Frontiers in Offshore Geotechnics M Meyer (Ed.) @ 2015 Taylor & Francis Group London ISBN: 978-1-138-02848-7 Axial capacity design practice for North European wind-turbine projects RJ. Jardine Jmperial College London London UK Dong Energy Wind Power; Gentofic Denmark N.V. Thomsen M. Mygind M.A. Liingaard & C.L. Thilsted sites. This paper reviews the technical and regulatory difficulties for design of axially loaded piles to German ABSTRACT: Improving foundation design is central to the offshore wind industry developing deeper water offshore windfarm projects. It is argued that moving towards reliable forward predictive pile design methods and away from °dynamic proving tests' will be vital to reducing unnecessarily high material and installation costs installation risks and disturbance to marine mammals. Steps are outlined to implement such a change either in bination with regional or international load and resistance factors. 1INTRODUCTION Pile capacity Pile-based multi-footed foundations are used widely installan Pile Jacket for offshore wind turbines and transformer platforms. niptalion True capacity Developments involving deeper water and larger tur- ICP capacity bines are likely to found a larger proportion of wind turbines on piled jackets. Piles supporting offshore Mximum cpacity transformer jacket structures experience parable Yfs = 1.59 shich can be measured loading conditions to oil and gas platforms where Poven capacity pression loading often dominates design. How- aconding s DIN1054 ever light-weight wind turbine jackets often expose Pile capacity before jacket installation (Resaikc) their piles to higher degrees of axial cycling; tension loading cases are also critical. Pile caucity aler pile driving (EeD) Offshore piles are monly designed within the Tdy 10d 100 days Tiae API and ISO frameworks. However the main text spues u Auqeau sood uou sians qudde IdV Figure 1. Schematie illustration of the challemge of proving and gives an unfortably high Coefficient of Varia- capacity according to DEN 1054:2010-12. It also delivers significant biases with respect to tion (CoV) ≥0.70 when assessed against field tests. Length/Diameter ratio (L/D) and sand state. It is over- prediction offers environmental and economic ben- conservative for shaft resistance with dense sands and third-party certifying bodies and national agencies. efits which can only be grasped if accepted by low L/D piles but potentially non-conservative for large diameter piles* end bearing; Jardine et al. (2005). More accurate CPT methods are now cited by APL. Lehane et al. (2005) showed that two of these the 1.1 Challenge of German regulations belorw 0.3. Considerable experience has been gained UWA and ‘full' ICP methods reduce predictive CoV's where the ICP and main text API/ISO approach give The German North Sea presents many dense sand sites in applying the ICP in the North Sea and elsewhere widely different predictions. Current local regulations (00 ) 5661 3s require conformance with the EC7 Framework. Con- The consequences of unnecessary conservatism cerns over the limited databases of very large pile tests reduces with L/D and the challenge of driving long extend beyond additional pile cost. Pile efficiency has led to the German national annex DIN 1054:2010- 12 specifying that the capacities of axially loaded piles two pile-per-leg configuration adding significantly piles into dense sand may force designers towards a must be proven by field measurements made afer pile to jacket costs. Other diffculties include increasing installation. installation failure risks and impacts on marine mam- Offshore static pile testing is usually considered mals. Moving towards more reliable axial capacity on capucities interpreted from dynamic monitoring excessively costly and it is more mon to rely 581
depending on number of piles tested N. Table 1. Correlation factors from DIN 1054:2010-12 1.3 Scope for reducing factors by adopting more reliable static predictive methods N≥ 2 N≥ 5 N ≥10 N ≥ 15 0N Current offshore German practice implicitly assumes 1.50 1.45 1.42 1.40 that the static capacities may be predicted more reli- fos 09′1 1.50 1.35 1.30 1.25 1.25 predictions based on modern site investigations and ably from dynamic measurements than from forward s（s △）- fo6 capacity models. As noted earlier systematic check- 5（6 △）- where no 0.85 and f = 0.1 ing shows this to be untrue. The CoV associated with dynamic interpretation (≥0.95 according to Likins & Rausche 2008) is greater than that applying to the API main text method (CoV 0.7) and far above those of of driving or later re-strikes. However driving mea- modern CPT based design methods ( A6.1 > E6.2 > E3.1 and their signals recorded sands. The ICP approach for clays calls for the Yield fully during driving All drove with 6080 blows per Stress Ratios (YSRs) and Sensitivities (S) of the clzy meter initially increasing to 120160 blows per meter layers to be specified. The YSRs were assessed from towards the final penetrations. A re-strnike test was the puted ss/o ratios and S; was taken conserva- performed on pile A3.2 six days after its installation. tively as 4.0 in the clays. The pression capacity GeoDrive carried out CAPWAP stress wave analyses results are set out in Table 2. Note that the ICP pro- ofthe EoD and the re-strike field data Their interpre- cedure leads to >≥30% lower shaft resistances under tation was designed to meet the DIN 1054:2010-12 tension loading pared to pression in sands requirements. The signal matching quality was good (Jardine et al 2005). It also treats any internal skin leading to the results and ICP pile capacity estimates friction as being relatively small and carried through given in Table 2 and developed as outlined below. the base to contribute a minor part of the overall base The Authors’ ICP axial capacity asessments capacity of large diameter semi-coring piles. adopted submerged unit weights (/) of 10.5 to 11 kN/m? in the main layers and s = 125 kPa in the two thin clay layers present. Pile-specific analyses 3.3 XPI and ICP static capacity estimates were undertaken with the layering tailored to match Borkum Riffgrund peofile led to the pile designers Applying the main text API approach to the generic each local PCPT log. The CPT g profiles were dis- cretized for shaft resistance at 0.5 to 1 m intervals for making the following capacity estimates: capacity assessment based on a moderately cautious interpretation of average values and an upper bound of Shaft outside shaft capacity = 17.81 MN 100 MPa leading to the example shaff-design qc-depth Shaft inside capacity = 16.03 MN profile shown in Figure 5 for CPT A3. Total shaft capacity = 33.84 MN Noting that local variations in strata impact more Annular base capacity = 3.61 MN significantly on pile base capacities than shaft resis- Fully plugged base capacity = 42.76 MN Inner outer shaft annular base = 37.45 MN bound approach was taken for end bearing assessment tance (which reflet the integrated profile) a lower the 36.345.5 MN (external only) range given by the The total API shaft capacities fall 7 to 26% below which assumed that the piles might tip into the lowest 584
below halfof the ICP values if the pile plugs fully. The ICP if fully unplugged coring conditions apply and fall (1) Recognizing that dynamic driving and re-strike tests do not measure medium or long term static API base capacity falls below the ICP estimate if an capacity: The EoD shaft resistances after correc- unplugged annular is assumed but far above it if the tion for dynamic effects are subject to substantial pressive API capacities fall 17 to 39% below the ICP pile is assumed to plug. The overall unplugged - 30% below the medium-term static capacities; scatter and their means are likely to fall 10 to estimates. Noting that foundation stiffness is usually more marked reductions (<70%) apply to end dominated by external shaft resistance and that any bearing- internal shaft resistance relies on mobilizing the rel- pue 1d Kqeai xq uo posq soeo (2) API results also imply a substantially softer response atively soft response of soil beneath the pile toe the more reliable forward predictions of field capac- effective stress based approaches should provide and quite different dynamic behavior during storms to ity. CPT design profiles should be established the ICP assessment. through a mildly conservative interpeetation of st o 3.4 Oufes from Borkum Riffgrumd analysis high resolution into shaft capacity calculations. All ‘high spikes” covering depth intervals less The key points from the moderately conservative ICP calculation outes and Geodrive CAPWAP ble low troughs’ (excluding starts of pushes from than 0.4 m should be eliminated and all credi- assessments listed in Table 2 are: (3) Continuous high quality measurements are inevitable borehole bases) included. (1) The CAPWAP EoD shaft capacities are 36.9 MN ±7% and base resistances 2.9 MN ±21%. required and caution should be given to q val- (2) The mean ratio of the EoD CAPWAP shaft assess- ues exceeding 90 MPa. An absolute upper limit ments to the ICP predictions is 0.93 which of 100 kPa is suggested until more experience is is marginally higher than the early age trend (4) Allowance must be made in any missing section obtained. adopted conservative interface shear angles. A on the full data set including geophysics and lab of PCPT profiles for potentially soft layers based more optimistic bination of = 29° and 20° in the sand and clay layers reduces the CAPWAP/ICP sar[-qns aes so pues Aue uuap o ax aq testing. Piezocones and sample descriptions can (3) The EoD-to-static ratio found applying the API ratio to 0.85. wo[ e Adde pnoqs suouenares Buueq pug (s) method is 1.09 far above that seen in field tests. bound CPT profile. The peobability that a pile tip (4) The base capacities show far more significant mis- moderately high and has to be addressed in design. will inadvertently terminate in a soft layer is often matches. The mean CAPWAP EoD/ICP ratio is 0.35 even though a lower bound q: selection was (6) Site specific interface ring-shear tests should be made. The API *annular’ estimates are closer to carried out wherever feasible. Such tests can the dymamic estimates. be highly cost-effective as even modest changes (5) The re-strike indicated pile A3.2's shaft resis- can affect capacity significantly. Conservative tance increased by 45% over 6 days rising more assumptions should be made if the data are sharply than the research tests plotted in Fig. 2. The (7) Allowance should be made for clay being dragged unavailable. restrike CAPWAP/ICP ratio α 1.50 and reduces to 1.35 if the more optimistic parameter set is down 1 m below any sand/clay interface reduc- adopted. The re-strike shaft capacity exceeds the ing the & angle applied over the top m of any API unplugged *coring* estimate by 69%. underlying sand layer to that of the overlying clay. (6) In this case the base capacity shows a remark- (8) Allowance should be made for the effects of lat- ably set-up ratio (1.83) over six days. However eral andaxial lod cycling as describedby Merrit the re-struck value still falls well below the ICP et al. (2012) and Jardine et al (2012). These fac- prediction. We recall that the latter is intended to tors are likely to reduce operational axial capacity. predict the base resistance available after a pile It is necessary to characterize the distributions of head settlement of D/10 (213 mm) which is orders cyclic forces developed at the head of each pile of magnitude greater than the set developed on made of cyclic loading tests Jardine and Standing develop in the design storm. Use may then be re-striking. peofile under extreme conditions to estimate how (2012) and the pile's lateral displacement-depth 4 ENGINEERING RECOMMENDATIONS (9) Pile ageing should be addressed when planning axial and lateral cycling degrades shaft capacity. The experience gained at Borkum Riffgrund and other sites led to ten recoeumendations as to how developers and interpreting SRD or re-strike data; see Jardine et al. (2006) and Rimoy & Jardine (2015). Any pile axial capacities from dynamic testing and towards can move with caution away from *proving’ driven analysis of the test data-bases should differentiate more accurate forward prediction methods supported m pue s-sng uaaaq Aqeap Kaan by dynamic monitoring. tested’ cases which can show quite different trends. 585
北欧风力发电机组项目的轴向容量设计实践.pdf
