Design of large diameter monopiles for offshore wind turbines in clay

Offshore wind power has great potential as a clean and renewable energy source that is capable of reducing our reliance on fossil fuels. The main drawback of offshore wind power is its comparatively high capital cost. One area in which this cost can be reduced is by optimizing the design of these structures. More efficient foundation designs is key in this regard. The p-y method is extensively used for the design and analysis of laterally loaded piles due to its simplicity and versatility. Matlock (1970) or the API RP 2GEO (2011) “soft” clay p-y model is the guideline of choice for normally consolidated to moderately overconsolidated clays. However, this p-y model is not yet verified for piles with very large diameters and low aspect ratios. Design of wind turbine monopiles is governed by serviceability limits such as the natural frequency of the structure and the accumulated tilt under long-term low-amplitude cyclic loads, but these guidelines have not been verified for serviceability limit state designs. The main objectives of this study were to: (a) assess the ability of the Matlock (1970) p-y model to accurately model the behavior of laterally loaded piles at both small and large displacements, (b) investigate the effect of gapping on the backside of laterally loaded piles and develop a theoretical framework to quantify its effect and predict its occurrence, (c) re-examine the derivation of lateral bearing capacity factors (N p ) used in published p-y models, (d) evaluate the effect of large numbers of small-amplitude cyclic load on the stiffness and the post-cyclic ultimate capacity of laterally loaded piles, (e) assess the ability of the Matlock (1970) p-y model to adequately account for pile diameter effects, (f) assess the ability of the Matlock (1970) p-y model to accurately predict the behavior of a pile in a variety of undrained shear strength versus depth profiles, (g) assess the ability of published p-y models to accurately predict the natural frequency of wind turbine structures. The methodology consisted of analyzing field tests, laboratory model tests (1-g and centrifuge), and numerical modeling. An extensive database of field tests and laboratory centrifuge tests was compiled. This data was then supplemented by a series of 1-g model tests in a variety of clay test beds (normally consolidated to heavily overconsolidated, kaolinite and Gulf of Mexico clay) carried out at The University of Texas at Austin and 3-d finite-elements models using Abaqus carried out by Ensoft Inc. The following conclusions were drawn from this study: (a) Matlock (1970) p-y model underestimates the lateral soil resistance on piles in normally consolidated and overconsolidated clays, regardless of pile diameter or aspect ratio, (b) the effect of gapping plays an important role in determining the pile response as it can lead to a loss of capacity and a reduction in stiffness, (c) lateral bearing capacity factors used in the Matlock (1970) model are too low, (d) the degradation in the stiffness of the pile response, when subjected to cyclic loading, was limited to approximately 30% and occurred within the first 100 cycles, (e) the method of normalizing used in the Matlock (1970) model successfully accounts for pile diameter effects, (f) estimates of the natural frequency of wind turbine structure based on the API RP 2GEO (2011) p-y model are lower than those based on the Matlock (1970) and Jeanjean (2009) p-y models.