Energy production with offshore wind turbines (OWTs) is becoming more popularand is predicted to continue to grow as an industry. OWTs are top heavy and dynamically sensitive structures, experiencing continuous environmental cyclic loadingthroughout the operational lifetime from the action of wind and waves. With construction sites in shallow water around the UK quickly reaching capacity, deeperwater solutions are being explored, where environmental conditions are more severe,necessitating an increased focus on the long-term cyclic behaviour of OWTfoundations.The successful design of OWT foundations involves load controlled analyses integratedin the time domain, and without the inclusion of the effects of very largenumbers of cycles of dynamic loading on the soil condition, the final design can beover-conservative and uneconomical. OWTs are designed with a natural frequency ina narrow band between two external excitation frequencies, and care must be takento ensure this frequency will not change significantly during the OWT’s lifetime. Ifit does, it could suffer from amplified motions caused by resonance, leading to overturningfailures or expensive repairs. Additionally, the accumulation of base rotationfrom continued environmental loading can result in the requirement of additionalcostly maintenance over the operational lifetime.The presented results from this PhD project propose modelling methods ofmonopile-mounted OWTs over the long term lifetime of their operation, with particular consideration of the interaction of the monopile foundation with the soil. The models discussed incorporate soil degradation effects, and quantify the subsequent impact on the natural frequency and other design drivers of the OWT.In this thesis, new models are proposed to reproduce observed experimental behaviour of piles undergoing both gapping and ratcheting due to soil degradation.The models are based on a beam-on-nonlinear-Winkler-framework (BNWF) frameworkapproach, using nonlinear springs to represent the effects of soil reaction. Novelmethods have been devised to recreate both gapping (separation of pile and soil dueto repeated loading) and ratcheting (accumulation of rotation from partial/one-waycyclic loading), and both provide good reproductions of laboratory and/or field experimental results. The models are subsequently combined to give a picture of thecoupled effects of both ratcheting and gapping.The importance of a full picture and model of gap formation is demonstratedwith comparison to a simple scouring model that does not include interaction withthe soil at the extremities of the environmental load. It is found that a completepicture of soil erosion and gapping down the entire length of the pile is required toaccurately assess the effect on the natural frequency.The ratcheting model is capable of recreating the accumulated displacement forloads of varying magnitude and shape. It is found that the ratcheting model canrecreate observed experimental behaviour of the dependence on load packet ordering.It is seen that large magnitude one-way cyclic loads following small magnitude oneway cyclic loading results in greater ratcheting than the reverse. Additionally, it isseen that large numbers of small magnitude cyclic loads reduce the impact of a laterlarge magnitude load. It is important for models used to capture the ratchetingof OWTs to include this impact of load packet ordering, since offshore loading ofwind and wave contains many loads of differing magnitude, so considering worst-casescenarios is important for the safety factor in design.In addition, the methods have been extended from 3 degrees-of-freedom (DoF)2-D planar motion, to a full 6 DoF 3-D beam element model. In the case of theratcheting approach, this includes a new method of cycle counting and partitioningto reproduce the increased ratcheting effect that results from multidirectionalloading, versus pure unidirectional loading seen in experimental literature. Thisis of particular importance to OWT modelling, since the external wind and waveloads will not necessarily be aligned, and this could result in additional ratchetingthan had been predicted at initial design and installation. By modelling an OWTand simulating harmonic wind and wave loading, it was found that in the case thatthe wind and wave loads are orthogonal, 2.4 times more ratcheting occurred at the mudline from 1000 wind cycles, when compared to the unidirectional load case. Byincluding the gapping model in the 6 DoF model, it is found that it can recreateboth the coupling that can occurs between orthogonal directions when a load directionchange occurs, as well as the loss of strength in the orthogonal soil that can beobserved after a load cycle in the initial direction.The ethos of the modelling techniques established in this project is to give designersa simple and expedient way to include soil degradation mechanisms into existingfinite element analysis (FEA) models, without the requirement to determine manycalibration parameters.
Nonlinear Soil Models for Predicting Long Term Dynamic Response of Monopile Supported Offshore Wind Turbines: (Alternative Format Thesis)
Williams, S. (Author). 27 Mar 2024
Student thesis: Doctoral Thesis › PhD