RESULTS
Expected results
The FLOATECH project aims to greatly improve the state of the art of coupled tools in order to enable the capture of important physical couplings between aerodynamics, structural dynamics and hydrodynamics which are incident upon large flexible FWT structures. This is achieved by developing higher order simulation tools for aerodynamics and hydrodynamics and coupling these together in the one-stop high order simulation package QBlade-Ocean. The work progress results are further described below.
Aerodynamic and Wake Modelling
FLOATECH proposes to introduce Lagrangian vortex particle method in the simulation of FWTs (WP1). This allows the resolution of only the region of interest in the wake, which contains active vorticity, and provides the ability to make use of adaptive grid refinement to ensure that regions of interest, such as shear layers or the turbulent wake transition, are correctly modelled without increasing too much the computational cost. Furthermore, the combination with a regular Eulerian particle grid provides an excellent solution for real-time modelling and allows for simple coupling to the aforementioned near-wake models. The use of the streamfunction-vorticity treatment of the flow field also allows the use of highly optimized spectral solvers , along with the implementation of Large-Eddy Simulation turbulence models . The newly developed Vortex Particle Multilevel Method at TUB, that will be embedded in QBlade-Ocean, allows for greatly accelerated simulations in combination with execution on parallel computing systems and GPU execution.
Structural Modelling
QBlade-Ocean shall contain the same high-order nonlinear structural model as that contained in the Chrono solver. This has been well validated and has shown superior performance both in terms of accuracy and high performance with regard to other solvers. This can hence be considered to be prior work within the pre-existing QBlade package. Although it would be possible to implement a higher-order structural model, such as a finite-element treatment, this would greatly complicate the necessary turbine definition for users and for general application to turbine design, and would be more suitable for a purely research-based simulation tool.
Hydrodynamic Load Modelling
The flexible and general beam element formulation of the Chrono structural solver allows for a simple development of the structural representation of the floater and the sub-structures. This provides an immediate coupling interface with a hydrodynamic model to allow for structural deflections as caused by diffraction or radiation wave loads.
In FLOATECH’s WP1, the hydrodynamic solver NEMOH will be integrated into the QBlade suite. NEMOH is a boundary element method code dedicated to the computation of first order wave loads on offshore structures (added mass, radiation damping, diffraction forces). While usually such a hydrodynamic solver is used as an external solver, NEMOH will be entirely integrated in the simulator. As an input, the user will have to provide a mesh of the floating substructure. As outputs the NEMOH will provide the hydrodynamic coefficients used for solving the equation of motion of a rigid floating body as well as the pressure impulse-response that will be used for calculating the structural loads in the floating sub-structure.
Wave Modelling
Wave data calculated by HOS-Ocean will be coupled to a strip theory approach for the calculation of the wave-induced loads, where at each time-step the pressure and velocity of fluid will be interpolated in time and space on each discretization points of the floating substructure. This shall be included in the integration within QBlade-Ocean to allow the user to select nonlinear incoming wave fields and therewith have a much more physically realistic representation of the sea-state.
FOW Control Strategies
The innovative feed forward wave-based control in WP3 exploits the deterministic nature of incoming wave behaviour together with detailed turbine structural definition using intelligent feed-forward self-learning control algorithms that maximize the power production, while constraining the fatigue loads on the FWT structure. The floater design will be optimized in conjunction with the controller. However, rather than the brute-force methods used therein, novel integrated optimization routines developed in this project will pave the way for commercialization of the technology, allowing the computationally efficient joint optimization of structural and controller parameters.
WP4 exploits the inherent dynamic nature of a FWT to its advantage. While often seen as a disadvantage of FWTs, the dynamic movement of the floater can be leveraged to excite fundamental instabilities in the wake structure behind the turbine rotor. Designing the floater to oscillate at a certain frequency is expected to produce similar effects as the concept of dynamic induction control, as presented in the wind farm control literature. This opens the door to advanced wind farm control methods at a much smaller cost (increase in actuator load) compared to conventional fixed-bottom wind farms, with the potential of increasing the annual energy production significantly.
Deterministic wave prediction applied to floating wind turbines
Of all components of a deterministic wave prediction chain, off the shelf remote wave sensors are by far the most affected by the multiphysics environmental conditions. The dynamic behaviour of FWT resulting from complex interaction of environmental forcing as well as control strategies prevents the usage of the state of the system to be exploited as a control parameter for the calibration of the input remote wave data acquired as done for ship motions. This project will thus bring at sea a remote deterministic wave sensor mounted a full-scale wind turbine at test on an offshore test site where in situ collocated and synchronized measurements from met-ocean sensors will provide control parameters for the off-line validation of the prediction chain. This will ensure the operational applicability of the feed forward wave-based control strategy.