Most icing research focuses on the high Reynolds number regime and manned aviation. Information on icing at low Reynolds numbers, as it is encountered by wind turbines and unmanned aerial vehicles, is less available, and few experimental datasets exist that can be used for validation of numerical tools. This study investigated the aerodynamic performance degradation on an S826 airfoil with 3D-printed ice shapes at Reynolds numbers Re = 2 × 10⁵, 4 × 10⁵, and 6 × 10⁵. Three ice geometries were obtained from icing wind tunnel experiments, and an additional three geometries were generated with LEWICE. Experimental measurements of lift, drag, and pressure on the clean and iced airfoils have been conducted in the low-speed wind tunnel at the Norwegian University of Science and Technology. The results showed that the icing performance penalty correlated to the complexity of the ice geometry. The experimental data were compared to computational fluid dynamics (CFD) simulations with the RANS solver FENSAP. Simulations were performed with two turbulence models (Spalart Allmaras and Menter’s k-ω SST). The simulation data showed good fidelity for the clean and streamlined icing cases but had limitations for complex ice shapes and stall.

Computational Fluid Dynamics (CFD) is widely used to predict wind conditions for wind energy production purposes. However, as wind power development expands into areas of even more complex terrain and challenging flow conditions, more research is needed to investigate the ability of such models to describe turbulent flow features. In this study, the performance of a hybrid Reynolds-Averaged Navier-Stokes (RANS)/Large Eddy Simulation (LES) model in highly complex terrain has been investigated. The model was compared with measurements from a long range pulsed Lidar, which first were validated with sonic anemometer data. The accuracy of the Lidar was considered to be sufficient for validation of flow model turbulence estimates. By reducing the range gate length of the Lidar a slight additional improvement in accuracy was obtained, but the availability of measurements was reduced due to the increased noise floor in the returned signal. The DES model was able to capture the variations of velocity and turbulence along the line-of-sight of the Lidar beam but overestimated the turbulence level in regions of complex flow.

This paper presents an analysis of stratification-dependent mean velocity profiles measured in a Norwegian coastal wind climate, and its comparison with models available in the literature. For this purpose, we use 3 years of observations from a 100 m meteorological mast located at the Frøya island (150 km west of Trondheim, Norway), equipped with a set of two-dimensional (2D) ultrasonic anemometers. The presented analysis is preceded by a general description of the site wind climate, the atmospheric stratification, the roughness length, and the surface layer height. Finally, the measured wind velocity profile is compared with selected models: the basic power and logarithmic law and the stability-corrected models: stability-corrected logarithmic wind profile, the Panofsky and Dutton model, the Peña boundary layer height corrected model, and the correlation-based Smedman-Högström model.