THE USE OF NUMERICAL ANALYSIS TO AID THE DESIGN OF MONOPILE FOUNDATIONSFORANORTHSEAOFFSHOREWINDFARM Felix C. Schroeder Geotechnical Consulting Group London UK f.c.schroeder@gcg.co.uk Angeliki Grammatikopoulou Geotechnical Consulting Group London UK Richard J. Jardine Imperial College London UK r.jardine@imperial.ac.uk David M. Potts Imperial College London UK d.potts@imperial.ac.uk ABSTRACT approach to support foundation design for the Triton Knoll array of monopile supported wind- turbine generators which are sited offthe UK's east coast and will deliver 855 MW from 2021. laboratory testing that allowed location-specific locally calibrated 3D FE analyses with the stiffness and yielding behaviours of the generally dense sands and stiff glacial till units encountered. The ability of the chosen constitutive models to accurately replicate the most critical features of soil response is demonstrated by paring single element FE analyses monopile analyses undertaken of a static loading case that provided predictions for ultimate and service limit states and allowed the extraction of soil reaction curves for use in further design optimisation. The advanced site characterisation and numerical analysis enabled the foundation designers to achieve substantial savings in the project's foundations' capital costs. Keywords: monopile foundations finite element analysis constitutive modelling INTRODUCTION and will deliver 855MW from 2021 onwards. Given the relatively large proportion of the project's capital cost invested in its foundations efficient design of TK's 90 monopiles in water lateral and moment loading capacities often offer efficient design solutions and have been adopted for most offshore WTGs installed to date.There has been a progressive trend over piles to resist lateral and moment loads is frequently based on calculations undertaken using APl (2011) are stil monly employed for sand and clay sites. These expressions are relatively easy to use and only require basic soil parameters such as undrained strengths in clay deposits and angles of shearing resistance in sands. However foundation monitoring. North Sea tills and dense sands. Far better Class A predictions (i.e. made before the field measurements Lambe 1973) were obtained forthese sites by performing finite element (FE) analyses with the Imperial College code ICFEP (Potts & Zdravkovic 1999) that recognised and
high quality samples (see Jardine and Potts (1988 & 1993) Kinley & Sharp (1993)). LINCON HIRE Fig. 1. Location of Triton Knoll Offshore Wind Farm (from .tritonknoll.co.uk) The recent PISA and PISA2 JIPs (e.g. Byron et al. 2019) developed new design prise:(i) accurate characterisation of ground conditions that enables the use of representative constitutive soil models (i) advanced 3D FE analyses conducted with ICFEP and (i) development from these analyses of four sets of calibrated soil reaction curves that design process for an OWF. The methods’ ability to deliver greatly improved predictions for monopile response under lateral and moment loading was demonstrated by the close agreement achieved between Class A predictions and the outes of large-scale field experiments conducted in glacial clay tils and sands as part of the PISA programme. This paper describes how the PISA methodology was applied at a relatively early stage of the summarises the work undertaken to characterise the soils encountered calibrate the determine the expected damping behaviour and the use of a numerical procedure which enables the determination of permanent deformation for given design storms. Due to space GROUNDCONDITIONS MATERIALMODELSANDASSOCIATEDPARAMETERS The Triton Knoll OWF project area is characterised by Quaternary sediments of varying thickness. The dominant Pleistocene deposits were deposited in alternating glacial and s monopile with an embedment depth of 21m at a location prising the following geotechnical formations: Holocene deposits;assumed to be removed by scouring following monopile installation (less than 1.3m thick). Bolders Bank Formation (BDK); an ablation till (primarily overconsolidated ductile
medium dense sand deposited in estuarine/intertidal and glaciomarine/ marine environments during the Elsterian period. Swarte Bank Formation (SBK);channel infill materials deposited in lacustrine to shallow marine interglacial conditions which can be separated into three principal units the basal till unit thought to contain a high proportion of reworked chalk. underlain by 6.5m of EG 11m of SH and an assumed 7.5m of SBK overlying basal chalk. The to specific locations on the basis of CPT profiling at each turbine position. (Pa) 80 R(S) 2(MP) BOK(S) 62 Fig 2. CPT profile and assumed stratigraphy BoldersBankFormation(BDK)and other clayunits(SH andSBK) in-situ soi testing at Cowden where the till section concerned is a member of the BDK the members present at TK should be split into upper and lower sections BDK(U) and BDK(L) with the latter being generally stiffer and stronger. plane. The MCC pressibility parameters A K and v were derived from 9 oedometer tests
extension tests were analysed to establish critical state angles of shearing resistance Φ'cs (see s) s jo (g 090 1. I (e b) Fig.3.Determination of critical state parameters (pcs)for BDK(U) and BDK(L) layers a) Drained triaxial test data and b) Assumed variation of 'cs with Lode’s angle 5000 • Lab data 1000 FE ideaisation 1000 0.0001 0.001 s [%] 0.01 0.1 Fig. 4. Derivation of stiffness-strain curve on the basis of laboratory test data for BDK(U) Figure 4 shows the variation of normalised secant undrained Young’s moduli Eu with shear strain invariant Es obtained from undrained triaxial pression tests on BDK(U) employing local axial strain measurements (where Es is equal to axial strain in undrained triaxial tests). Isotropic small strain stiffness modelling was adopted in the analyses in line with the PISA reproduce stiffness non-linearity. Undrained triaxial tests were simulated in single element FE analyses; the predictions for pared with a typical CAUc triaxial pression test on overconsolidated tillin Fig. 5 which shows ductile behaviour up to 30% strain. The initial K values for these simulations were chosen to match the stress conditions of each sample at the beginning of undrained shearing. S measured in each test.The agreement between the simulation and the test data is
125 Lab date 100 -FE simulato 00 09 lab-data =FE simulation 25 Axial strain(5) -100 s'(KPa) 100 4 000 Lab data 50 1 000 100 (uge eoey 10000 0.001 Axial atrain(%) 100 0.1 Fig.5. Single element ICFEP simulation of CAUC test for BDK(U) 001 Su (IPa) 009 800 1000 (nwo8 (1) x08 FE-scoer SK(U) SN(L) (0=30ng (s =20g (e b) Fig. 6. a) Undrained strength profile and b) assumed YSR profile
数值分析在北海近海风电场单体基础设计中的应用.pdf
