System-Level Simulation of Floating Platform and Wind Turbine Using High-Fidelity and Engineering Models
Paterson, Eric G.
MetadataShow full item record
Compared with inland wind farm, the offshore sites has higher energy density, and less environmental and visual impact. It's an attractive energy source for populous coastal cities. However, the marine environment adds more uncertainties in the system, especially for floating wind turbines. Due to physical conditions, full-scale testing is usually unpractical, thus system-level simulation is essential in design stage. Cyber Wind Facility (CWF) project aims to provide highly resolved 4D cyber data to answer the key performance questions: structure and turbine loading, shaft torque, platform motion. In this study, a full-scale NREL 5MW wind turbine and OC3 spar buoy with mooring lines (figure 1) are simulated in open-source code OpenFOAM. The whole system is calculated by three different models: actuator-line model (ALM), which calculates aerodynamic force from turbine blades; quasi-static model, which estimates the restoring mooring force in each time-step, and high-fidelity model for the floating platform in waves. The ALM model is less computationally expensive than resolving full turbine geometry, it represents the blades as a set of actuator elements and the loading is distributed along the lines (figure 2). Sectional force at each blade element is computed according to local flow conditions and airfoil lookup table, the lift and drag forces are projected onto the flow as body forces in the momentum equation. The original ALM was developed for xed-bottom turbines (Jha et al., 2014), modifications are needed to incorporate 6DOF motions from the turbine tower. The floating platform, i.e. OC3 spar buoy, measures 130m in length with 120m of total draft (Jonkman, 2010). Single harmony linear waves are generated by wave2Foam library (Jacobsen et al., 2012), numerical beach is included for absorbing waves. The quasi-static mooring-line model is developed from catenary-line equations (Faltinsen, 1990) and it is also implemented in engineering tool hydrogen (Jonkman, 2007). By calculating the anchor and fairlead position, together with known physical properties of the cable, we can solve both the mooring-line configuration and restoring force. In general, it 1) models individual taut or slack mooring lines; 2) accounts for weight and buoyancy, axial its ness, and friction from seabed; 3) ignores bending and torsional sti nesses, cable inertia and hydrodynamic forces, and 4) Solves for cable positions and tensions under static equilibrium given the instantaneous fairlead location. Due to the numerical instability in standard dynamic meshing multiphase solve in OpenFOAM, especially in the presence of large displacement, a modified tightly-couple RANS solver (Dunbar, 2013) is used for the high-fidelity simulation. It features sub-iteration in each time-step to ensure simulation is converged with respect to mesh motion and dynamic relaxation is introduced for faster convergence. Furthermore, it is validated by experimental data of simple 2D cylinder and OC4 semisubmersible platform (Robertson et al., 2014). In summary, this study integrates ALM and the quasi-static mooring-line model. By using wave generating tool and tightly-coupled solver we can study the 6DOF motion of floating platform in waves with less computational resource than fully-resolved high-fidelity model.