Integration of ARIMA and software models for wind speed forecast and noise map prediction in a wind farm

Wind turbine installation for electricity production is growing all around the world. The large energy demand, in fact, together with the environmental issue, leads to the promotion of renewable energy sources, with low (or zero) polluting impact. Wind turbines represent a sustainable solution, especially in areas with strong wind speeds and low density of obstacles. Anyway, there is a problem of annoyance related basically to noise production and landscape degradation. Regarding noise, it is almost evident that the greater is the wind speed, the larger is the noise emission, because of higher rotation speed of the blades and higher aerodynamic noise. Of course higher wind speed corresponds also to larger electricity production, thus a compromise must be found between energy and noise. For this reason, models able to exploit the relation between wind speed and noise map in the surrounding area of any wind farm, represent a strong tool to help policy makers in monitoring operating wind farms and in designing turbines placements. In this paper, the wind turbine noise issue is faced by different point of view. After presenting the noise maps in a wind farm in Italy, produced in the framework of a predictive software, in different wind speed and directivity conditions, a seasonal ARIMA model is presented. This model is used to predict wind speeds, to be given as input of the software for noise map drawing. This method will show that, starting from the time series of wind speeds in the wind farm area, strong information about the noise levels map can be obtained by means of accurate modelling of the area under investigation.