As ocean current turbines move from the design stage into production and installation, a better understanding of both oceanic turbulent flows and localized loading is needed by researchers and members of industry. In this study, an ocean current turbine is simulated for a variety of realistic ocean turbulence environments using the National Center for Atmospheric Research (NCAR) Large Eddy Simulation (LES) model. The ocean current turbine is represented in this model by an actuator disk and flow characteristics are measured at the turbine as well as in the turbine wake. Turbulent inflow conditions are generated using two methods: (i) a precursor simulation where the NCAR LES model is run without the ocean turbine present and (ii) a synthetic turbulent flow generator adapted for the ocean from the National Wind Technology Center TurbSim tool. In both cases, the inflow and boundary conditions will be designed to represent conditions during an observational campaign at Admiralty Inlet in Puget Sound; comparisons will be made between the simulation results and available measurements. Significant results from this study will be discussed, including the effects of cyclic gravity waves on turbine stresses, statistical comparison of the two inflow generation methods, examination of turbulent mixing in the turbine wake, and analysis of turbine loading. Future areas of research are identified, including incorporation of more complex boundaries, improvements to the ocean synthetic turbulence generator, and representation of the turbines using actuator lines.
OpenFOAM has shown great promise as a computational tool for wind energy applications. In this presentation, we discuss the work done at CU Boulder to develop a computational testbed in OpenFOAM based on numerical tools created by the U.S. National Renewable Energy Lab (NREL). We generate neutral and unstable atmospheric boundary layers in a precursor simulation, and compare those turbulent fields to synthetic inflow turbulence generated by the NREL National Wind Technology Center (NWTC) stochastic turbulence generator TurbSim. We then simulate the performance of a 1.5 MW utility-scale wind turbine using actuator line simulations as well as the NWTC FAST aeroelastic code. These simulations are used to quantify the value of higher-fidelity models both for inflow turbulence and turbine dynamic response. Ultimately, the computational results will be validated against remote sensing data from an experimental campaign at NWTC involving a scanning LIDAR upwind of the turbine and a tethered aerostat downwind of the turbine. We discuss how these measurements will be used to examine LES model accuracy for a variety of atmospheric stability conditions. Finally, we outline future extensions of the computational testbed that will include actuator elements for the wind turbine nacelle and tower structure, as well as enable wind farm level control schemes.
In recent years, there has been increasing interest in using tides as a clean, renewable energy source, but in order to move from the design stage to production, a more comprehensive characterization of turbulence is needed to predict loads on a turbine in a tidal environment. The commonly used turbulence intensity metric does not describe turbulent length or time scales, nor does it give any directional information. Using acoustic Doppler current profiler (ADCP) data from Admiralty Head, WA, several additional statistics are calculated to more fully characterize the turbulence that a turbine would experience in a tidal flow. These metrics include coherent turbulent kinetic energy, time-frequency (wavelet) properties, measures of anisotropy, and structure functions. Coherent turbulent kinetic energy (CTKE) highlights intermittent turbulent events and wavelet analysis of the CTKE can be used to predict loading events on the turbine. Anisotropy invariant analysis allows the anisotropy to be quantified in terms of one, two, or three (isotropic) component turbulence. Structure functions are typically used to predict spectral slopes or dissipation rates that are then used as inputs to synthetic turbulence generators for simulations of tidal turbines, but a more in-depth analysis can differentiate between isotropic, low-turbulence events and their counterparts which cause unwanted loads on a turbine. Not only do these results provide useful information about large, anisotropic eddies that affect tidal energy production, but they exemplify the wide range of possible quantities available from the simple velocity component observations of an ADCP. These quantities can be used to generate realistic inflow and boundary conditions for simulations of tidal turbines, and can also be used to validate results from the simulations.