Quantifying the numerical uncertainty for CFD simulation of Floating Offshore Wind Turbines in pitch decay motion

Compared with the in-situ experiment-based techniques, numerical tools have been widely used in predicting the responses of floating offshore wind turbines (FOWTs) owing to their low cost and higher accuracy levels. Recently, several engineering models with mid-fidelity tools have been cross-verified and validated with both numerical and experimental data. Unfortunately, these mid-fidelity tools proved ineffective as they significantly underpredicted the responses of semi-submersible FOWTs, especially under low-frequency nonlinear wave excitations. Lately, there has been a gradual development of high-fidelity CFD simulations capable of accurately capturing all the nonlinearities and viscous damping effects. Besides, it has been established that using CFD tools for cross-verification and validation could help improve the efficiency and accuracy of engineering-level tools.

The compatibility of accurate hydrodynamics prediction of floating offshore wind turbines with CFD tools has been extensively studied. Specifically, these tools have been used to evaluate the Hywind floater, a spar-type floater used as the primary model in many studies. CFD tools have proved beneficial in tunning the inputs of low-cost mid-fidelity tools and quantifying the viscous effects and nonlinearities that are difficult to quantify numerically. However, most of these studies have performed grid refinement studies and sometimes time-step studies, but almost never numerical uncertainty quantification. To accurately conduct the validation against experimental measurements, it is important to evaluate the total numerical uncertainties of numerical models. This is yet to be fully realized to date.

Herein, PhD candidate Yu Wang and Professor Hamn-Ching Chen from Texas A&M University in collaboration with Dr. Arjen Koop from Maritime Research Institute Netherlands and Dr. Guilherme Vaz from Blue Ocean Sustainable Solution performed an innovative verification and validation study to accurately estimate the numerical uncertainty of CFD simulations for a semi-submersible floating offshore wind turbine in pitch free-decay motion. The simulations were performed with both dynamic and linear mooring models to derive the quadratic and linear damping coefficients of the semi-submersible. A total of 20 simulations were performed. They started by estimating the spatial and temporal discretization uncertainty, followed by validations with uncertainties in pitch motions. Their work is published in the journal, Ocean Engineering.

The research team reported that the iterative uncertainty had little impact on the total numerical uncertainty and was neglected. The mean and maximum discretization uncertainty values for pitch motion were obtained as 0.236 and 0.531 degrees, respectively. Using a dynamic mooring model significantly improved the validation results. It accurately predicted the natural pitch period though it slightly under-predicted the quadratic damping coefficient. It was found that the total numerical uncertainty originated purely from the discretization uncertainty. Furthermore, a guideline was provided to guide future verification and validation studies of CFD tools for marine structures subjected to decay tests.

In summary, a detailed verification and simulation study was conducted towards CFD simulations of FOWT with a semi-submersible platform in free decay motion. The numerical uncertainties were successfully validated against the experimental measurements. Importantly, the predicted linear damping coefficient and pitch natural period were in good agreement with the experimental results. In a statement to Advances in Engineering, first author Yu Wang said the new study provided valuable insights that will advance free-decay dynamical analysis of floating marine structures.