PhD Defence

Erik Janzon´s PhD Defense

Title: Local Effects On Icing Forecasts for Wind Power In Cold Climate

Speaker: Erik Janzon

Description: The local effects of the atmospheric response to land surface properties are examined with an emphasis on the impacts to wind turbine icing predictions.

Time: 10 February 2023 10:00

Supervisor: Anna Rutgersson

External examiner: Dr. Alfredo Peña, Senior Scientist, DTU

Place: Hambergsalen, Geocentrum

Abstract:

Understanding the risk of atmospheric icing on wind turbines is crucial for the operation of wind farms in cold climate regions. Processes leading to atmospheric icing in the planetary boundary layer are subject to interactions with the land surface--the details of which are largely unknown in current numerical weather prediction models and must be parameterized. This thesis examines the impact of the representation of land surface cover in these models on meteorological variables related to wind turbine icing. First, a sensitivity analysis is conducted using a single-column model to test the relative impacts of the vegetation fraction on ice accretion at the elevation of a modern commercial wind turbine. The impact of the representation of surface roughness due to the vegetation fraction and parameterization of processes related to the forest canopy is found to have the largest impact on icing in the simulations. This effect is combined with an important secondary role due to the albedo of the surface, which impacts the evolution of the icing event under early season solar insolation. Next, large eddy simulations (LES) are conducted to test the impact of the land cover heterogeneity length scale on wind profiles during a semi-idealized diurnal cycle in dry, subarctic conditions. The effective surface roughness is found to decrease as a function of the land cover heterogeneity length scale and the blending height is found to be limited by the height of the atmospheric boundary layer. Using the findings of this study in combination with previous work in the literature, a dynamic blending height model is presented as a possible coupling strategy for surface parameterizations. Following the results of the aforementioned study, LES simulations of different heterogeneity length scales are used to validate an analytical model that is used to simulate mean wind profiles and wind stress over arbitrary patterns of surface roughness. Lastly, semi-idealized LES simulations of supercooled low-level clouds are used to test the sensitivity of meteorological variables related to wind turbine icing to land cover heterogeneity length scales. Here, the presence of snow cover in bare patches has a significant effect on the icing prediction. The results of this thesis show that the representation of the surface can have a significant impact on wind turbine icing forecasts and that these findings will be helpful in identifying uncertainties in these predictions.