No abstract

The design of floating wind turbines needs the validation of numerical models against measurements obtained from experiments that accurately represent the system dynamics. This requires solving the conflict in the scaling of the hydrodynamic and aerodynamic forces that arises in tests with wind and waves. To sort out this conflict, we propose a hybrid testing method that uses a ducted fan to replace the rotor and introduce a force representing the aerodynamic thrust. The force is obtained from a simulation of the rotor coupled in real time with the measured platform displacements at the basin. This method is applied on a test campaign of a semisubmersible wind turbine with a scale factor of 1/45. The experimental data are compared with numerical computations using linear and non-linear hydrodynamic models. Pitch decays in constant wind show a good agreement with computations, demonstrating that the hybrid testing method correctly introduces the aerodynamic damping. Test cases with constant wind and irregular waves show better agreement with the simulations in the power spectral density's (PSD's) low-frequency region when non-linear hydrodynamics are computed. In cases with turbulent wind at rated wind speed, the low-frequency platform motions are dominated by the wind, hiding the differences from hydrodynamic non-linearities. In these conditions, the agreement between experiments using the proposed hybrid method and computations is good in all the frequency range both for the linear and the non-linear hydrodynamic models. Conversely, for turbulent winds producing lower rotor thrust, non-linear hydrodynamics are relevant for the simulation of the low-frequency system dynamics.

The design of floating wind turbines requires both, simulation tools and scaled testing methods, accurately integrating the different phenomena involved in the system dynamics, such as the aerodynamic and hydrodynamic forces, the mooring lines dynamics and the control strategies. In particular, one of the technical challenges when testing a scaled floating wind turbine in a wave tank is the proper integration of the rotor aerodynamic thrust. The scaling of the model based on the Froude number produces equivalent hydrodynamic forces, but out of scale aerodynamic forces at the rotor, because the Reynolds number, that governs the aerodynamic forces, is not kept constant. Several approaches have been taken to solve this conflict, like using a tuned drag disk or redesigning the scaled rotor to provide the correct scaled thrust at low Reynolds numbers. This work proposes a hybrid method for the integration of the aerodynamic thrust during the scaled tests. The work also explores the agreement between the experimental measurements and the simulation results through the calibration and improvement of the numerical models. CENER has developed a hybrid testing method that replaces the rotor by a ducted fan at the model tower top. The fan can introduce a variable force which represents the total wind thrust by the rotor. This load is obtained from an aerodynamic simulation that is performed in synchrony with the test and it is fed in real time with the displacements of the platform provided by the acquisition system. Thus, the simulation considers the displacements of the turbine within the wind field and the relative wind speed on the rotor, including the effect of the aerodynamic damping on the tests. The method has been called “Software-in-the-Loop” (SiL). The method has been applied on a test campaign at the Ecole Centrale de Nantes wave tank of the OC4 semisubmersible 5MW wind turbine, with a scale factor of 1/45. The experimental results have been compared with equivalent numerical simulations of the floating wind turbine using the integrated code FAST. Simple cases as only steady wind and free decays with constant wind showed a good agreement with computations, demonstrating that the SiL method is able to successfully introduce the rotor scaled thrust and the effect of the aerodynamic damping on the global dynamics. Cases with turbulent wind and irregular waves showed better agreement with the simulations when mooring line dynamics and second order effects were included in the numerical models.