Using Large-Eddy Simulations to Estimate Spatial Coherence and Power Spectral Density of a Major Hurricane for Wind Turbine Design Applications

Offshore wind development is underway along the US East Coast, where damaging hurricanes commonly occur. Even the most stringent wind turbine designs (IEC Class I) may not be able to withstand winds greater than a Category 3 hurricane.  Characteristics of the hurricane boundary layer (HBL) that can influence turbine loads, especially in major hurricanes (≥ Category 3 on the Saffir-Simpson hurricane wind scale), are poorly understood due to a lack of adequate observations across the rotor diameter (~126 m). We use a large-eddy simulation (LES) model (Figure 1), Cloud Model Version 1 (CM1), to simulate hurricane-like wind profiles of an idealized Category 5 hurricane, the strongest class of hurricanes and thus a “worst case” scenario for wind turbine designers. The model can simulate hurricane winds at high spatial (< 10 m) and temporal resolution (< 0.1 s). By comparison to the limited available observations, we find that a relatively “simple” and inexpensive configuration of the CM1 model accurately represents hurricane behavior in terms of mean wind speeds, wind speed variances, and power spectra (Figure 2). Comparisons of HBL power spectra to the Kaimal and von Kármán spectra suggest that the Kaimal model is a good representation of turbulence in the HBL. However, comparisons of spatial coherence from the model to theoretical coherence functions suggest that modifications are needed to accurately represent coherence in the HBL. Results also indicate that flow in the HBL remains highly coherent (≥ 0.6) for much larger separations (≥ 15 m) than is seen onshore and in a non-HBL atmosphere. Coherence of turbulent structures with characteristic sizes similar to 1/4th and 2/3rd the typical blade length (~ 63 m) are prevalent in the HBL which might increase loads on mechanical turbine components.