On Design Load Prediction and Extreme Response Modelling of Floating Offshore Renewable Energy Structures

Abstract

FLOATING offshore renewable energy (FORE) is a marginal industry and as such cost reductions are essential in order for it to become competitive with other forms of electricity generation. Design practices for evaluating the extreme responses of FORE devices are still under consideration by standards authorities, thus methods that enable potential speeding up of the design and development phases, particularly physical model testing are needed. Probabilistic design approaches have proved effective in bringing down costs in the aerospace and automobile industries and so provide a potential route to viability by allowing partial safety factors to be reduced. However, the scarcity of real world data and early development stage of the sector mean that the suitably large data sets needed to evaluate extreme responses and uncertainties have to be generated numerically. This poses a challenge when there is a trade-off between fidelity and run time. This thesis seeks to understand the extent to which fast mid-fidelity models such as WEC-Sim (a potential flow, Cummins equation based model) can model the extreme responses of FORE devices compared with physical experiments and where and how short design waves can be used in speeding up the prediction of design loads.

A WEC-Sim numerical model was found to perform reasonably accurately in evaluating extreme mooring responses of a generic point absorber wave energy convertor (WEC), with a median error relative to physical experiments of around 10%. It was then used to develop a methodology for using short design waves during physical model tests, conducted in the COAST lab at the University of Plymouth, and evaluated for a hinged raft type WEC and semi-sub floating offshore wind turbine (FOWT). The approach used constrained focused waves scaled to an inflated target percentile amplitude or response in place of the one to three hour irregular wave time series commonly used in the prediction of design loads. Recommendations were given on the suitability of the method in different sea conditions; for different response types; and on how to calibrate the waves in physical experiments. Constrained NewWave (CNW) and conditional random response wave (CRRW) profiles were compared and found to perform well in different situations, with the CNWs being a better choice when modeling snatch loads in steep sea states for a hinged raft type WEC. It was found that a single, frequency specific phase correction calculated from a focused wave could be used to calibrate the constrained focused waves, greatly speeding up the calibration time. Difficulties in the calibration of the short design waves in steeper sea states make the developed method increasingly difficult to apply close to the wave breaking limit.

Physical experiments on a semi-sub FOWT were conducted and the results compared with a WEC-Sim model for the responses of pitch, nacelle acceleration and mooring load. It was found that an additional drag term had to be added to the numerical model to improve the accuracy of the low frequency surge motions and mooring loads. Furthermore, it was found that the wave theory used to generate the wave surface elevation input to the edited numerical model had a significant impact on the surge response, resulting in a difference of over 30%. A constrained wave group was proposed as an alternative short design wave when studying the moorings of Semi-subs. This thesis has developed a short design wave method to improve the efficiency of the design process of FORE devices, provides several case studies on the methods applicability and has assessed the extend to which mid-fidelity numerical models are able to model extremes.