Automatic estimation of the sound emergence of wind turbine noise with nonnegative matrix factorization
Résumé
In many countries, the acoustic impact of wind farms is often constrained by a curtailment plan to limit their noise,which spreads in their surroundings. To update the plan, on/off cycle measurements are performed to determine theambient noise (wind turbines in operation) and residual noise (wind turbines shut down), but these shutdownoperations are limited in time, which reduces the representativeness of the estimated in situ emergence.Consequently, a machine learning technique, called nonnegative matrix factorization (NMF), is proposed to estimatethe sound emergence of wind turbines continuously, i.e., without stopping the machines. In the first step, theapplication of NMF on a corpus of various simulated scenes allows the determination of the optimal setting of themethod to better estimate the sound emergence. The results show the proper adaptation of the method with regard tothe influence of the propagation distance and atmospheric conditions. This method also proves to be efficient incases in which the real emergence is less than 5 dB(A) with a mean error lower than 2 dB(A). The first comparisonwith in situ measurements validates these performances and allows the consideration of the application of thismethod to optimize wind farm operations