A first approach to the optimization of landing and take-off operations through intelligent algorithms for compliance with the acoustic standards in multi-runway airports. (October 2021)

• Record Type: Journal Article
• Authors: Salata, Ferdinando; Falasca, Serena; Palusci, Olga; Ciancio, Virgilio; Tarsitano, Anna; Battistini, Vincenzo; Venditti, Andrea; Cavina, Lorenzo; Coppi, Massimo
• Is Part Of: Applied acoustics. Volume 181(2021)
• Publication Date: 2021-10
• Subjects: Airport noise -- Acoustic limits -- Neural networks -- Environmental protection -- Predictive simulations -- Balanced approach; Acoustical engineering -- Periodicals; Periodicals; 620.2
• DOI: 10.1016/j.apacoust.2021.108138
• Languages: English
• ISSNs: 0003-682X

Abstract

Highlights: Airport facilities and surrounding areas: economic opportunity and environmental issue. Multi-runway airports and regulatory thresholds for acoustic noise. Prediction of the acoustic impact on the surroundings according to the landing/take-off schedule and to the wind. Neural networks and their application in the prediction of aircraft noise based on historical data. Abstract: Noise limits for aeronautical traffic near airport infrastructure refer to energy contributions of sound and flight distribution among runways essentially depends on weather conditions. Therefore, the acoustic impact of air traffic on the surrounding area can be predicted in real time as a function of the runways use thanks to dynamic control. The problem can be solved thanks to the development of a predictive calculation model (based on machine learning and, specifically, neural networks) implemented from historical data obtained from monitoring systems and correlated with the monitored acoustic parameters. This approach borrows from other sectors new possibilities for optimizing and managing airport traffic in order to contain the noise generated by aircraft in transit, a possibility that until a few years ago was unexplored in these terms. As a first approach, an IT tool has been created for the identification in real time of a configuration of the runways use that guarantees the maximum airport operation and noise levels within the regulations. In this preliminary phase, the number of variables analyzed and the historical database used for learning the neural network are limited and an approximation of less than 1.3 dB is established with respect to the data recorded at the noise control units.