Uncertainty Quantification for an Automatised RANS-Based Wind Resource Assessment Approach

Abstract

Abstract In order to assess the model uncertainty of our RANS-based tool chain for wind-resource assessment “E-Wind”, we have simulated a large number of real wind farm projects with multiple measurement locations. Using one measurement location as reference and all others as target position, we systematically evaluated and analysed the horizontal extrapolation error for the annual mean wind speed introduced by the RANS method. Similar to previous studies, we found the model error to be strongly correlated with the simulated speed-up factor between two locations. We derived a method for site-specific prediction of the model uncertainty based on terrain complexity, simulated speed-up and distance between the measurement location and the turbine.