Artificial Neural Networks Applied to the Simulation of Complex Aircraft Noise Scenarios

Abstract

The intensification of air traffic and urban intrusion generate critical situations of noise annoyance around airports and the estimation of aeronautical noise gains importance in the evaluation of traffic scenarios and in the definition of new landing and takeoff procedures aiming at reducing this annoyance. This estimation has been made by segmentation models of trajectories that generally do not present the temporal history of the estimated levels. To overcome this limitation and the lack of a complete analytical model, a dynamic approach is proposed, directly related to 4D aircraft trajectories. For this, the differential platitude of the aircraft's handling dynamics is used to generate, from the trajectories, the values ​​of some of the noise causal factors that are inputs to the estimation model. Thus, an artificial neural network allows the representation of the temporal evolution of noise at points in the vicinity of airports. The results are obtained and validated based on the Integrated Noise Model (INM). The tool is promising for complex analysis of the nuisance scenario in particular for problems associated with the dispersion of complex trajectories.