Noise Prediction for Aircraft during Approach by Machine Learning Using Measured Sound Source Spectra, Flight Parameters, and Aircraft Specifications
- Resource Type
- Journal Article
- Authors
- Takehisa TAKAISHI; Taro IMAMURA; Tomohiro KOBAYASHI; Yuho IKUTA
- Source
- TRANSACTIONS OF THE JAPAN SOCIETY FOR AERONAUTICAL AND SPACE SCIENCES, AEROSPACE TECHNOLOGY JAPAN. 2023, 21(APISAT-2022):1
- Subject
- Aircraft Noise Prediction
Machine Learning
Neural Network
Sound Source Modeling
- Language
- English
- ISSN
- 1884-0485
Noise prediction for aircraft during approach is necessary for environmental assessment around airports. This study aims to develop a component-wise sound source model of aircraft during the approach using flight parameters and publicly available aircraft specifications. By using the aircraft specifications as input parameters, different types of aircraft can be described in a single model. The component-wise sound source spectra measured using a phased microphone array are related to the flight data and publicly available aircraft specifications using a neural network to obtain the sound source model. The developed model can predict component-wise sound power spectra, and the variation of the spectra in terms of different flight conditions can be evaluated. The presented model predicts the sound source spectrum with an accuracy of 2–3 dB. Based on the developed sound source model, noise on the ground is calculated with an accuracy of approximately 2 dB.