Influence of atmospheric conditions on measured infrasound from wind turbines

Scientific interest towards low frequencies and infrasound from wind turbines has renewed over the last decade as size of number farms have continuously increased. However, few comprehensive studies are found assessing such noise emitted by in response to meteorological parameters, including turbulent characteristics atmospheric boundary layer. The current work presents a detailed long-term measurement campaign combining recordings close high-resolution parameters at three heights. An artificial neural network (ANN) model, developed based on data, was able adequately predict both blade passing frequency levels broadband infrasonic noise. model allowed discriminating between inflow remaining microphone-induced under wind-shielding dome for wide range conditions. At same time, ANN deriving physically plausible relationships turbulence turbine components. Dominant predictors production were mean speed, intensity vertical heat flux.