum is located between boreholes B3 and B2 of which B2 capacity of the driving system. Only pile D2 could indicates a greater thickness of relatively low penetrate to its 65.3 m target depth without exceeding resistance clay. The driving records indicate that clay the blow count limit; for example Pile B4 reached a persisted to approximately 40 m at Leg D2 whereas maximum of 1504 blows/250 mm over its final drive the other piles encountered sand from apporximately and came close to refusing at penetrations of 56.2 m 28m as indicated by their closest boreholes. and 56.45 m after electrical problems developed with the hammer that each took approximately 5.5 hours to The steady increases in the D2 piles’ SRDs between rectify. Counts rose to 500 blows/250 mm when 42-44 m and much sharper rise at 52-54 m correlate driving remenced indicating marked set-up. The with the presence of a dense sand layer that develops back-analysed SRDs exceeded the initial estimates higher CPT (qc) values. The initial forward-predicted for the BLP piles’ static capacities based on the 1990s SRD profiles generally pared well with the site investigations. back-analysed profiles except for Leg D2 as explained above. The idealised CPT profiles specified The predictions and measurements could only be 65 MPa for qc in the sand layers found generally reconciled by assuming CPT q.
> 65 MPa over the between 26 m and 60 m. The pile D2 predictions were sand sections in which the cone operators had improved by abandoning the initial assumption that terminated their strokes due to reaching the devices' thespatially closest orehole3should beappliedt max-out values and so failed to achieve reliable these piles and assuming instead that the more clay- continuous profiles. The qc profile was adjusted dominated second closest borehole B2 was more upwards from 65 MPa to 80 MPa over the last 5 m of applicable for the SRDandstatic capacity penetration within the sand layer to achieve a good assessments. blow-count match for Pile D2 which developed the
lowest final SRD at the BLP location. More radical been applied in multiple design and assessment increases in q were required to match the three other studies: see Overy (2007) Aldridge et al (2010) or piles’blow counts.However recognising the Merritt et al (2012). potential problems of SRD interpretation at load levels close to refusal the latter were not adopted at The ICP methods rely on good quality CPT data and the interim asessment stage. ring shear interface measurements of pile-surface- soil-friction angles (8). Yield Stress Ratio (YSR) SRD in MN determinations are required with clays as are 506070 8090100110120 reconstituted oedometer tests and/or shear strength 2 tests to assess clay sensitivity (Si). In the initial Pile Blowcount 5730 Pen(m)] End Bls/250 Time (hm) 64.80 03:22 assessments sand interface friction angles in sands 9180 3861 02:15 05:48 were derived from Jardine et al (2005)'s correlations 10 12 1887 62.70 ST:20 with mean particle size dso values. The clay interface 14 friction angle (8) interpretation was based on lower 16 18 bound empirical relationships between residual and 20 22 plasticity index.Sensitivity was estimated from 24 liquidity index while YSR （or apparent over 26 consolidation ratio) was estimated using SHANSEP E 30 28 relationships that were deemed appropriate for the 32 clay types encountered. 36 38 Soil'set-up'from Jardine et al (2015) argue that end bearing 2 stoppages calculationsshouldinvolveaconsciously 46 conservative assessement of design CPT qe values. 50 48 Noting the BLP piles’ consistently hard driving to final penetration an average q profile was assessed 56 for this jacket’s piles over a depth range 1.5 times the 58 pile diameter above and below the pile tip. A more cautious end bearing was adopted for WPP location 66 due to the presence of numerous thin clay layers and the greater scatter shown by the SRD back-analysis. B2B4-D2D4 Taken together the revised shaft and base Figure 4: BLP 4-pile back-analysis results assumptions providedaconservative interim assessment of pile capacity and stiffness that allowed 6.Static capacity predictions engineering to advance and foundation integrity to be 6.1 General assessed in interim re-assessments of jacket- The original pile design undertaken in the 1990s was foundation interaction. The latter adopted updated performed to the API main text method which tends jacket in-place design loads that accounted for the to underpredict axial capacity for piles with L/D<40 higher wave loading indicated by new Metocean driven in dense to very dense sand; Jardine et al measurements. Meanwhile new site investigations (2005). Both conditions apply at the site considered were planned and conducted to check the revised soil where the L/D range was 23.5-26.8 and axial capacity parameters including the sand layers’ qc profiles and was dominated byvery dense sands.The conservatively biased API procedures led to axial so allow an updated and more secure final foundation analysis. capacities of 81.5 MN and 87.0 MN for the WPP and BLP piles at their target penetrations of 57.5 m and 6.2 ICP static capacities 65.3 m respectively. The unplugged ICP static capacities assessed as outlined above are presented in Tables 1 and 2 with CPT-based capacity methods offer fundamental the BLP piles showing significantly higher capacities advantages in sands and have shown statistically that reflect the cautiously raised qc profiles and the closer predictions for pile load test databases; see piles’ hard driving conditions. Yang et al (2017). Four CPT-based methods are presented in API-RP2GEO for sands. The re-analysis adopted the ICP sand and clay methods which have Table 1: WPP ICP capacity predictions using SRD Following field observations by Byme et al (2012) back-analysis results 57.5 m penetration case the shaft SRD has been puted by deducting from the total a third of the static base capacity ponent. Borehole Shaft End ICP Tension The resulting ratios indicate EoD shaft SRDs that fall Cluster Friction Bearing Comp- (MN) (MN) Capacity Capacity 13 to 36% short of the ICP estimates. However the (MN) (MN) latter predict the static capacity available 10 days B1 B2 80.9 13.9 94.8 59.7 after driving. The field tests presented in Fig. 5 show that shaft capacities set-up strongly in sand and that B2 D2 72.9 11.1 84.0 57.4 EoD resistances can be expected to fall 10 to 40% B3 B4 D4 80.0 13.9 93.9 60.9 below the medium term (nominally 10 day) ICP predictions. Table 3 lists the piles in each corner in the order in which they were installed. As noted Table 2: BLP ICP capacity predictions using SRD earlier the second pile driven in each group back-analysis results final penetration cases developed ahigher SRD with an average increment of 4.2MN (around 6%) implied in shaft resistance. Borehole Pile Shaf Bearing End ICP Comp- ICP (Pen) No. Frietion (MN) (MN) Capacity Tension Capacity The equivalent data for the BLP piles are given in (MN) (MN) Table 4.In this case the ‘EoD SRD-to-static shaft B7A B2 100.8 26.0 126.8 69.8 capacity ratios² are above or close to unity (100%) for (64.8) three of the four piles suggesting that their stati B7A B4 99.0 26.0 125.0 69.5 capacity may be 20-30% higher than suggested by the (64.4) ‘consciously’conservative ICP calculations.This B6C D2 106.6 13.0 119.6 68.9 means the intial ICP static capacity has probably been (65.3) underpredicted. The fourth (D2) pile’s SRD was 73% B7A D4 26.0 of the calculated ICP static capacity which is in line 98.0 124.0 69.1 (62.7) with the WPP results and Fig. 5 field load test data. This is an automatic consequence of adopting the lowest SRD pile case when re-matching the design q 6.3 SRD versus ICP capacity profile for the BLP static capacty estimates. The ICP ‘shaft-only’ capacity values are pared Table 4: BLP SRD-to-static shaft capacity ratios with those derived from the SRD back-analysis in Table 3. The last column presents the ratio of the EoD Piles EoD ICP. Shaft SRD Ratio = SRD Comp shaft SRD as derived by back-analysis of the first (MN) Shaft : =SRD. Shaft pile driven for each WPP leg to the ICP static shaft ICP-Base/3 Base SRD/ Shaft capacity. (MN) ICP Table 3: WPP SRD-to-static shaft capacity ratios B2 114.0 100.8 : 26.0 105.3 01 Piles EoD SRD'. ICP. Shaft Ratio = SRD SRD' Comp SRD B4 114.9 99.0 : Shaft 26.0 106.2 1.07 (MN) (MN) A Shaft : Base SRD/ Shaft D2 81.7 106.6: (MN) ICP 13.0 VLL 0.73 B2-S2 66.4 80.9 : D4 101.5 98.0: 0.2 61.8 0.76 26.0 91.5 0.95 B2-S1 66.6 13.9 B4-S1 74.4 9.4 80.0: B4-S2 88 13.9 869 0.87 6.4 Long term capacity D2-S1 50.7 72.9 : Rimoy et al (2015) discuss how the ageing trends D2-S2 55.0 4.3 11.1 47.0 0.64 shown in Fig. 5 continue long after full pore pressure D4-S1 65.1 80.0: equalization through processes that include: a) radial D4-S2 68.0 2.9 13.9 60.5 0.76 effective stresses increasing steadily due to creep Note A: SRD-SRD represents the difference in SRD between processes relaxing circumferential arching around the the second and first pile driven in pairs at each jacket leg. shaft b) increased shaft dilatancy developing on loading and c) physiochemical and biological activity that may impede shearing at the interface in sand and disrupt residual shear surfaces in clay.
