Abstract
In the field of stochastic dynamics of marine structures, the determination of long-term extreme responses is a crucial aspect to ensure the desired level of structural reliability. The calculation of these responses requires precise knowledge of the environmental conditions and reliable methods to predict the values associated with a reliability target level. While there is a very precise method to determine the value of these extreme values, e. g. the full long-term analysis (FLTA), this approach is computationally expensive. Then, approximated methods are needed.
One practical approach for the determination of the most relevant environmental conditions for extreme calculation is the environmental contour method (ECM). However, some limitations have been detected when this method is used for offshore structures that consider survival strategies e. g. offshore wind turbines (OWT). Lastly, a modified ECM procedure (MECM) has been developed with the purpose to bypass the limitations of the traditional ECM. This method is based on short-term simulations and through an iterative process by testing many environmental contours in the operational range allows finding an important wind speed with its corresponding return period and thus, the problem that traditional ECM has, is avoided.
The environmental conditions, which are represented by a large number of parameters, are also an important aspect of extreme calculation. Whereas some of them are treated as stochastic values, some are considered deterministic and, therefore the existence of uncertainties in their measured/estimated values is inevitable. These uncertainties are addressed by adopting values recommended by standards and guidelines and, in practice, it is often necessary to be conservative when there is a lack of information about the specific site studied. Therefore, the understanding of the impact that these uncertainties can have on the loads/responses that govern the design of offshore structures, especially wind turbines, is of great relevance. In this work, the influence of uncertainty in the wind shear coefficient (WSC) is studied. This parameter is directly related to one critical environmental condition i. e. wind speed at hub height, and its influence in power production and fatigue loads has been documented in the literature, but, few cases have addressed their influence in bottom fixed OWT responses.
This work seeks to highlight the relevance of an accurate selection of shear coefficient and, its influence on the probabilistic analysis of a bottom fixed OWT taking into account that considerable variations from recommended values may occur. Through the use of coupled simulations in FAST, the NREL 5MW wind turbine will be subjected to varying wind shear conditions, and the corresponding 50-yr long-term responses will be calculated considering the MECM to take into account the influence of the wind turbine survival mode. The extreme values are fitted from a Global Maxima Method (GMM). Finally, it is sought to relate the uncertainty in a relevant input parameter (i. e. WSC) with the uncertainties propagated to the output parameters (i. e. extrapolated long-term extreme responses).
Neural Networks Applied to the Wave-Induced Fatigue Analysis of Steel Risers
