Floating wind turbine drivetrain multidimensional reliability evaluation by Gaidai hypersurface methodology

Structural safety and reliability evaluation is crucial for renewable energy installations, as Wind Turbines (WT) are built to endure excessive loads inflicted by ambient environmental loads. The presented case study introduces a state-of-the-art multivariate hypersurface risk assessment methodology, particularly suitable for the reliability study of high-dimensional structural systems with multiple failure modes. The presented analysis considers one of the most critical parts of WT, namely, the drivetrain and gearbox. As the presented case study is primarily methodological, a deeper structural reliability study that includes bearings has not been performed. The advocated multivariate reliability method can account for a virtually unlimited number of gearbox hotspot stresses, as opposed to existing reliability techniques that are mostly limited to univariate or bivariate cases. Novel Gaidai Hypersurface (GH) multivariate structural risk evaluation methodology presented in the current case study is particularly appropriate for evaluating multimodal structural risks pertinent to large semi-submersible floating wind turbines (FWT) operating under excessive hydrodynamic and aerodynamic loads. The main objective of this study has been to establish a baseline for a novel multimodal hypersurface design approach that would enable relevant data to be extracted from numerically simulated or measured time histories, jointly representing FWT drivetrain dynamics. The methodology of the current study offers a generic way to predict failure and damage risks for nonlinear, nonstationary multimodal FWT structural dynamics in an accurate yet efficient manner. The non-dimensionalized design values (also called characteristic values) constitute the necessary structural design input, provided the return period of interest. The structural failure (damage) probability and Confidence Intervals (CI) are reported. The parametric extrapolation scheme was verified against the conventional Gumbel fit scheme.