Analytical assessment of wind turbine noise impact at receiver by means of residual noise determination without the farm shutdown

The present thesis, after an examination of the actual wind energy situation around the world, will summarize what is known in the scientific literature about wind turbine noise (WTN) generation, propagation, wind-noise interaction and the well-known phenomenon of the amplitude modulation. Extensive studies have been dedicated to the collection of international regulations, to the effects that WTN and annoyance have on human health, with a special attention to low frequencies for the WTN, considered strongly annoying by some authors and not by others. Particular attention will be given to the direct and the indirect health effects of noise, underlying that WTN is significant for the indirect ones, particularly for the annoyance. Since WTN results much more disturbing than the other most common noise sources, the reasons to consider annoyance as a health effect and its effects on people will be discussed. One of the aims of this work is to propose a specific limit for WTN and it will be reached by considering the annoyance dose-response relationships for different sources and deriving a limit value according with ones already issued for the other sources. Then, all the issues relating to the background noise measurement in a wind field will be analysed in details. By means of 9 long-term measurement campaigns performed ​​in Tuscany and Puglia , the purpose of providing suitable parameters to assess noise impact of a wind farm in a complex terrain is achieved. Moreover, the lack of a proper and complete regulation to consider the masking sound contribute to the noise level at the receiver, suggested the main objective of this thesis: a standard procedure to measure and to analyse noise and wind data, with the aim of characterizing wind turbine and background noise in all kinds of terrain. The procedure will be able to compare the results with national law limits, but it can be adapted to other international requests. Through iterative steps, the method provides the evaluation of noise impact produced by operational wind farms, without stopping the energy production for measurement purpose. The outputs of the procedure are the noise immission levels at receiver as a function of Neq, a new parameter that simultaneously considers the blades rotational speed of the whole wind farm, the residual noise at receiver as a function of vgr and the total immission and residual noise at receiver. The levels are evaluated on the measurement period distinguishing between night-time and daytime. In a specific Chapter a confidence level for the outputs of the procedure and an estimate of their uncertainty is acquired through a Monte Carlo procedure. Due to the complexity of the subject, where both the immission and the residual are not reliably predictable, the validation of the procedure outputs results a complex task. Nevertheless, it is performed in three alternative ways. The first is a comparison of the immission levels of the procedure with the noise prediction models, the second is a check of the difference between the measured 10 min environmental noise level and the environment levels on 10 min from the procedure. The third is a method based on the implementation of a computational model for simulated scenarios, where hypothetical set of measured noise levels are simulated summing a theoretical residual noise function to a theoretical immission noise and to a random noise. When applied to this set of data, the procedure should return the two inputs theoretical functions. At the end, a sensitivity analysis on the Neq parameter and a deeper explanation of some numerical parameters involved are presented, aiming to reduce the procedure uncertainty